DOU. E 1.28: WIT-2295T9- 7 I Ee'^^ -lACKS MIT-2295T9-7 COMPUTER-AIDED INDUSTRIAL PROCESS DESIGN, THE ASPEN PROJECT Functional Specifications for Aspen Sixth Quarterly Progress Report, Appendix 1 December 15,1977 Date Submitted «« 'T:^.^^'^^ ql u\e^S^^ VO«:Jr^t*" Work Performed Under Contract No. EX-76-C-01 -2295-009 Department of Chemical Engineering and Energy Laboratory Massachusetts Institute of Technology Cambridge, Massachusetts U. S. DEPARTMENT OF ENERGY NOTICE This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Department of Energy, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. This report has been reproduced directly from the best available copy. AvaUable from the National Technical Information Service, U. S. Department of Commerce, Springfield, Virginia 22161. Price: Paper Copy $9.50 Microfiche $3.00 UNIVERSITY OF ILLINOIS LIBRARY AT URBANACHAMPAIGM STACKS COMPUTER-AIDED INDUSTRIAL PROCESS DESIGN The ASPEN Project FUNCTIONAL SPECIFICATIONS FOR ASPEN Appendix I to The Sixth Quarterly Progress Report ^IT-2295T9-7 s UC-90C, UC-90d Department of Chemical Engineering and Energy Laboratory Massachusetts Institute of Technology Cambridge, Massachusetts 02139 Date Submitted: December 15, 1977 PREPARED FOR THE UNITED STATES DEPARTMENT OF ENERGY FOSSIL ENERGY PROGRAM Under Contract No. E (49-18) -2295 Task No. 9 The following members of the Project staff contributed to this report: Mr . Aiftarquaye Armar Dr. Joseph F. Boston Dr. Herbert I. Britt Mr. Chau-Chyun Chen Professor Lawrence B. Evans Dr. Paul W. Gallier Dr. Prem Gupta Dr. Babu Jos3ph Dr. Vladimir Mahalec Dr. Elaine Ng Mr. George Randall Professor Warren D. Seider Mr. Siri Tia Mr. Hiroshi Yagi m TABLE OF CONTENTS Page 1. INTRODUCTION 2. FUNCTIONAL REQUIREMENTS SPECIFIED IN THE CONTRACT 3. THE ASPEN USER 13 4. ASPEN EXECUTIVE SYSTEM FUNCTIONS 16 5. COMPUTATIONAL ALGORITHMS FOR HEAT AND MATERIAL BALANCE 37 6. PHYSICAL PROPERTY AND DATA REGRESSION SUBSYSTEMS 61 7. UNIT OPERATIONS SUBSYSTEM 149 8. COST ESTIMATION AND ECONOMIC EVALUATION SUBSYSTEM 199 IV 1 . INTRODUCTION The foL-mai statement of the functional requirements of the ASPEN system is contained in the statement of work for the contract between M.I.T. and the Department of Energy (Contract No. E(49-18)-2295 Task No. 9). A summary of these requirements is presented in the next section. The purpose of this document is: (1) to expand on the work statement and resolve ambiguities, (2) to describe what functions the system will perform in sufficient detail to permit system design, and (3) to serve as a communication vehicle with the Advisory Committee and the Department of Energy. It is important to note that these functional specifica- tions are intended to define what the system will do, but not necessarily how it will be accomplished. In practice, some idea of how the functions might be accomplished is necessary to understand the functions. Therefore, where we know of ways in which functions might be in.plemented, we have stated them to improve understanding of tha functional specifications. Separate sections discuss the ASPEN executive functions, the computational architecture, the physical property and data regression subsystems, the unit operations subsystem and the cost estimation and economic evaluation subsystem. The results of the survey of the Advisory Committee on the Preliminary design criteria and the discussion by the User/Use Subcommittee of the System Design Task Force at their meeting on August 21-22, 1977 have been used in preparing these func- tional specifications. This document will serve as the basis for ASPEN system design specifications. 2. FUNCTIONAL REQUIREiMENTS SPECIFIED IN THE CONTRACT This chapter summarizes the functional requirements of ASPEN as taken directly from the statement of work in the contract between M.I.T. and the Department ot Energy (Contract No. E(49-18)-2295 Task No. 9). Objectives The main objective of this program is to provide for the Department of Energy a rapid, efficient and consistent means of performing its process evaluation functions. The capability will be embodied in a computer based process simulator and process economic evaluation system known as ASPEN (Advanced System for Process Engineering). The system so developed will be designed to meet the specialized requirements characteristic of fossil energy conversion processes: extensive data base for coal physical properties, compatibility with conversion reactor models currently available and/or being constructed, routines for solids handling, and routines for waste product recovery systems. ASPEN will be capable of providing detailed heat and material balances and detailed economic analysis for process plant construction and operation. In developing ASPEN a key objective will be to assure that information requirements (physical properties, processing equipment, etc.) for the system are the same as would be supplied to any group of competent engineers involved in a process evaluation. Additionally, the results generated by the system shall be in a form normally associated with assessing process economic viability (cost of products, capital and operating costs, rate of return on investment, etc.). The ov-rall objectives cited above will be met through the development of a steady state simulator with the capability for evaluating fossil energy processes (gasification, liquefaction, etc.) ASPEN System Structure The staady-state simulator will compute the steady state performance of a process, including the process flows and physical and chemical properties for all intermediate and product streams. it will also compute equipment sizes, and utilities requirements. An executive system will be developed to provide flexi- bility, and yet convenience for the user, to model different process configurations for fossil energy process design. The executive will control: data input reading, physical property and other data bank retrieval, execution of calculations to model process units, sizing and costing of process equipment, cost analysis calculations and output reports. The input language will be made to be as familiar and con- venient to fossil energy process designers as possible. To accommodate different computer systems, the ability to run in batch and on-line operations will be provided. A variety of options will be provided for computing physical properties (different equations for vapor pressure, equations of state, activity coefficients, etc.) and it will be possible for the user to specify the computational method to be used. The user may also add new physical properties and their methods of computations. Default options will be available for the user who prefers to use standard methods. The executive system will allow insertion of proprietary programs for use by industry. It will provide for handling multiphase flows (solids, vapors, and liquids) and multiphase handling in unit operations, especially for coal. Equipment sizing and costing calculations will retrieve information from data banks and the heat and material balance. calculations. In addition various types of economic analysis will be possible for the process model. Physical Property Subsystem The physical property subsystem will compute properties of pure species and mixtures important in fossil fuel conversion processes. Properties to be computed include volumetric properties (density and specific volume), thermodynamic prop- erties (enthalpy, entropy, free energy), transport properties (viscosity, diffusivity, and thermal conductivity), and equilibrium properties (phase-equilibrium ratios for gas- liquid, gas-solid, liquid-liquid and liquid-solid and multiphase systems) . The syctem may be used independently through calls on sub- routines and will be accessed by the steady-state simulator discussed. One or more system data banks will be provided containing pure-compound constants (such as molecular weight, critical properties, coefficients in correlations, etc.) and interaction parameters for deterimining mixture properties. The user may add constants for species not in the data bank or may change the constants in a private file for use in a specific simulation study. A data bank containing the important properties of at least (but not limited to) twenty standard types of coal will be provided and there will be provision for adding properties of new types of coals and physical properties not in the original system. Unit Operations Subsystem Building-block subroutines will be developed to model all of the common unit operations in fossil energy processes including but not limited to stream adders, steam splitter, distillation column, adsorption columns, extraction units, pumps, compressors, expanders, turbines, equilibrium flashes, heat exchangers, heaters/coolers, condensers, furnaces, cyclone separators, crushers, grinders, screens, kilns, fluidized bed heat exchangers, filters, centrifuges, driers, electrostatic precipitators, crystallizers, etc. The best available algor- ithms for convergence of racycle streams will be implemented. There will be provision for automatically adjusting process equipment parameter (s) to achieve specified value(s) of stream variables for the purpose of design of processes. Subprograms will be included to model standard reactor types such as the equilibrium reactor, well-mixed reactor, homogeneous tubular reactor, catalytic fixed-bed and fluidized- bed reactors. Where models exist of specific reactors involved in fossil fuel conversion processes (such as gasifiers, lique- faction reactors, oil-shale retorts, etc.) they will be pro- grammed and included in the system. programs will be provided to model the major process units for sulfur recovery (such as the Claus reactor, amine scrubbers, etc. ) . Cost Estimation and Economic Evaluation Subsystem The cost estimation and economic evaluation subsystem will size the equipment, and estimate the capital investment, oper- ating costs, and profitability of the process. Two levels of accuracy will be provided: (1) Order-of-magnitude (within 100%) with capital investment based on power-law cost correlations; (2) Study estimate (within 30-50%) in which costs of individual items of equipment are determined based on weight of steel, factors for construction, etc. The input to the Cost Estimation and Economic Evaluation Subsystem will be the process flowsheet with operating con- ditions of all items of equipment, flows of feed and product 10 streams, and utilities required by each unit. The subsystem will size and cost equipment and calculate utility require- ments. The level of detail of the flowsheet description must be consistent with the required accuracy of the estimate. The user may specify economic parameters (such as project life, production rate profile, inflation rate, etc.) and may specify the measure of profitability (discounted cash flow rate of return, payout time, return on initial investment, etc.). A data bank will be provided containing size and cost data for process equipment, maintenance costs, labor categories, raw materials, product values, utilities, etc. There will be provisions for alternative sets of data from different sources (such as the American Society of Cost Engineers, government data compilations, proprietary company sources, etc.) and the user may specify the set to be used for each or all items. A mechanism will be provided to update this data bank to keep the cost data current. Data Regression Subsystem A user may wish to simulate a process for fossil fuel con- version where he may find physical property correlation con- stants unavailable for some pure components or coal liquid fractions. A data regression subsystem is to be developed to assist the user. The ASPEN system, as with many process simu- lators, will use correlating equations for physical properties over certain ranges of temperature, pressure, etc. Constants in these equations will be required as input. 11 For pure components this subsystem will be able to use laboratory, plant, or literature data to regress the corre- lating equations for best fits. Data may be used to correlate such properties as density, vapor pressure, solubility and heat capacities. Similarly the correlating equations for phase equilibria (i.e., vapor-liquid, liquid-liquid, vapor-solid and solid- liquid) require input constants. Again, laboratory, plant, or literature data may be used to determine the best correlating constants. A wide variety of correlating equations will be made available. Specification of various objective functions will be made possible for obtaining the desired accuracy of fit. The system will automatically determine the best regres- sion constants from the experimental data. Integration, Te stin g, and Documentation In order to demonstrate the ultimate capabilities of the system developed, three representative processes will be subjected to detailed analysis by the simulator. These bench- mark problems will be used to comprehensively check the validity of the simulator and all of its subsystems. The benchmark problems that will be developed will be characteristic of those for which the simulator will be intended to solve. The problems will be designed to cover a range of types of flowsheet analysis including: . 12 (a) Steady-state simulation (b) Equipment sizing and economic evaluation (c) Determination of process parameters from plant operating data. Each of the problems will refer to a specific process and will be stated in sufficient detail that one could obtain a solution manually or by writing special purpose programs. Documentation will be supplied at the conclusion of the project for proper use and maintenance. A user's manual will be provided which will guide the process design engineer. A systems manual will be provided which will facilitate mainten- ance by computer programmers and permit modification or enhancement of the system. .•>.:• y^a 13 3. THE ASPEN USER 14 ASPEN is intended to be a tool for the Department of Energy for performing its process evaluation functions. Hence the main users of ASPEN will be D.O.E. personnel and contractor personnel engaged in evaluating, developing, designing, opera- ting and improving fossil energy conversion processes. In addition, because the systen will become publicly available, prospective D.O.E. contractors, private industry and consul- tants may have access to ASPEN for simulating fossil energy processes. ASPEN will have users at three levels. The first level will be for users who are process engineers familiar with heat and material balance calculations around a flowsheet. These users need not be familiar with computer programming and/or computer methods for process flowsheeting. Their use will be limited to models already built into the system. They may however supply their own physical property and cost data. The second level users are process engineers who are familiar with computer programming and computer process flow- sheeting methodology. These users will be capable of adding their own models for their simulation. They will be able to make use of the advanced features of ASPEN in computational methodology and flexibility to deal with new kinds of pro- cesses. They may extend the system to do design calculations and process optimizations ordinarily not done by ASPEN. 15 The third level users will be engineers and scientists who are interested in making modifications to the system. These users will have to be familiar with the system documentation. They may change any subprogram within ASPEN or add new sub- programs to ASPEN. Typically this may represent a simulation group who has implemented and maintains an in-house version of ASPEN. 16 4. ASPEN EXECUTIVE SYSTEM FUNCTIONS 17 4.1 Introduction The description of the ASPEN Executive System requires some understanding of the basic functions of a process simulator. The purpose of this introduction is to outline the simulation process from the user's point of view with the objective of defining some of the terminology used in the rest of this docu- ment. The user begins with a process flow diagram of the process that he wants to simulate. A process flowsheet is a collection of process units which are interconnected by streams to repre- sent the flow of material and information. These process units may represent actual pieces of hardware such as heat exchangers, reactors, distillation columns. Sometimes one process unit may represent more than one piece of hardware (and vice versus) as in a unit for a separation column which represents a tray dis- tillation tower with reboiler, condenser and interheaters. Figure 4.1 shows a typical flow diagram. The first step in the simulation of the process is the preparation of a block flow diagram of the process. Here the unit operation elements of the process flow diagram are replaced by blocks. Blocks also represent physical mixing or separation of streams which are not shown as equipment in the process flow diagram. A block flow diagram consists of blocks interconnected by streams. Streams represent flow of information between blocks. This information represents data such as flow rate and composition of a material flowing from one block to another. The function of a block is to transform the inlet streams to ••V*' 18 METH. REACTOR #1 PRODUCT GAS i RECYCLE GAS Figure 4.1 Example of a Process Flow Diagram. Figure shows methanation section of a coal gasification plant, 19 produce the outlet streams. This transformation is related to the actual unit operation represented by the block. The trans- formation may be simple such as a change in stream temperature as in a heat exchanger or complex as in a change in composition in a plug flow reactor. Figure 4.2 shows a block flow diagram of the process of Figure 4.1. The transformation in the blocks is accomplished by a model of the actual unit operation taking place in the process. In ASPEN there is a library of subroutines which represent the various types of transformations. These subroutines are here- after referred to as the unit operations models and are discussed under the chapter on Unit Operations Subsystem. The connections between various blocks via streams is referred to as the connectivity information . This information is important in determining the sequence of computation. Each block and stream in the block flow diagram is assigned a unique identification and the connectivity is specified in terms of these identifications. The user determines what unit operations model is to be used for each block in the process. The composition of the streams is expressed in terms of components selected by the user. These components may be elements, compounds or mixtures depending on the process. During the simulation, ASPEN may require physical property constants such as heat capactiy and vapor pressure for these components. These data may be selected from the data bank or supplied by the user. 20 Sll Figure 4.2 Example of a Block Flow Diagram for the process shown in Figure 4,1. 21 Another aspect of the simulation is the block data . Each block will require a set of variables necessary to characterize its unit operations model. These may be fixed parameters such as number of trays in a distillation column or temporary vari- ables such as temperature and pressure on each tray. Other information required may be the type of physical property methods to be used and convergence tolerances for iterative calculations within the block. The data structure of a process consists of component data, connectivity data, stream data and block data associated with a process. The first step in the simulation of a process is the building of this data structure from user supplied information, data bank and the results of earlier simulation (if available and if requested by the user) . 4.2 Information Flow in an ASPEN Simulation Figure 4.3 shows the sequence of steps performed by ASPEN during the simulation of a process. The first step is the processing of the input data supplied by the user to set up the data structure. This data structure will have space allocated to store data regarding streams, blocks and components. In addition this data structure will contain other information applicable to the simulation such as level of output calculation control, etc. The executive will extract information from the data bank, such as component data and enter these into the data structure. The data structure then becomes a pool of informa- tion concerning the process being simulated. The executive will / \ hJ z < o CO en hH UJ q: LU o S [-• u w Oi a. i-H H <: Z E- 3 =^ ^J S J UJ UJ Q 3 Z3 S UJ "S H O ^ d o'^ u h ? ^ J uj ex O W Cu J < C/5 < Oi c/3 s: < L3 D- CU o H < CO UJ H w tu < uj a: Q u CO O 2: • Qi O D- CU hJ • 22 2 tu CO 3 o o I— I c o •H a 5-1 O c ro ■H ind.M N0I13V annvD S3NIinOH ifidino 23 store both the user input data as well as the data structure created, on secondary storage for future use. The executive will check the input data for obvious errors and inconsistencies and generate appropriate error messages. The user may halt the program before proceeding to the next phase of the simulation. This is to facilitate checking of the messages generated by the input processing phase. The next phase of the simulation is to set up the computa- tional algorithm to do the heat and material balance. This involves determining the sequence of calculation and the inser- tion of additional convergence blocks to facilitate the solution of recycle problems and control blocks to facilitate the solu- tion of design problems. The user may override any or all of these decisions through inputs to this phase of the simulation. The inputs may specify certain calculation sequences, specific tear variables and partition the process block diagram into independent parts over which the computations are to be per- formed. The output will be a specific order in which the heat and material balance calculations are to be performed. This is added to the data structure and stored in the secondary files also. The user may examine the output, make changes and repeat the step to obtain a satisfactory result. The heat and material balance calculations are performed next, and this would require program modules in the Unit Operations Subsystem as well as the Physical Property modules which are called by the unit operations subroutines. The user may input at this stage, the process feed variables, certain 24 block parameters, and calculation control variables. The user may specify the level of intermediate results generated. The executive will update the data stored in the files to reflect the new calculated values and the user may specify how often this is done. The output from this system will be a flowsheet with all stream variables calculated such that heat and material balances are satisfied. The next phase is the sizing of the various pieces of equip- ment. This would require additional sizing parameters to be supplied by the user. The sizing modules may access the data bank for obtaining specific information. The size and utility requirements thus calculated is appended to the data structure. This may now be input to the economic evaluation subsystem to do the profitability analysis of the process. The above discussion identifies the functions which will be performed by the ASPEN Executive. These are (1) Process Input Data (2) Set up Data Structure (3) Set up and Execute the Computational Algorithm to Solve Heat and Material Balance (4) Manage Files in Secondary Storage (5) Produce Output Reports 4.3 PROCESSING INPUT DATA In this section, some g2neral requirements on the input language are discussed. In chapter 3, it was pointed out that the typical user of ASPEN will be a process engineer familiar ■AvvKKlnatkr^s 25 2) 3) with heat and balance calculations around a flowsheet. He is not expected to know computer progranuning and/or computer methods for process flowsheet. Hence the input language for ASPEN should be easily understood by a process engineer. There are three (conflicting) requirements on the structure of the language 1) The language must have enough structure and verbosity for someone else (or the process engineer after he has been away from the application for a few months) to be able to quickly understand the essence of the model. The capability for the process engineer who is familiar with the system to input the model quickly with a minimum of excess typing. Input should be free format. The potential for a f ill-in-the-blanks capability which would help lead the novice through the modeling process, The final design of the input language must necessarily find a compromise between the three goals. The other requirements on the input language relating to its content are; 1. User should have the flexibility in choosing input and output dimensional units, 2. The input may include arithmetic and logic statements (similar to those in FORTRAN) manipulating stream and block data. The requirements on input data processing are 1. The data should be checked thoroughly for consistency, engineering soundness and out-of-range variables. 2. Whenever possible, the system should supply default values for missing data. 26 3. The system should provide a copy of all input data to the system with the default values entered and clearly marked. This should serve as a permanent record on process being simulated. The input data for ASPEN consists mainly of the data regarding the process (components, streams and blocks) and computational data for controlling the execution. These are discussed next. 1. Component Data Component data consists of the names and abbreviations for the components present in the system, the name of the data bank from which data regarding the components are to be extracted, data for items not in the data bank and/or items for which the user wishes to supply his own data and binary constants. 2. Stream Data Stream data identifies the names and abbreviated names for streams present. Since ASPEN would allow different types of streams the user may specify what phases are present in the stream and what data he likes to carry as stream variables. In addition ASPEN will allow the user to specify what components are present in a stream through the concept of a stream class . All streams belonging to one class have the same phases and components present in it. The default will be that all com- ponents are present in all streams. Similarly if the user does not specify the phases present in a stream it will be assumed to be a vapor-liquid stream. The stream data may include values of known stream variables. 27 3. Block Data Block data consists of names and abbreviations for each block in the block flow diagram. The user must specify what model he is going to use for each block and this in turn would determine what additional data is required for each block. The user specifies the connectivity by including the names of streams entering and leaving the block. Each block will have associated with it a list of parameters which may be specified by the user. Some blocks will also allow the user to specify guess values for certain variables internal to that block. An example of this is the initial temperature profile for a column. The user may specify these parameters using identifying labels. Since the unit operations models may call upon physical properties, the user may specify what estimation methods are to be used in the model for the block. If convergence calculations are to be used in the model for the block, then the parameters and methods for convergence may also be specified by the user. If the user is interested in debugging then he may specify the level of output desired during the execution of the block. 4. Computational Data The user may specify the computational order in which the system will perform the heat and material balance calculations. These include data on partitioning of the system, specification of tear streams, and guess values for tear stream variables. 28 5. Control of Execution These input include the starting and ending point of calcu- lation, execution time limits and the maximum number of errors which will be tolerated before the system aborts execution. 6. Input/Output Control The user may specify the input and output dimensional units. There will be standard sets of units available such as SI units or Engineering units with user option to override the unit as any specific quantity. The user may change the set of units during the input, that is he will not be required to use the same set of units throughout the input. The user may name the different files which will be created during the run and indicate their ultimate disposition (whether to save them after execution) . The user may specify the level of intermediate results end diagnostics described during the simulation. There will be provision for using the results of a previous run to start a new run. In this case most of the input data will be read from a previously created file. In the case of output reports, the user will have many options including summary of stream and block data, detailed values of streams and block information, physical propepcty data used, details of specific streams and blocks, overall heat and material balance and heat and material balance for each block in the flowsheet. There will be options for number of streams/page and the page size of the output reports. 29 4.4 Data Structure Requirements The term data structure refers to the representation of various pieces of data in core during the simulation of a process, ASPEN consists of many different modules and each module needs access to certain items of data. The way the data structure is set up determines the way the modules communicate with each other. This also includes input/output modules and modules which are responsible for storing information on files for future use. It is important to have a data structure that will permit easy communication links between various modules. One of the drawbacks of today's simulators is their rigid data structure. For these simulators the data structure is such an integral part of the system that it is difficult to extend the system to simulate new types of processes such as those encountered in coal conversion. The data structure of ASPEN should permit the description of new types of streams and com- ponents. These include streams containing multiphase solids. During the course of a simulation study, the user may want to make modifications to the process structure such as adding or deleting streams and/or blocks. The data structure should allow these types of modifications easily. There is another conflicting requirement on the data structure, that is it should make efficient use of time (fast execution) and space (minimum use of core memory). The criteria to be used is that ASPEN should perform comparable to todays exsiting simulators for conventional vapor-liquid processes. Finally, the data structure should not impose restrictions on problem sizes such as maximum number of streams blocks and components. 30 4.5 File Management Function s ASPEN will have the capability of storing input data, inter- mediate results and final results on files. In addition, the data banks will normally reside in secondary storage and the ASPEN Executive will access these banks for data during a simu- lation. The details regarding the creation and maintenance of the physical property data bank is discussed in chapter 6 and that of size and cost data bank is discussed in chapter 8. During the simulation a number of different types of files are created by the ASPEN Executive. These are (1) Input Data File (2) Problem File (3) Output Report File (5) Calculation History File 1. Input Data File This file contains a description of the problem as input by the user. It will have an echo of the user's input data. In addition this file will contain a problem definition, complete with defaults and data labels. This file will serve as a per- manent record of the problem for future reference and contain enough iniiormation for somebody other than the user to under- stand the problem. The user may use this file to rerun a simulation. 2. Problem File This is also a description of the simulation problem but in terms understandable to ASPEN. This file contains an image of the data structure for the process, and any other information 31 generated during the simulation. There will be space allocated in the data structure for all streams, blocks and components in the process and hence all this data will be saved in the problem file. If the executive creates calling programs as in a pre- compiler, such information will be stored in this file. The results of the algorithm to determine the computational sequence are saved in this file. During the course of a simulation this file will contain a current description of the problem. The file serves as an information transfer medium between the different sections of ASPEN such as heat and material balance, sizing, costing and economic evaluation. 3. Output Report File The output reports requested by the user are saved in this file. The user may opt to send the reports directly to the printer if he so chooses. However having this option will enable the user to examine selected parts of the output through a time-shared terminal and then get a hard copy of the report. The different kinds of reports available are described in section 6 of this chapter. 4. Calculation History File During a run, ASPEN will generate many messages, inter- mediate results and diagnostics. All these output are saved on this history file. For example, during a test run, the user may request extensive printout of intermediate variables. The results are output to this history file. Again the user may request that all history be sent directly to a printer thus avoiding the creation of a new file. 32 Figure 4.4 shows the interaction of various parts of ASPEN with these files. The problem file may form the link between different sections of the simulation. In addition ASPEN will have three libraries of program modules. These are 1. Library of Unit Operations Modules 2. Library of Physical Property Modules 3. Library of Sizing, Costing and Economic Evaluation Modules Each of these libraries are discussed in more detail in subsequent chapters of this report. 4.6. Report Writing Requirements Report writing routines will extract information from the Problem Data File and output them in a form generally used by process engineers, and process evaluation groups. It will produce the following types of reports at the user's request: 1. Problem Description (process configuration, unit operation models used, feed variables, block parameters, constraints, etc.) 2. Overall Material and Energy Balance Summary (feed stream and output stream variables, closure check) 3. Stream Summary (component flows, temperature and pressure for all streams. User may include or exclude specific streams.) m^rn^ 33 4. Table of Stream Variables (component flows, temperature, pressure and other user requested properties of all streams. User may include or exclude specific streams and specific stream attributes) 5. Table of Block Parameters uUser may include or exclude specific blocks) 6. Material and Energy Balance Check for Each Block 7. Sizing and Cost Data for Each Equipment in the Process 8. Economic Evaluation Report (capital costs, operating costs, profitability analysis, etc.) 9. Table of Physical Property Data Used 10. Computation Sequence for Heat and Material Balance 11. Analysis of Streams Such as Heating and Cooling Curves The reports will be numbered and there will be a table of con- tents for each report. 34 X \ USER'S INPUT DATA , PREVIOUSLY ^ ( PREVIOUSLY^ CREATED PROBLEM ^ FILE \ ^ T - " \ CREATED INPUT Flowsheet Description J xwruT \ DATA FILE ' A / / / MESSAGES Problem Definition Echo of Input Cards FIGURE 4.4a. Processing Input Data User Specified Constraints Flowsheet Description DETERMINE COMPUTATIONAL ALGORITHM MESSAGES Flowsheet with Computational Order FIGURE 4.4b. Determining Computational Algorithm -nDaaKHM^EApvyn(i^K»ria>t)ir' 35 'lowsheet with Heat and Material Balance 'PROBLEM^ FILE FIGURE 4.4c. Heat and Material Balance Calculations Flowsheet with Heat and Material Balance 1 SIZING AND COSTING OF EQUIPMENT Flowsheet with Sizes and Costs FIGURE 4.4d. Sizing and Costing of Equipment 36 Flowsheet with sizes and costs ECONOMIC EVALUATION Economic Evaluation Reports OUTPUT^ REPORT FILE FIGURE 4.4e. Economic Evaluation Flowsheet FIGURE 4.4f. Report Writing 37 5. COMPUTATIONAL ALGORITHMS FOR HEAT AND MATERIAL BALANCE 38 5.1 INTRODUCTION The simulation or design problem to be solved by ASPEN is represented by a block diagram. Each element of the block diagram is modeled by a subroutine which describes either a part of, or an entire process unit or a group of process units. Blocks are connected by streams. Streams can be either material or energy or information or any user defined stream type. The solution of the flowsheet heat and material balance problem involves the determination of stream variables and possibly block parameters so that a heat and material balance exists for each blpck in the system and the user's specific- ations are satisfied. Several computational schemes have been proposed to solve this problem but only the sequential modular technique has been widely implemented. Since sequential modular is known to work, has been thoroughly studied, and is easy for process engineers to understand, this technique will form the basic ASPEN computational procedure. Other techniques will be considered with concern about how they could co-exist with the basic sequential modular system, so that ASPEN can be extended at a future date to incorporate these new techniques as they are developed. The ASPEN system will utilize the process structure in order to solve efficiently simulation or design problems. A process flowsheet can often be divided into subsystems which can be computed individually without any information from downstream subsystems. This approach is useful regardless of the actual 39 numerical algorithm. Algorithms in ASPEN will identify sub- systems which can be computed individually (so called partitioning of a process flowsheet) . These subsystems are either single blocks or maximal cyclic subsystems (portions of a flowsheet with recycle flow of material, energy or information) . Each maximal cyclic subsystem identified by the partitioning algorithm can be analyzed further to determine what variables determine the cyclic structure (e.g. recycle streams) . Such variables will be called tea r variables , since one can convert a cyclic system into a serial system by assuming values for the tear variables. Tear variables can be used either to generate a sequential computational procedure, or to create efficient com- putational and/or storage schemes for some other algorithms. The ASPEN system will determine a set of tear variables according to the requirements of a particular computational algorithm (minimal tear set, optimal convergence properties, etc.). It should be emphasized that the procedure to determine the tear set is always the same; only the criteria change. The partitioning of process block diagram and tear variable selection is common to all computational procedures in ASPEN. From that point, algorithms proceed differently. We shall describe the structuring ol process computations (partitioning and tearing) common to all algorithms, the sequential modular algorithm and the extensions of the computational methods in ASPEN (simultaneous modular and implicit convergence methods) . 40 5.2 STRUCTURING OF PROCESS COMPUTATIONS 5.2.1 PARTITIONING OF A PROCESS BLOCK DIAGRAM DESCRIPTION The structure of the process block diagram is analyzed in order to determine what subsystem can be computed individually. The simplest such subsystem is a single block but often the entire flowsheet may represent a system which has to be con- verged simultaneously. INPUT DATA Connectivity information about the process block diagram, including design specifications. OUTPUT DATA Maximal cyclic subsystems ordered in a sequence corre- sponding to direction of flow at the block diagram level. REQUIRED SPECIFICATIONS None OPTIONAL SPECIFICATIONS None 5.2.2 TEAR VARIABLES SELECTION DESCRIPTION Each maximal cyclic subsystem identified by the partitioning algorithm is analyzed to determine what set of variables intro^ 41 duces cyclic structure. Typically, these correspond to recycle streams as well as to variables manipulated in order to achieve design specifications. Branch and bound algorithms will be employed to find a tear set which satisfies the requirements of a particular computa- tional algorithm. The default option is to find a tear set which cuts feedback loops (fundamental circuits) a minimal number of times. INPUT DATA Connectivity information for a maximal cyclic subsystem. Information streams specifications. OUTPUT DATA Set of tear variables ADDITIONAL RESULTS Set of fundamental circuits (loops) in the maximal cyclic subsystem. REQUIRED SPECIFICATIONS None OPTIONAL SPECIFICATIONS - Criterion to be used to define the tear set. Possible choices are: system default, minimal number of tear set variables, tear set with a minimal total weight of tear variables (weights specified by the user) . 42 - Partial or complete set of tear variables Type of computational procedure to be used to compute material and energy balances. If more than one category of optional specifications is provided, they should not conflict with each other. ERROR CONDITION Non-fatal. Specified (sub) set of tear variables defines a tear set which is not optimal for the chosen computational procedure. ■AviniiikiKK/'/*^ 43 5.3 SEQUENTIAL MODULAR ALGOR I THM Based upon selected tear streams, the computational sequence for a serial structure will be generated. The user will be able to: request the system to generate the computational sequence supply his own specifications for some or all phases of the procedure to generate the compu- tational sequence. Regardless of whether the user supplies data for some or all of the above steps, the system will perform a complete analysis of the flowsheet. If needed, the user's specifications will be complemented with additional sequencing. If a user's sequence contains untorn recycles the system will print a message but will still accept the user sequence. Results of algorithmic analysis of the flowsheet by the system are always delivered to the user for future work. 5.3.1 DEFINITION OF COMPUTATIONAL SEQUENCE DESCRIPTION Partitioning and tear stream selection enables construction of serial computational procedures for the process block diagram. Maximal cyclic subsystems are computed one after the other in a sequence determined by the partitioning algorithm. A separate computational sequence is determined for each maximal 44 cyclic subsystem prior to the subsystem computation. A sequence for computation of individual blocks within a maximal cyclic subsystem is generated. This enables serial computation of the subsystem starting with assumed values for tear variables and known values for the subsystem inlet streams. The user may specify partial or complete computational sequences, either by entering the computational (sub) sequence or by specifying nesting of computational loops. If the user has entered a feedback controller this creates a loop or loops that must be converged independently of any recycle loops in the same cyclic subsystem. In such case ASPEN will establish a sequence that, will converge the control loop(s) either inside or outside of the recycle loop(s) according to user specification. Simultaneous convergence of control and recycle loops is considered in 5.4.2. INPUT DATA Partitioned block diagram with a set of tear variables. OUTPUT DATA Computational sequence REQUIRED SPECIFICATIONS None OPTIONAL SPECIFICATIONS Partial or complete computational sequence 45 - Partial or complete specification of nesting of computational loops. ERROR CONDITIONS Fatal. - Unrealizable specified computational (sub) sequence 5.3.2. STREAM CONVERGENCE DESCRIPTION This capability is used to converge the tear streams of a flowsheet containing recycles. Any number of tear streams may be converged simultaneously and a system may contain any number of stream convergence blocks. The number of convergence blocks in a system depends on the number of maximal cyclic subsystems and the degree of nesting, :.f any, the user chooses for each maximal cyclic subsystem. The ASPEN system default is to set up a convergence block for each cyclic subsystem to converge the tear streams of that subsystem simultaneously. The following methods will be considered for implementation in the ASPEN system. (1) Direct substitution with specified damping or accel- eration (2) Bounded Wegstein v^ith adjustable bounds and direct substitution sub-i deration. (3) Bounded Wegstein on selected streams and direct sub- stitution on the remaining streams. 46 (4) A quasi-Newton method (5) General Dominant Eigenvalue method In addition to convergence promotion, the convergence pro- cedure tests for stream convergence. When all tear streams in the system are converged, an overall balance is obtained for the entire system as well. INLET STREAMS Any number and type including information streams OUTLET STREAMS One-to-one correspondence with inlet streams REQUIRED SPECIFICATIONS The user must specify the convergence method. In addition, in the case of the acceleration methods, the user must specify which stream variables are to be accelerated. For example, in most energy balance problems the variable tear set will most often be component flows and enthalpy. If, as is usually the case, pressure is specified, these variables are sufficient to continue sequential computation of the blocks. For all methods, the user may specify convergence tolerance information. ADDITIONAL RESULTS Acceleration method parameters PPS REQUIREMENTS None OUTLET ATTRIBUTES Same as corresponding inlet 47 5.3.3. CONTROL CAPABILITY DESCRIPTION The control capability allows at least partial solution of the flowsheet design problem in which variables other than system inlet stream attributes and block parameters are specified. For each such design specification an inlet attribute or block parameter must be "freed." Specifically, the ASPEN control subprogram samples the values of selected variables and adjusts the values of freed variables in such a way that user specifications are met. The number of user specifications must be equal to the number of adjusted variables. Any identifiable simulation variable may be sampled or adjusted. The user specifications may be any function of sampled variables. A control procedure could be used, for example, to adjust a feed flow rate so that a specified product rate is achieved. The mathematical problem can be expressed as the solution of N simultaneous nonlinear equations in N unknown variables f £ (Sj^, . . . ,Sj|,X2^, . . . ,x^) = ith specification, i = 1, ..., N where s represents M samples of variables that result from an evaluation of a flowsheet segment, x represents adjusted (unknown) variables, and f represents nonlinear functions specified by the user. 48 It should be emphasized that the following assumptions are implied: (1) the response functions f and thus the sampled vari- ables s are independent functions of the adjusted variables x and only of x. This means that any loops (control or convergence) inside the flow sheet segment under control must be converged each iteration of the control block. (2) the system of equations has one and only one solution in the region of adjusted variables x specified. These assumptions are important as it is very easy to mis- specify a control block. An ASPEN flowsheet may contain any number of control blocks. STREAMS Control block streams are information streams implied by the sampled and adjusted variable identifications. REQUIRED SPECIFICATIONS Sampled variable identifications Adjusted varible identifications and bounds The functions f described above (often, simply the identity function) Convergence method Inside or outside convergence if in same cyclic subsystem as recycles. ADDITIONAL RESULTS None PPS REQUIREMENTS None 49 5.3.4. INFORMATION TRANSFER The information transfer capability of ASPEN allows trans- fer of information between flowsheet elements. For example, block or stream variables may be set before they are used depending on results from other parts of the flowsheet. Care should be taken that the information transfer is not a feedback of information, thereby creating a loop that should be torn and converged. As currently specified, the ASPEN automatic sequencing capability will not detect such a loop. Specifically, the subprogram samples the values of selected variables, computes values of user specified functions of them, and uses these computed values to set other selected vari- ables. Any identifiable simulation variable may be sampled or set and the values of the set variables may be any function of the sampled variables. There is no limit to the number of sampled or set variables in information transfer blocks. As an example it might be desired to specify the vapor flow rate from a fixed pressure flash to be 10 moles/hr. The infor- mation transfer procedure could be used immediately before the flash to: (1) sample feed flow rate (2) Compute the function (vapor fraction) = 10/ (feed flow) (3) Set the vapor fraction of a fixed PV flash to the computed value. 50 STREAMS An information transfer block creates implicit information streams implied by the sampled and set variable identifications REQUIRED SPECIFICATIONS Sampled variables identifications Set variable identifications and bounds The functions to apply to the sampled variables (often simply the identity function) ADDITIONAL RESULTS None PPS REQUIREMENTS None 51 5.4 IMPLEMENTATION OF NEW ALGORITHMS Data structure in the ASPEN system will readily accommodate computational procedures different from the sequential modular algorithm. We shall discuss two algorithms which can be imple- mented with comparatively small additional efforts provided the sequential modular method is already implemented. 5.4.1 SIMULTANEOUS MODULAR ALGORITHM Let us assume that each process unit can be represented by a linear algebraic model. Each block is modeled by Y = AX + b where - Y is a vector of variables of outlet streams, - X is a vector of variables of inlet streams, A is a matrix of coefficients describing the linear relationships, b is a constant vector. The entire flowsheet can then be represented by a set of linear algebraic equations. Solution of the linear system provides variables for every stream which is an outlet of some block. Therefore, a single solution of the linear system provides values for variables of all recycle and information feedback loops. If the linear model truly represents the system behavior, the solution is found. However, in order to generate 52 the linear model one has to assume a solution and generate linear relationships describing each block. Since the initi- ally assumed solution is invariably erroneous, the generated linear model is not correct. Therefore, the solution of the linear model is not the true solution. Solution of the non-linear model can be used to find a better linear model (matrices A). New values of the tear vari- ables can be taken as the initial guess for serial computation in sequential modular simulation and the computational sequence is executed only once. This leads to a new set of values for (Y,X) pairs of each block. A new linear model of the blocks is generated and the entire linear flowsheet model is solved again to obtain new values for the outlet stream variables. The above procedure is repeated until the convergence is achieved. In order to implement simultaneous modules procedure a linear model of each block has to be created. Given a set of input and output block variables, i.e. an exact solution of the block model, the linearization procedure generates coefficients of the linear block model. The simplest linearization procedure employs split frac- tions whenever possible. This results in very sparse, almost diagonal linear block models. More complex block procedures (e.g. distillation columns) are represented by specially derived linear models. The user will have to supply a linearization procedure for his own nonlinear block which he may wish to include in the simulation. 53 To summarize, exact models of each block (these are the modules employed in sequential modular procedure) solve for true values of outlet stream variables for a given set of input stream variables and block parameters. This is a basis for the derivation of linearized block models. The linear flowsheet model provides new values for tear variables, thus enabling the subsequent solution of each nonlinear model (see figure below) . TEAR VARIABLES ? LINEAR NONLINEAR MODEL MODEL PARAI4ETERS OF LINEAR / N BLOCK MODELS Figure 5.1 Interaction between a linear and a nonlinear model in simultaneous modular simulation. 54 This solution procedure is applied separately to each max- imal cyclic subsystem determined by the partitioning procedure. Nevertheless, the number of variables in a linear model can be quite large (several thousands) , but the equations are sparse. To store and repeatedly solve the linear system of the same structure, special sparse matrix techniques have to be used. The remainder of this paragraph describes algorithms required to implement simultaneous modular methods. 5.4.1.1 NON-ZERO PATTERN OF THE LINEAR MODEL DESCRIPTION This procedure determines a nonzero pattern of the linear model and finds such permutation of the system equations and variables that the minimal work space is required for the solution of the algorithm (niinimal fill-in). The algorithm permutes variables and equations according to streams. Stream connectivity information is represented by the stream adjacency matrix where in a position (I, J) there is an entry 1 provided the stream I is an immediate predecessor of stream J. Stream adjacency matrix for each maximal cyclic subsystem is permuted in a bordered lower triangular form with the nonzero pattern as shown on Figure a. mwm^. 55 "\ x\ XX xx\ XX XX x\ XX xxxx\ XX xxxxx\ XX xxxxxx\ XX xxxxxxxv XX xxxxxxxx XX xxxxxxxx XX Figure a. Figure b. Right hand columns correspond to tear streams. Since each unit is incident with a comparatively small number of streams, the matrix will, most of the tine, be of the structure similar to Figure b. The procedure systematically examines each block. The number of variables in each inlet and outlet stream is deter- mined and corresponding locations are reserved for the block matrix in the overal systera matrix. Elements of the block matrix are entered according to the nonzero pattern for a particular block. This also includes design specifications. 56 The user has to provide procedures which will determine non- zero pattern for his own modules. INPUT DATA Adjacency matrix for streams. Minimal set of tear streams. Nonzero pattern for linear model of each block in the maximal cyclic subsystem. OUTPUT DATA Bordered lower triangular nonzero pattern of the linear model of the maximal cyclic subsystem. ADDITIONAL RESULTS Bordered lower triangular form of the stream adjacency matrix REQUIRED SPECIFICATIONS User has to specify nonzero pattern for its own modules OPTIONAL SPECIFICATIONS None ERROR CONDITIONS Fatal - Insufficient storage for the linear system. - Missing nonzero pattern for user supplied module. 57 5.4.1.2 REPETITIVE SOLUTION OF SPARSE LINEAR ALGEBRAIC SYSTEM OF THE SAME STRUCTURE DESCRIPTION Successive iterations between the linear and nonlinear models require repetitive solution of the sparse linear system of the same structure. The identity of the structures can be utilized to increase the speed of the subsequent solutions. During the first solution of the linear system a pivot sequence is chosen. The saise pivot sequence is employed in subsequent solutions. This implicitly assumes that the element appearing subsequently in pivot positions are not zero. INPUT DATA Nonzero pattern of the linear system. Corresponding numerical values. OUTPUT DATA Solution of the linear system REQUIRED SPECIFICATIONS None ERROR CONDITIONS Fatal. Insufficient work space to solve the system of equations. Zaro pivot element. 58 5.4.2. IMPLICIT CONVERGENCE ALGORITHM DESCRIPTION Sequential and simultaneous modular algorithms converge iteration loops of block models within the iteration loop at the flowsheet level. It is possible to converge simultaneously outer iteration loops of some (or all) block models and the flowsheet energy and material balances iteration loops. Each block model is essentially the same as the module for sequential modular simulation, except that the outer iteration loop limit is set to one. Variables and equations of the outer iteration loops within the block models become part of the overall system of equations. From an equation solving viewpoint, implicit convergence is equivalent to partitioning and tearing the system equations as implied by the solution procedures built into the block algor- ithms. Therefore, the method does not offer as many possibil- ities (in particular for design variable selection) as a true equation solving procedure. On the other hand, the advantages of the algorithm are: 1. ability to satisfy a broad spectrum of design specifications simultaneously with the block and flowsheet convergence. 2. most of the existing software for a sequential modular simulation can be employed with minor changes. 59 In order to implement the implicit convergence algorithm, the system establishes locations continuing values of iteration varibles and corresponding functions. If a procedure employs an algorithm utilizing an (approximate) Jacobian matrix, special procedures to estimate the initial Jacobian matrix in compar- atively small execution time are required. 5.4.2.1 IMPLICIT CONVERGENCE BLOCK DESCRIPTION The implicit convergence block receives function values corresponding to errors in the current solution of the system equation. A numerical procedure is employed to find new values of iteration variables. The user can either choose a system library routine, or insert his own algorithm. INPUT DATA Initial guesses for iteration variables. Functional values for each iteration step. OUTPUT DATA Values for iteration variables. REQUIRED SPECIFICATIONS None 60 OPTIONAL SPECIFICATIONS Maximum number of iteration steps. Maximum allowed error at the solution point Numerical solution procedure type. ADDITIONAL RESULTS Approximate Jacobian matrix at the solution point. ERROR CONDITIONS Fatal. Overspecif ied design requirements 61 6. PHYSICAL PROPERTY SUBSYSTEM 62 6 . 1 INTRODUCTION The primary purpose of the ASPEN Physical Property Su)3- system (PPS) is to calculate the physical properties of materials as functions of temperature, pressure and composition when accessed by the Unit Operations or Cost Estimation Sub- systems. The PPS is itself composed of a set of subsystems which consist of a property calculation subsystem/ regression and constant estimation subsystems and data banks. The prop- erty calculation subsystem is the central component of the PPS, while the others are peripheral components that are needed to develop and store the information required to perform the calculations. The property calculation subsystem of ASPEN will contain a wide range of methods for calculating the properties of the vapor, liquid and solid phases of pure compounds atid mixtures as functions of temperature, pressure and phase composition. Methods will be included for homogeneous solid mixtures and for liquid mixtures containing electrolytes. It will be able to calculate the properties of pseudo-compounds representing boiling fractions of petrol-sum and coal liquids, as well as the coal and heterogeneous solid properties that are important in fossil fuel conversion processes. The properties that may be calculated include volumetric, thermodynamic, phase equilibrium and transport properties of homogeneous materials, and other properties related to the physical processing of heterogeneous solids. The routes which 63 determine the specific methods to be used may be completely specified by the user. Different routes may be employed in different parts of a process. The available property models will represent the state of the art in equations of state, corresponding states correla- tions, activity coefficient models, and other types of corre- lations. New models may be readily added to the system by the user. The property calculation subsystem will be directly access- ible by the ASPEN Unit Operations and Cost Estimation Subsys- tems, and by the physical property data regression subsystem. In addition, it may be used independently of other parts of ASPEN through appropriate calls on its subroutines. The data banks provided as part of ASPEN will contain a comprehensive set of physical property constants for the vapor, liquid and solid phases of pure compounds, the important prop- erties of a number of standard types of coal, and the param- eters of certain classes of property correlations for both pure compounds and coal. The PPS will be designed to readily accom- odate additional data banks provided by the user. All of the data banks will be accessible by the simulator executive system and the PPS peripheral subsystems. The data regression subsystem enables the user to supply information in the form of raw data from which property con- stants and/or correlation parameters are determined by linear or nonlinear regression. The user may specify which models are to be employed for this, and which constants and/or parameters 64 are to be treated as regression variables. The results may be placed in an output file or stored directly in one of the data banks. The constant estimation subsystem estimates property con- stants not otherwise available using mainly structural infor- mation. These results may also be routed to a data bank or stored in an output file for later retrieval. A variety of physical property input modes will be pro- vided. The information required for a simulation consists of property constants, correlation parameters and calculation routing information. Any subset of this information may be provided by the user as direct input. The remainder may be obtained from the data banks, or from the constant estimation or data regression subsystem output files. In the next section, the architecture of the PPS and its interaction with other subsystems of ASPEN are discussed. The sections that follow then describe in detail the major compon- ents of the PPS: the data regression subsystem, the data banks, the constant estimation subsystem and, finally, the property calculation subsystem. 6.2 PHYSICAL PROPERTY SYSTEM ARCHITECTURE The major components of the physical property subsystem (PPS) , other components of ASPEN that interact with them, and the flow of physical property information are shown schematic- ally in Figure 6.1. In this section, discussion is limited 64a z o H EH M cu O < SOU S OMt< U H a H o w a< i< D JH O Q « C* P4 (A o s W tn tn u 65 mainly to the functions of the components of ASPEN that are not strictly members of the PPS, but directly influence mudh of the physical property information flow, in Figure 6.1 these are shown as the Simulator Executive System (SES) , and the Unit Operations and Cost Estimation Subsystems (UOS and CES) . Figure 6.1 indicates three functions of the SES with regard to physical properties: receiving input data from various sources, creating a physical property data structure and trans- mitting property calculation routing information to the UOS and CES. In the following discussion each of these functions will be considered in turn. The physical property information required to perform a simulation consists of property constants, correlation param- eters, and calculation routing information. Any subset of this information may be provided by the user as direct input. All other required information raust be available from other sources specified by the user using prescribed key words. The allow- able sources are the data banks and the output files of the Data Regression and Constant Estimation Subsystems. The SES will be capable of directly accessing these sources, thereby precluding the necessity of manual transfer of information. The SES will create a Physical Property Data Structure containing all constants and correlation parameters, such as critical properties and Antoine constants, needed by the Physical Property Calculation Subsystem (PPCS) . In addition, it will contain information defining all property calculation routes to be used in the simulation. Since the data structure 66 contains all the physical property information required for a simulation, it will be stored by the SES in a designated output file. It may be retained for subsequent use for any process based on the same set of components and property calculation models without repeating the file-building procedure. The SES will transmit the property calculation routing information to the UOS and CES, which will in turn directly access the PPCS. They will provide state variable information such as temperature, pressure and composition, and the PPCS will return to them the values of properties such as phase equilibrium ratios, enthalpies and densities. Wh6n calculation routes are not specified by the user standard default routes will be used. The Data Regression and Constant Estimation Subsystems and the Data Banks are' peripheral components of the PPS, providing information to the other components of Figure 6.1, but receiving no information back from them. The functions and attributes of these components, as well as the PPCS, are discussed in detail in the next few sections. 6.3 DATA REGRESSION SUBSYSTEM The purpose of the Data Regression Subsystem (DRS) is to use linear or nonlinear regression to determine the values of physical property constants and/or correlation parameters to achieve a best fit of a model or correlation to a set of raw data. It is anticipated that the most common applications will 67 be to fit single variable polynomials to a set of raw data, to determine Antoine constants from pure compound vapor pressure data, and to determine binary interaction parameters for activity coefficient models from binary temperature-pressure- composition data. The DRS will be designed to handle these particular applications reliably and efficiently, and with minimum effort on the part of the user. Any special purpose features required to achieve this, such as phase equilibrium calculation algorithms, will be included as part of this sub- system. In addition, the DRS will have an extended capability to handle a wide range of unanticipated data processing problems. It will be able to access the same PPCS as the Unit Operations and Cost Estimation Subsystems, including the control moni- tors. For each data point, the user may specify uniquely any calculable quantity in the PPCS as a basis for an objective function. For example, for some data points, the preferred objective function might be a sum of squares of differences between observed and calculated equilibrium ratios, while for others in the same data set it might be in terms of observed and calculated system pressures. The user may also specify any combination of property con- stants and property model correlation parameters as the set of regression variables. Values of all other constants and corre- lation parameters required to calculate the objective function must be provided by the user as direct input, or must be avail- able from one or more data banks or from the output file of the 68 Constant Estimation Subsystem. The source from which each set of required information is to be obtained by the DRS executive will be specified by the user. The DRS results may be placed directly in an output file, which may be accessed later by the Simulator Executive System. The results may also optionally be directed to the User Primary and Auxiliary Data Banks, but not to either of the Master Data Banks. 6.4 DATA BANKS There will be two types of data bank in ASPEN, referred to as "Primary" and "Auxiliary." In the following sections, the functions of these data banks and the types of information stored in them are discussed, 6.4.1 Primary Data Bank (PDB) The PDB will consist of two parts. The first part will contain a comprehensive set of physical property constants for the vapor, liquid and solid phases of pure compounds. The second part will contain the important properties of a number of standard types of coal and coal derived materials. Since these two parts of the PDB contain different kinds of infor- mation, they are discussed separately in the following sub- sections. mm^ 69 There will be a Master PDB which will be provided as part of ASPEN, and will be internally self-consistent and well- documented. In addition, ASPEN will be designed to readily accomodate any number of User PDB's, having the same structure as the Master version, but not necessarily constrained by the same set of ground rules regarding consistency and documenta- tion. Pure Compounds A preliminary list of constants contained in the pure compound part of the PDB is given in Table 6.1. The general ground rules are listed below. Those that refer only to the Master PDB are indicated by an asterisk (*); the others pertain to both the Master and User PDB: (a) All constants have precise, universally accepted definitions, and are generally available for most compounds included, (b)* Pure compound constants are internally self- consistent. For example: ' Pc^c = ZcR^c ' The liquid vapor pressure equation passes through the critical point, triple point and normal boiling point. ' The vaporization enthalpies at a reference temperature and at the normal boiling point satisfy the extended Watson equation. 70 (c) Properties that depend on temperature and/or pressure, such as vapor pressure, are represented by correlations that are known to provide accurate descriptions of the behavior over wide ranges of the independent variable(s). (d) For properties that depend on temperature and/or pressure, such as vapor pressure, the best avail- able constants of the correlation equation are included, along with information indicating the upper and lower limits of the range of applic- ability of the equation with those constants. (e)* All informaton :.s accompanied by a documentation code which refers to a location in a reference file. This file contains sufficient information to enable the information in the data bank to be exactly reproduced, (f)* Information may be entered by direct input only, using prescribed data formats, AS shown in Figure 6.1, the User PDB(s) may receive input directly from the Data Regression and/or Constant Estimation Subsystems. Coal and Coal Derived Materials Each set of information contained in this section of the PDB pertains to a particular coal, usually originating from an individual bed or seam such as Illinois No. 6 or Alabama Black 71 Creek, that may be regarded as having more or less uniform properties* It is recognized that coal is a highly hetero- geneous material, and that no two samples - even from the same seam - have precisely the same properties. The values included in the Master PDB for each coal will be typical, well-docu- mented values. Each of the 13 major classes of coal, from lignite to anthracite, will be represented by one or more individual coals in the Master PDB. The User PDB may have constants for any number of additional coals, as well as different sets of constants for any of the coals included in the Master PDB. A preliminary list of the coal property constants is given in Table 6.2. Ground rules (e) and (f) listed above are applicable to the coal part of the Master PDB as well as to the pure compound part. 6.4.2 Auxiliary Data Bank (ADB) The purpose of the ADB is to store certain frequently used correlation parameters that have generally available (usually published) sets of values that are more or less universally applicable. Examples are BWR constants for a group of com- pounds, and selected binary interaction parameters for the Soave modified Redlich-Kwong equation of state. There will be both a Master ADB provided as part of ASPEN, and any number of User ADB's that may contain different values of the same param- eters as well as additional sets of parameters. 72 Since binary compound interaction parameters for activity coefficient models are usually used only for the system for which they were developed, they are not "universal" in the same sense as the Soave binary interaction parameters, for example, and hence are not included in the ADB. They will, of course, be stored in the file containing the Physical Property Data Structure, as discussed in section 6.2., after initially being provided as direct input, and may therefore be directly retrieved for subsequent runs. A preliminary list of correlation parameters included in the Master ADB is given in Table 6.3. While values may be entered into the User ADB{s) by direct input or directly from the Data Regression Subsystem, the Master version is restricted to direct input according to prescribed data formats. 6.5 PHYSICAL PROPERTY CONSTANT ESTIMATION SUBSYSTEM The purpose of this subsystem is to estimate property con- stants not available in the data banks, from the literature, or elsewhere, using mainly structural information. It will have a comprehensive set of methods for pure compounds, selected methods for the pseudo - compounds that represent boiling frac- tions of petroleum and coal liquids, and methods for estimating some of the data bank constants for coal (see Table 6.5). It will also have the capability to estimate binary pair infinite dilution activity coefficients. 73 A preliminary list of estimation methods for pure compounds is given in Table 6.4, After critically evaluating the methods listed, some may in fact not be included, and others may be added. The criteria on which the selection of methods will be based will include accuracy and the ranges of applicability with respect to conditions (temperature and pressure) and com- pound types. Particular attention will be paid to compound types that are common in coal conversion processes, such as aromatics and high molecular weight polycyclic compounds. Methods that provide redundant capabilities will be eliminated. It may be noted that methods for surface tension and binary diffusion coefficient are not included in Table 6.4. Predic- tive methods for these properties exist and selected ones will be included after further evaluation. While most of these methods depend mainly on structural information, which must be provided as direct input, many methods also require other constants, such as the critical temperature and pressure and normal boiling point. This infor- mation may be provided as direct input, obtained from the Primary Data Bank, or estimated where possible by other methods in the subsystem. When a sequence of methods is required, and alternate routes are possible, the route used is prescribed by the user. Three predictive methods will be included for estimating infinite dilution activity coefficients for binary pairs: Pierotti-Deal-Derr, ASOG and UNIFAC. All three of these methods depend on structural information, which must be pro- 74 vided as direct input. The predicted values may be used to determine the binary interaction parameters of any specified activity coefficient model using the Data Regression Subsystem. Several methods will be provided for estimating the pseudo- compound properties of hydrocarbon fractions that are defined by the average specific gravities and boiling points of ASTM or arbitrary distillation cuts. The Cavett and Kesler-Lee corre- lations will be included for petroleum fractions, and other state-of-the-art correlations will be available for fractions that arise from a Similar treatment of coal liquids. Results of all estimations will be stored in a file that may be directly accessed by the Data Regression Subsystem or the simulator executive system. The results may also be routed directly to the User Primary Data Bank, but not to the Master Primary Data Bank. 6 . 6 PHYSICAL PROPERTY CALCULATION SUBSYSTEM The purpose of the Physical Property Calculation Subsystem (PPCS) is to calculate physical properties for pure compounds and mixtures of gases, vapors, liquids, homogeneous solids and heterogeneous solids (including coal) as functions of tempera- ture, pressure and composition, using standard thermodynamic relationships and correlation models. As shown in Figure 6.1, these calculated properties are available to the Unit Opera- tions, Cost Estimation and Data Regression Subsystems. 75 The PPCS may be thought of as the central component of the PPS, since it contains the physical property models on which all property calculations are based. There will be a Master PPCS that will contain both the set of property models to be provided as part of ASPEN, and the monitors which control the calculation of properties according to the user specified routes. Modification of the Master PPCS is expected to be infrequent, and will require familiarity with the details of the system architecture and operation. The User PPCS, on the other hand, will contain only property models supplied by the user. New models may be added to this subsystem (and existing ones modified) with relative ease, requiring no modification of the control monitors. In the following subsections, the structure of the Master PPCS is described in detail. In addition, lists of properties calculated by it are given, along with preliminary lists of the methods to be included. 6.6.1 Gases, Vapors, Liquids and Homogeneous Solids A list of general calculation requirements is given in Table 6.6, where it may be noted that only a few of the prop- erties are directly required for process calculations. The remaining ones are largely subordinate properties, combinations of which are used to calculate those that are needed for process calculations. 76 The various types of calculated properties fall logically into one of the following three classes: (1) Principal derived (2) Subordinate derived (3) Basic The relationships between these three classes of properties and the flow of information between them are shown schematically in Figure 6.2, The principal derived properties (PDP) are: (1) Phase equilibrium ratios of components in mixtures. (2) Pure compound and mixture enthalpies, free energies, entropies, heat capacity ratios and sonic velocities. All of the PDP are required directly for process calculations, and all may be expressed in terms of other properties by definition or by standard thermodynamic relationships. The relatively simple expressions used to calculate them are given in Table 6.7. Calculation is controlled by the PDP monitor (PDPM) , which implements the various options listed in Table 6.7 according to user specifications, and obtains the values of other required properties from other appropriate monitors. The subordinate derived properties (SDP) are combinations of basic properties, and in some cases other SDP, that are required for calculation of the PDP, but are not themselves needed for process calculations. They are expressed entirely in terms of other properties and state variables (temperature, pressure and composition) using standard, rigorous thermo- dynamic relationships. The SDP and the fundamental expressions by which they are calculated are given in Table 6.8. 77 The SDP monitor (SDPM) implements the options listed in Table 6,8 according to user specifications, and obtains the values of the required basic properties from the basic property monitor. In most cases, the SDPM also performs the arithmetic and mathematical operations indicated by the expressions in Table 6.8. Exceptions to this arise when a user specifies that one of the derived properties, fugacity coefficient, enthalpy departure or free energy departure, is to be obtained from a specified equation of state. In this case, it is most efficient to perform the indicated integrations analytically, resulting in unique expressions for each equation of state model. It is preferable to include these expressions in the respective equation of state subroutines, where the derived properties are calculated as options. The SDPM then accesses the appropriate equation of state model directly. In order to accomodate this strategy, every equation of state module in the PPCS will have the capability to calculate any combination of the properties ln(}). Ah, Ag, 0Ah/9T)p, (3Ag/9T)p, P, T, and V. The properties calculated on any given access depend on the properties required and options specified by the user. Integrations of molar volume with respect to pressure are required for the various Poynting corrections indicated in Table 6.8 by the symbol "6", as well as the enthalpy and free energy corrections denoted by the symbol "A". When the molar volume model specified by the user is independent of pressure, the SDPM will evaluate the integral by a simple multiplication of the molar volume times the appropriate pressure difference. 78 If the molar volume model includes a pressure effect, the SDPM will integrate numerically over the appropriate pressure range using a Gaussian quadrature formula. The number of points to be employed may be specified by the user, but will not be allowed to exceed some small number such as 4 or 5, with a default of 1 or 2. The basic properties (BP) are properties that are calcu- lated directly from models specified by the user. They depend directly on temperature, pressure, phase composition, pure compound constants and correlation parameters, and in some cases other basic propertied:. They are all described by models, the forms of which are not necessarily prescribed by thermodynamics. Calculation of the BP is controlled by a monitor, the main task of which is to properly implement the model choices specified by the user. The BP and model types available for each are listed in Table 6.7. Note that the fogacity coefficients, enthalpy departures and free energy departures were included in Table 6.8 as SDP. They are also listed here as BP because they may be obtained both from equation of state models as derived properties, and from other types of models as direct correla- tions. Enthalpies and equilibrium ratios have a similar dual availability as either PDP or BP. This is a particularly important feature, since both of these properties are sometimes represented by highly empirical models having specialized forms. In the case of equilibrium ratios, information will be 79 available to the model subroutines to enable them to identify specific components that may be treated as "key" components, or in some other special manner. The standard reference state for the thermodynamic functions enthalpy, free energy and entropy is an important consideration. For reasons that are discussed under Unit Operations Subsystem (Section 7.2.5), it is desirable to have a dual basis, with element- based functions for reactor calcula- tions, and compound-based functions for non-reaction separation calculations. Since the thermodynamic functions are normally calculated as the sum of an ideal gas value and a departure function (see sections 2, 3 and 4 of Table 6.7) , a dual system may be achieved by simply allowing the bracketted terms in section 5 of Table 6.8 to be optional. These terms represent the standard enthalpy and entropy of formation, respectively, of the pure compound ideal gas at T^.^^ and 1 atm based on zero enthalpies of the constituent elements in their standard (stable) states at the same conditions. (The value of T ^ is taken to be 298.15«K, in accordance with the usual convention.) If the bracketted terms are not included, then all thermodynamic functions will be based on pure compound ideal gas enthalpies, entropies and free energies equal to zero at 298.15*>K. A preliminary list of candidate BP models is given in Table 6.10. In each area a critical evaluation will be conducted to determine which methods will be included in the Master PPCS. In some areas, it is anticipated that some of the methods 80 listed will not be included, notably equations of state and transport properties. However in other areas, such as activity coefficients of electrolytes and solids, methods not presently listed will be added. Assessment of particular methods will be based on ranges of applicability with regard to pressure, temperature, polarity and other factors. For equations of state, the criteria will also include suitability for the calculation of fugacity coefficients, enthalpy departures, free energy departures, and liquid properties as well as vapor. With the exception of the derived quantities obtained from equations of state, all calculations of a general nature, using definitional or standard thermodynamic relationships are performed by the appropriate control monitors, rather than in the method modules. The repetitive coding of common calcula- tions is thereby eliminated, and the requirements associated with adding new method modules to the system are reduced. It is notable that the ??CS is based on the direct calculation of enthalpy and free energy, rather than enthalpy and entropy. The rationale for this choice is explained in Appendix 6.1. The use of the unsymmetric convention to represent liquid fugacities for vapor-liquid systems having one or more super- critical components is an important feature of the PPCS. Coal conversion processes in particular frequently involve systems in which supercritical gases such as H2, N2, CO, CO2 and CH^ coexist in phase equilibrium with coal liquids or other 81 subcritical solvents. When the pressure is sufficiently high to cause gas solubilities greater than a few mole per cent, commonly used corresponding states correlations, such as the Chao-Seader correlation, do not provide accurate estimates of the reference liquid fugacities of the gas components. In such cases it is usually possible to obtain more accurate results using the unsymmetric convention with the Henry's constant as the reference liquid fugacity. The modified unsymmetric convention employed in the PPCS is discussed in detail in Appendix 6.2. 6»6.2 Coal and Coal-Derived Materials A preliminary list of calculated physical properties of coal and coal-derived materials is given in Table 6.11. These properties are all functions of temperature and presssure. Together with the constants included in the coal data bank (Table 6.2), they provide a wide range of information useful for both the physical processing of coal, such as crushing, grinding, drying and f luidization, and chemical processing such as oxidation and hydrogenation. The proper handling of the enthalpy effects that accompany reactions involving coal, coal liquids, and other materials characterized by something other than composition in terms of pure compounds, is an important requirement of the PPS. The materials that may be involved in a reactor enthalpy balance calculation, as reactants and/or products, may be classified according to the following three types: 82 (A) Pure compounds or pseudo™ compounds, and mixtures thereof. The pseudo-compounds arising from petroleum and coal liquids would be included in this type. (B) Gases, liquids or solids that may or may not participate in reactions, and are characterized by elemental composition, such as the "ultimate analysis" of coal. (C) Inerts that do not undergo phase change or participate m reactions. In some cases, the ash constituents of coal would be an example of this type. In general, in order to perform a reactor enthalpy balance, the unit operations Subsysteir. uses the enthalpies of the inlet and outlet streams, evaluated at their respective conditions of temperature and pressure, relative to the constituent elements in their standard states*. For materials of type A, element- based enthalpies may be calculated using enthalpy departure functions and standard ideal gas heats of formation, as discussed in section 6.6.1. For type C materials, the ability to calculate sensible heat changes is all that is required. Since the ideal gas state may be meaningless for materials of type P, the PPS will provide an alternate, element-based method that will be compatible with the enthalpy departure methods for type A. This method will be based on the assumption that the elemental analyses will be known for type B reactant materials. * The stable states at 298.15''K and 1 atm, 83 and that the reactor model will calculate the elemental analyses for type B product materials. The information required to calculate the enthalpies of these materials, whether reactants or products, consists of their respective heats of formation, in their stable states at standard conditions, from the constituent elements at standard conditions. These heat of formation values may be obtained from the usually available heats of combustion of the type B materials. 84 6 . 7 NOMENCLATURE P AC. V f g Ag Agf h Ah Ahf H constant pressure heat capacity departure of constant pressure heat capacity from ideal gas value at same temperature constant volume heat capacity coefficient of molecular diffusion fugacity Gibbs free energy per mole departure of Gibbs free energy per mole from ideal gas value at same temperature standard Gibbs free energy of formation per mole* enthalpy per mole departure of enthalpy per mole from ideal gas value at same temperature standard enthalpy of formation per mole* Henry's constant *For compound as ideal gas at 298. 15° K and 1 atm based on the constituent elements in their standard (stable) states at 298.15''K and 1 atm. mm 85 k K N n P P P R s As AS, ref T T u V "V w X y z z ref s thermal conductivity phase equilibrium ratio molecular weight number of moles vapor pressure system pressure (absolute) ideal gas reference pressure (1 atm) universal gas constant entropy per mole departure of entropy per mole from ideal gas value at same temperature standard entropy of formation per mole* temperature ideal gas reference temperature (298. IS** K) sonic velocity molar volume partial molar volume weighting factor mole fraction in liquid phase mole fraction in vapor phase mole fraction in solid phase compressibility factor *For compound as ideal gas at 298.15°K and 1 atm based on the constituent elements in their standard (stable) states at and 1 atm. 298.15**K 86 Greek Letters 3 Y 6 A e/k n e p a a 0) coefficient of extended Watson equation exponent activity coefficient solubility parameter enthalpy or free energy pressure correction function Lennard-Jones parameter viscosity fugacity pressure correction function fugacity coefficient (f/P) molar density Lennard-Jones parameter surface tension Pitzer's acentric factor aipole moment Stiel polar factor Subscripts 87 1 ' k m b c r TP BP component index supercritical component index subcritical component index normal boiling point critical property reduced by corresponding critical property triple point bubble point vector over components Superscripts o (o) IG V L C V/L v/c L/L L/C E VAt> FUS SUB 1, 1 1 pure compound infinite dilution ideal gas gas or vapor phase liquid phase crystal or solid phase coexisting vapor and liquid phases coexisting vapor and solid phases coexisting liquid phases coexisting liquid and solid phases excess property vaporization property fusion property sublimation property coexisting liquid phases pertains to unsymmetr ic liquid reference fugacity convention 88 FIGURE 6.2 FLOW OF CALCULATED PHYSICAL PROPERTY INFORMATION BP Method Modules - SDP Combinations of BP & other SDP — ^ PDP Combinations of SDP & BP ' ^ BP = basic properties SDP = subordinate derived properties PDP = principal derived properties 89 TABLE 6.1 MASTER PRIMARY DATA BANK PURE COMPOUND CONSTANTS Compound name Molecular formula Molecular weight Critical temperature Critical pressure Critical volume Critical compressibility factor Critical viscosity Critical thermal conductivity Normal boiling point Normal freezing point Triple point pressure Triple point temperature Solid vapor pressure at reference temperature Extended Antoine equ. liquid vapor pressure coefficients Inp oL = A + B T + C + D In T + ET + FT^ Antoine equ. solid vapor pressure coefficients Inp oC = A + B T + C Pitzer's acentric factor 0) = -log^Q(p' oL Tr=.7/Pc'-1 Solubility parameter Liquid molar volume at normal boiling point Saturated liquid molar volume at reference temperature Vapor molar volume at normal boiling point 90 TABLE 6^1 (continued) Saturated vapor molar volume at reference temperature Coefficient and exponent of modified Rackett equ. for saturated liquid molar volume .oL In Ad-Tj,) n Solid density at normal freezing point Solid density at reference temperature Ideal gas heat capacity temperature polynomial coefficients (and related constants) Enthalpy of vaporization at normal boiling point Enthalpy of vaporization at reference temperature Extended Watson equation coefficient Ah VAP = Ah VAP \1-T'/ .38 + B(l-T^) Enthalpy of fusion at normal freezing point Enthalpy of fusion at reference temperature Enthalpy of fusion at triple point Enthalpy of sublimation at triple point Liquid heat capacity at normal boiling point Saturated liquid heat capacity at reference temperature Solid heat capacity at normal freezing point Solid heat capacity at reference temperature Standard enthalpy of formation of ideal gas @ aS^C Standard entropy of formation of ideal gas @ 25°C Standard Gibbs free energy of formation of ideal gas @ 25®C 91 TABLE 6.1 (continued) Ionic valence Dielectric constant Dipole moment Lennard-Jones parameters (a and e/k) Liquid viscosity at reference temperature Coefficients of modified Andrade liquid viscosity equation In n' A + I + C ^w T Liquid surface tension at reference temperature Liquid surface tension equation temperature coefficient and exponent a^ = A(l-T^)'' Liquid thermal conductivity at reference temperature Liquid thermal conductivity equation temperature coefficients k^ = A[l + B(l-T^)^/"^] Solid thermal conductivity at reference temperature Gas self-diffusion coefficient at 1 atm and at reference temperature ::;-<%::>:c:^:; TABLE 6.2 MASTER PRIMARY DATA BANK COAL CONSTANTS 92 Coal name ASTM rank based on equilibrium moisture content (e.g., bituminous, anthracite) Proximate analysis, wt. per cent Ultimate analysis, wt. per cent Petrographic or maceral analysis, vol. per cent Mineral species analysis, wt. per cent (e.g., kaolite, dolomite) Major compounds in ash, wt. per cent of ash (e.g., SiOj, CaO) Sulfur content, wt. per cent of coal (organic, pyritic, sulfate) Analysis of trace elements in coal, wt. per cent Analysis of trace elements in ash, wt. per cent Reflectance of coal macerais Equilibrium moisture content Free swelling index Bulk density at reference temperature Coefficients of bulk density temperature dependence relationship True density at reference temperature Coefficients of true density temperature dependence relationship Microhardness 93 TABLE 6.2 (continued) Hardgrove grindability, crushing indices Bond's work index Washability analysis Gray-King coke type and yield Gieseler plasticity analysis Moisture diffusivity Friability Shape factor (grain microstructure) Flammability and explosive limits Specific gravity of ash Ash fusibility (initial deformation point, fusion point, flow point) Enthalpy of fusion of ash Slag viscosity temperature dependence coefficients Slag surface tension temperature dependence coefficients Coefficients of heat capacity temperature dependence relationships for coal, ash, liquid slag, solid slag Coefficients of thermal conductivity temperature dependence relationships for coal, ash, liquid slag, solid slag Enthalpy of drying at reference temperature Enthalpy of flash oxidation at reference temperature and pressure 94 TABLE 6.2 (continued) Ethalpy of complete oxidation at reference temperature and pressure Coefficients of correlation relating enthalpy of limited oxidation (pretreatment) to proximate and/or maceral analysis Enthalpy of flash hydrogenation at reference temperature and pressure Enthalpy of complete hydrogenation at reference temperature and pressure Coefficients of correlation relating enthalpy of limited hydrogenation to proximate and/or maceral analysis Coefficients of correlation relating volatile matter to temperature at reference rate of heating Temperature of onset of devolatilization (based on 5 per cent of original volatile matter) Enthalpy of flash devolatilization at reference temperature and pressure Enthalpy of complete devolatilization at reference temperature and pressure Coefficients of correlation relating enthalpy of limited devolatilization to proximate and/or maceral analysis Carbonization (Fischer) assay from pyrolysis at reference temperature and controlled heating rate; resulting gas heating value, gas yield, light oil yield and tar yield 95 TABLE 6.3 MASTE R AUXILIARY DATA BANK UNIVERSAL MODEL AND CORRELATION PAPAMETERS Original BWR coefficients for pure compounds Chao-Seader pure compound liquid fugacity correlation coefficients Grayson-Streed pure compound liquid fugacity correlation coefficients Robinson-Chao pure compound liquid fugacity correlation coefficients Binary interaction parameters for the Soave modified Redlich-Kwong equation of state Solubility prarmeters for the Scatchard-Hildebrand activity coefficient model van der Waals molecular area and volume parameters for the UNIQUAC activity coefficient model Parameters of the Yen-Mexander enthalpy departure correlations Parameters of the Yen-Woods molar volume correlations TABLE 6.4 ESTIMATED PURE COMPOUND PHYSICAL PROPERTY CONSTA^ITS AND ESTIMATION METHODS 96 CONSTANT METHOD REQUIRED INFORMATION* Lydersen Nokay mod. of Lydersen T, / sp.gr. @ 60°F Lydersen Forman-Thodos Lydersen Riedel Vetere M, SI SI SI ^c' ^C ^b SI Definition Garcia-Barcena T , P , V c' c c SI 0) Definition Edmister Lee-Kesler ^c' ^C ^b ■^C ^c' "^b Stiel-Thodos Ogata-Tsuchida SI SI U. Minkin SI o , e/k Tee-Gotoh-Stewart Stiel-Thodos ^c' ^c' ^C ^c TABLE 6.4 (CONTINUED) 97 CONSTANT P°^(T) METHOD Reidel Har la Cher- Brown Gomez-Nieto-Thodos Smith-Winnick-Abrams- Prausnitz REQUIRED INFORMATION* Correlation Parameters fc' ^c' 'b' " Tj^(or B.P. § lOmmHg) , SI ah°^^{T^,) Ah°V*^(T) Vetere Watson (extended) Pitzer Halm-Stiel •^c' ^C ^b Ah^VA^T^^f), T^, 6 X 6 Definition Viswanath-Kuloor P°^(T ^), P , T , 0) *^ ^ ref c c ^j^oVAP ^^^j ^ ^^ Cp'^(T) Benson Benson'O'Neal SI SI C°^T) Ah' OIG Lyman- Danner Benson Thinh-Trong Anderson-Beyer-Watson •^c' "^b' C^^''^'^)' SI SI SI SI ASfIG Benson Thinh-Trong SI SI CONSTANT Ag olG em 98 TABLE 6.4 (CONTINUED) METHOD REQUIRED INFORMATION vanKrevelen-Chermin Standard relation SI As^, Ah° Definition and Haggenmacher T , p , Ah^^^^ d°^ v°^ Lyckman-Eckert- -Praunitz c c 1°^(T) Reichenberg M, T^, SI n°''(T) Orrick-Erbar M, V°^, SI Thomas T , vO^, SI Morris T^, SI vanVelzen-Cordozo-Langenkamp SI oV Rcy-Thodos T^, P^, M, SI oL Sato Robbins-Kingrea M T^, VOL, c^^ Ah°^^^ SI = structural information 99 TABLE 6.5 ESTIMATED PROPERTY CONSTAN TS OF COAL AND COAL-DERIVED MATERIALS Flammability and explosive limits Enthalpy of flash oxidation at reference temperatute and pressure Enthalpy of complete oxidation at reference temperature and pressure Enthalpy of flash hydrogenation at reference temperature and pressure Enthalpy of complete hydrogenation at reference temperature and pressure Enthalpy of flash devolatisation at reference temperature and pressure Enthalpy of complete devolatilization at reference temperature and pressure The following properties, using mean maximum reflectance: Free swelling index Initial volatile matter content Coke yield Gas heating value at reference temperature Gas yield at reference temperature Tar and light oil yield 100 TABLE 6.6 PHYSICAL PROPERTY CALCULATION REQUIREMENTS FOR GASES (OR VAPORS) , LIQUIDS AND HOMOGENEOUS SOLIDS PROPERTY APPLICATIONS ^^^ Fugacity coefficient Fugacity coefficient temperature derivative Fugacity Fugacity temperature derivative Enthalpy < 2) Enthalpy temperature derivative ^^^ Enthalpy departure function Excess enthalpy Entropy ^^^ Entropy temperature derivative ^^^ Entropy departure function Excess entropy Free energy (2) Free energy temperature derivative Free energy departure function Excess free energy Pressure^-^^ Volume Temperature ^^^ Saturation pressure ^^^ (2) V°, v., L' V^, v., L' L°, C° L°, C° V^, V, L° V^, V, L° V^, V, L° L, C V^, V, L° V^, V, L° V^, V, L° L, C V^, V, L° V°, V, L° V°, V, L° L, C V^, V, L° V^, V, L° V^, V, L° V°, V, L° , L. L, C", C L, C°, C L, C°, C L, C*^, C L, C°, C L, C°, C L, C*^, C L, C°, C L, C°, C L, C", C L L, C°, C TABLE 6.6 (continued) 101 Saturation volume '^^ Saturation temperature^"' Equilibrium ratio^^^ Activity coefficient Activity coefficient temperature derivative Partial molar volume Viscosity ^^^ Surface tension^^' Thermal conductivity'^^ Diffusion coefficient^^' Heat capacity ratio^^^ Sonic velocity ^^^ v°. V, L°, L, C°, C v°. V, L°, L, c°. C ViAi, Vi/Ci, WH' Li/Ci H- Ci Li, Ci V, L°, L L°, L v°, V, L°, L, C°, C v°, Vi , L°, L i v°, V v°. V (1) SYMBOL DEFINITIONS: V^ pure compound gas or vapor ^i component of gas or vapor mixture gas or vapor mixture pure compound liquid ^1 component of liquid mixture L CO liquid mixture pure compound hoiT.ogeneous solid Ci component of homogeneous solid mixture homogeneous solid mixture 102 TABLE 6.6 (continued) Note: Unless otherwise noted, pure compound properties are assumed to be functions of T and P; component of mixture and total mixture properties are assumed to be functions of T, P, and X, ^ or z. (2) Directly required for performing process calculations, (3) Function of T, v and Xf Z ^^ 1- (4) Function of P, v and x, x ^^ ^• (5) Function of T and x, ^ or z. (6) Function of P and x, ^ °^ ?l 103 TABLE 6.7 PRINCIPAL DERIVED PROPERTIES FOR GASES (OR VAPORS) , LIQUIDS AND HOMOGENEOUS SOLIDS * 6.7.1 EQUILIBRIUM RATIOS OF COMPONENTS IN MIXTURES VAPOR/LIQUID (SYMMETRIC CONVENTION) V/L K L ^ OL ' 1 1 Independent options: L (a) = 1 (b) :>' 6.7.3 FREE ENERGY OF GAS, VAPOR; LIQUID OR SOLID Pure Compound. 107 oIG g (T,P) = g"""(T) + Ag^(T,P) Option: Ag = Mixture IG g(T,P,0 = g (T,0 + Ag(T,P,£) (i = ^' Z o^ £^ Option: Ag = 6.7.4 ENTROPY OF GAS, VAPOR, LIQUID OR SOLID Pure Compound s°(T,P) = s°^^(T) + As^(T,P) Option: As = Mixture S(T,P,C) = S-'''(T,e) + Ls{T,P,a (i = 21' Z o^ z) Option: As = 108 6.7.5 HEAT CAPACITY RATIO OF GAS OR VAPOR Pure Compound Cf" + Lcl" P P (cf - cf) Mixture _E E. P P ^e>arvx. 113 2^ w = summation over subcritical components only m Option: 6. = 1 im Note: The coefficients w are weighting factors that may be area, volume or mole fractions, normalized to unity over the subcritical components. Restriction: The option chosen for the w's must be the serme as that used for the activity coefficients by the unsymmetric convention. 6.8.5 IDEAL GAS HEAT CAPACITY, SNTHALPY, FREE ENERGY AND ENTROPY PURE COMPOUND . OIG,_, h (T) „oIG,_, s (T) ■I, C°'^ dT T P ^ref ■/ C°^^ d £nT ref +Ah£°-'-^ (Tref) ■¥Ls^^^ (Tref) g°^^(T) = h°^G(T) -Ts°^G(T) 114 MIXTURE Cp^(T,i) ZqC°^^(T) h^^(T,5) zqh. 1 s^^{T,a .11 ^ 1 ^ g^^(T,i) IG = h^^(T,U - Ts^-(T,i) (^ = X, ^ or z) 6.8.6- HEAT CAPACITY DEPARTURE FUNCTION PURE COMPOUND GAS, VAPOR OR LIQUID AC ■(^), GAS, VAPOR OR LIQUID MIXTURE AC. /9Ah\ (i = X or ^) 6.8.7 HEAT CAPACITY DIFFERENCE PURE COMPOUND GAS, VAPOR OR LIQUID* 115 C°-C° P V TM° ^9P/9T)pO ~^ (9P/8P°). GAS, VAPOR OR LIQUID MIXTURE* C - C P V TM ^^^/^T)p.g 2 OP/9p) T,C (£ = X or ^) 6.8.8 ENTHALPY DEPARTURE FUNCTION PURE COMPOUND GAS, VAPOR OR LIQUID USING EQUATION OF STATE Pressure-explicit equation of state: Ah' ■/:['-'(i?}J dv + RT(z - 1) Required partial derivatives obtained by finite difference estimates using specified equation of state or other specific model. 116 Volume- explicit equation of state; »>° ■ /'['-'(»)> GAS, VAPOR OR LIQUID MIXTURE USING EQUATION OF STATE Pressure-explicit equation of state: Ah =rt-K4.ih^'^''^-" Volume-explicit equation of state Ah = ■rhKsvj (£ = X or y) PURE COMPOUND GAS OR VAPOR USING FUGACITY COEFFICIENT Ah°^ = -RT' K'-l GAS OR VAPOR MIXTURE USING FUGACITY COEFFICIENTS Ah^ = -RT' W^My^ 117 PURE COMPOUND LIQUID USING FUGACITY Ah°^ = -RT- \ 3^ A PURE COMPOUND LIQUID USING VAPORIZATION ENTHALPY Ah°^(T,P) .oV,„ _oL oVAP = Ah-^(T,p--) - Ah°^^(T) + A°(T,P) Independent options: (a) Ah°^ = (b) A" = o . , oL (c) A° = Ah°^(T,?) - Ah°^(T,p°^) (d) A° = -RT^ (^[^X°^v°^dp]j LIQUID MIXTURE USING VAPORIZATION AND EXCESS ENTHALPIES WITH SYMMETRIC CONVENTION Ah^{T,P,x) = Ex,Ahf(T,pf) i -^ OVAP ' ^x. ZLh. -^-(T) + Ex. A° + h^ (T,P,x) Independent options: 118 (a) Ah°^ = (b) L? = OL oL, (C) A? = Ah°^(T,P) - Ah^;^(T,p^") (d) AT = - RT' \^9T[RTjpOL 1 J^j (e) h' = Restriction: No component should be supercritical at mixture conditions LIQUID MIXTURE USING VAPORIZATION AND EXCESS ENTHALPIES WITH UNSYMMETRIC CONVENTION Ah^(T,P,x) = [SUBCRITICAL CONTRIBUTIONS] • , E' + Ah + h where (SUBCRITICAL CONTRIBUTIONS] refers to the terms appearing in the symmetric convention expression applied to the designated subcritical components only, and Ah = - RT^ E X, 119 = - RT*- E X, {^), /X k summation over designated supercritical components only Independent options: (a) Same as listed for symmetric convention for subcritical components (b) h^' = LIQUID MIXTURE USING REFERENCE FUGACITY DERIVATIVE AND EXCESS ENTHALPY Ah (T,P,x) = 2 /^lnf?\ L '^'^ ^^i( 3T ) + h^ (T,P,x) Option: h = PURE COMPOUND SOLID USING ENTHALPY OF SUBLIMATION At system temperature, Tj .oC Ah°^(T,P) = Ah°^(T,p°C) _ Ah^SUB^^j ^ ^o A° = -Rt2 \3T L^T J C-'°°"L •]), 120 Independent options: (a) Ah°^ = (b) A° = At triple point temperature, T^pt Ah' oC oV oSUB (T,P) = h°^°(T^p)-h°^°(T)+Ah°^T,p,P^p)-Ah"="''(T^p) TP •/m + I C°^ dT + A° L p A° = -RT^ W-C-"""]) Independent options: (a) Ah°^ = (b) A° = SOLID MIXTURE USING SUBLIMATION AND EXCESS ENTHALPIES At system temperature, T: C ._ . - ^ .^oV. oC, oSUB Ah^(T,z) = Ez. Ah°^T,p^") - Zz. Ah---"(T) + E z. A? + h^ (T,z) i ^ ^ A? = -RT^ 1 1 3T IrT J oC 1 J / MjnfmfiiiUiiM'iMii'iii 121 Independent options: (a) Ah°^ = (b) A^ = (c) h^^ = At triple point temperatures, T, TP. 1 Ah (T,P,z) = IZihf°(T^p ) - hI°(T,z)+Ez^Ahf (T^p ,P ) ^ 1 i i i i + 2 z. A° + h^ (T,z) i TP. -^^] Independent options: (a) Ah°^ = (b) A^ = (c) h^ = 122 fi.fi.g EXCESS ENTHALPY LIQUID, SYMMETRIC CONVENTION E^ = - RT I i\ ST /P,x LIQUID, UNSYMMETRIC CONVENTION h^' = - RT^ I X, (-^U Note; E = summation over supercritical components only SOLID = - RT ^il-— jp.z 6.8.10 FREE ENERGY DEPARTURE FUNCTION PURE COMPOUND GAS, VAPOR OR LIQUID USING EQUATION OF STATE Pressure-explicit equation of state; Ag' = /■" (p - M) dv - RT In z° + RT {z° - 1) + RT tn P/P^ef ■lit ■ nmfiririiMWiiiiir riiil tm VnriiVMMMyi- 123 Volume-explicit equation of state; Ag° = ■/;- RT ^) dP + RT In P/P ^ GAS, VAPOR OR LIQUID MIXTURE USING EQUATION OF STATE Pressure-explicit equation of state: Ag RT (P - ^) dv - RT £n 2 + RT (Z - 1) + RT tvi P/P ^ Volume-explicit equation of state Ag -/',v- Jo RT ^) dP + RT In P/P ^ P ' ref PURE COMPOUND GAS OR VAPOR USING FUGACITY COEFFICIENT Ag°^ = RT £n°^/3T)p EOS, CSC 4>°^(T,P), 0<(.°V3T)p EOS, CSC «|)i(T,P,^), 0*I/3T)p^^ EOS, CSC (J>^{T,P,x), 0*^/3T)p^^ EOS oV OAh°^/3T)p EOS, CSC, OSM Ah°^(T,P), OAh°V3T)p EOS, CSC, OSM Ah^(T,P,i) , OAh^/3T)p^^ EOS, CSC, OSM Ah^(T,P,x Ag°^(T,P) Ag°^(T,P) Ag''(T,P,^ , OAhV3T) P,x (9Ag°^/3T)p OAg°V3T) , (3AgV3T) P'Z Ag^(T,P,x), OAgV3T)p K^/^(T,P,x,:^) CV/L) K^/^(T,P,x,z) (V/C) K^/^(T,P,x^,x^^) (L/L) EOS EOS EOS OSM OSM OSM K^/^(T,P,x,z) (L/C) OSM 130 TABLE 6.9 (CONTINUED) BASIC PROPERTY h°V(T,P) h°^(T,P) h'^d.P) h'^(T,P,2.) h^(T,P,x) h*^(T,P,z) ih°V*^T) ih°*'"S(T) Ah°S"B(T) C°^°(T) C°=(T) P°''(T)* p°St)* «iMP„)<'' MODEL TYPES ** OSM OSM OSM OSM OSM OSM OSM OSM OSM OSM OSM OSM OSM OSM TABLE 6.9 (CONTINUED) 131 BASIC PROPERTY Yi(p )(T,x). (3Yi(P^)/^^^P,x Y^(p^){T,z), OYl(p^)/3T)p^, n°^(T,P) n°^(T,P) n^(T,P,i;) n''(T,P,x) k°^(T,P) k°^{T,P) k°^(T) lc^(T,P,^) k''(T,P,x) k^(T,z) t?Y. (T,P) 11 ^ij(T,P) pI^(t,p,z) MODEL TYPES AC AC CSC, om CSC, OSM OSM OSM CSC, OSM CSC, OSM OSM OSM OSM OSM OSM OSM OSM ** TABLE 6.9 (CONTINUED) 132 BASIC PROPERTY MODEL TYPES** "ij c^' OSM OSM a°^(T,P) OSM a (T,P,x) OSM * Temperature and pressure inverse properties also optionally available, e.g., T(v°^,P) and P(v°^,T) ** EOS = equation of state model CSC = corresponding states correlation AC = activity coefficient model OSM = other specific model .v^aiuttrJaaainrjinrbibOiiv. 133 TABLE 6.10 BASIC CALCULATED PHYSICAL PROPERTY MODELS FOR GASES (OR VAPORS) , LIQUIDS AND HOMOGENEOUS SOLIDS EQUATIONS OF STATE Original Redlich-Kwong Modified Redlich-Kwong: Soave Wilson Barnes-King Chueh-Prausnitz Original BWR Modified BWR: Starling Lee-Kesler Hayden-O'Connell virial Vetere virial Halm-Stiel virial Tsnopolous virial Barner-Adler Lee-Erbar-Edmister Sugie-Liu Peng-Robinson 134 TABLE 6.10 (CONTINUED) ACTIVITY COEFFICIENTS Nonelectrolytes Original Scatchard-Eildebrand Modified Scatchard-Kildebrand: Flory-Huggins athermal Binary interaction parameter Margules Van Laar Null modified Van Laar Wilson NRTL FLOWTRAN modified NRTL UNIQUAC Null-Chen Electrolytes Guggenheim-Turgeon (strong) Edwards-Prausnitz (weak) ''A-A-' 135 TABLE 6.10 (CONTINUED) ENTHALPY DEPARTURE (OTHER THAN EQUATION OF STATE) Yen-Alexander Edmlster-Persyn-Erbar Chueh-Deal hydrocarbon mixtures containing Hj) Huang-Daubert (petroleum fractions) Cavett (liquid) PURE LIQUID FUGACITY COEFFICIENT (OTHER THAN EQUATION OF STATE) Chao-Seader Grayson-Streed Robinson-Chao Prausnitz-Shair MOLAR VOLUME (OTHER THAN EQUATION OF STATE) Vapor Edward s-Thodos (saturated, nonpolar) 136 TABLE 6.10 (COI^TINUED) Liquid Goldhaminer (pure) Modified Rackett (pure saturated) Cavett Spencer-Danner modified Rackett (saturated) Gunn-Yamada (saturated) Yen-Woods Chueh-Prausnitz Chueh-Deal (hydrocarbon mixtures containing H2) PARTIAL MOLAR LIQUID VOLUME Infinite Dilution Brelvi-O'Connell Lyckman-Eckert-Prausnitz Tiepel-Gubbins VAPOR PRESSURE Liquid Extended Antoine Cavett Gonez-Ni eto-Thodos TABLE 6.10 (CONTINUED) Solid Antoine BINARY HENRY'S CONSTANT Polynomial in reciprocal temperature IDEAL GAS HEAT CAPACITY Polynomial in temperature ENTHALPY OF VAPORIZATION Watson Chueh-Deal Chueh-Swanson ENTHALPY OF FUSION Constant 137 ENTHALPY OF SUBLIMATION Constant 138 TABLE 6.10 (CONTINUED) ENTHALPY (OTHER THAN FROM DEPARTURE FUNCTION) Polynomial in temperature for pure compound Mole fraction weighted average of component polynomials in temperature for mixture VISCOSITY Low Pressure Pure Gas Chapman-Enskog (Lennard-Jones, nonpolar) Chapman-En skog (Stockmayer, polar) Thodos Reichenberg Chapman-Cowling Hirshfelder-Bird-Spotz Low Pressure G a s Mixture Wilke Hern ing- Zipper er Brokaw Thodos Dean-Stiel 139 TABLE 6.10 (CONTINUED) High Pressure Pure Gas Chapman-Cowling Stiel-Thodos Reichenberg Uyehara-Watson High Pressure Gas Mixture Dean-Stiel Giddings Low Temperature Pure Liquid Orrick-Erbar Thomas Morris van Velzen-Cardozo-Langenkamp High Temperature Pure Liquid Letsou-Stiel Liquid Mixture McAllister 140 TABLE 6.10 (COHTINUED) THERMAL CONDUCTIVITY Pure Gas Kinetic theory (monatomic) Eucken Modified Eucken Mason-Monchick Bromley Misic-Thodos Roy-Thodos Low Pressure Gas Mixture Lindsay-Bromley Mason-Saxena Stiel-Thodos Wassiljewa Brokaw Pure Liquid Robbins-Kingrea Sato-Riedel Missenard ZAi^. 141 TABLE 6.10 (CONTINUED) Liquid Mixture Fillipov Li Jordan-Coates NEL Vredeveld DIFFUSION COEFFICIENT Low Pressure Binary Gas Chapman-Enskog (Leonard- Jones, nonpolar) Chapman-Enskog (Stockmayer , polar) Wilke-Lee Fuller-Schettler-Giddings Mathur-Thodos Gilliland Slattery-Bird Othmer-Chen Brokaw Multi component Gas Mixture Wilke Toor Hi r schf elder-Curt iss-Bird 142 TABLE 6. 10 (CONTINUED) Liquid Stokes-Einstein Wilke-Chang Scheibel Reddy-Dora iswamy Hayduck-Laudie Louis-Ratclif f King-Hsueh-Mao Perkins-Geankoplis Electrolytes Nernst-Haskell (single electrolyte) Gordon Vinograd-McBain SURFACE TENSION Non-Aqueous MacLeod-Sugden Brock-Bird-Miller Aqueous Me issner -Michaels Tamura-Kurata-Odani 143 TABLE 6.11 CALCULATED PHYSICAL PROPERTIES OF COAL AND COAL-DERIVED MATERIALS Enthalpy of coal, ash, liquid slag, solid slag as function of temperature Thermal conductivity of coal, ash, liquid slag, solid slag as function of temperature True density as function of temperature Bulk density as function of temperature Agglomerating tendency as function of temperature and free swelling index Equilibrium volatile matter content as function of temperature and pressure Enthalpy of flash oxidation as function of temperature and pressure Enthalpy of limited oxidation as function of temperature, pressure, and extent Enthalpy of complete oxidation as function of temperature and pressure Enthalpy of flash hydrogenation as function of temperature and pressure Enthalpy of limited hydrogenation as function of temperature, pressure and extent Enthalpy of complete hydrogenation as function of temperature and pressure Enthalpy of flash devolatillzatlon as function of temperature and pressure 144 TABLE 6.11 (CONTINUED) Enthalpy of limited devolatilization as function of temperature, pressure and extent Enthalpy of complete devolatilization as function of temperature and pressure Viscosity of slag as function of temperature Surface tension of slag as function of temperature Conversion of maceral analysis to wt. per cent Conversion of proximate or ultimate analysis to the following bases: (a) specified moisture content (b) dry (c) dry ash-free (d) dry mineral matter - free (Parr) '>?':;% ■*^'*»'^™'™?««i»B«!gV'; 145 APPENDIX 6.1 ENTHALPY-FREE ENERGY BASIS OF PROPERTY CALCULATION SUBSYSTEM The Property Calculation subsystem as represented in Tables 6.7, 6.8 and 6.9 is designed around the direct calculation of enthalpy and free energy, rather thah enthalpy and entropy, with entropy being obtained from simple relationships. In principle, it makes no difference whether entropy or free energy is the second quantity directly calculated, along with enthalpy, if it is assumed that any or all of these three quantities and their temperature derivatives may be required, the following table shows the appropriate relationships for the two alternative cases: Quantities Directly Case Calculated Relationships to Obtain Other Quantities hr(3h/3T)p,s,Os/3T)p h,Oh/3T)p,g,(3g/aT)p g = h - Ts (3g/3T)p = -s s = - (3g/3T)p (3s/3T)p = (1/T) (3h/3T) In Case 1, the two quantities h and s must be calculated in order to obtain g. Furthermore, (3h/3T)p and (9s/3T)p are in effect redundant, since they are related by (3h/3T)p ^ T(3s/3T)p. In Case 2, on the other hand, only one quantity, (3g/9T)p, must be calculated if s is desired. While Og/3T)p and Oh/8T)p are related, the relationship is not simple, hence these are not redundant quantities. 146 APPENDIX 6.2 MODIFIED UNSYMMETRIC FUGACITY CONVENTION The Property Calculation subsystem as represented in Tables 6.7, 6.8 and 6.9 uses a modified unsymmetric convention for handling vapor-liquid systems having one or more supercritical components. The expression for the equilibrium ratio is given in section 1 of Table 6.7, and Table 6.8 contains expressions for the activity coefficient (6.8.3), Henry's constant (6.8.4), liquid enthalpy departure (6.8.8), excess enthalpy (6.8.9), free energy departure (6.8.10) and excess free energy (6.8.11). These relationships are based on an unsymmetric convention that is slightly different from the usual one, in which the fugacity of a supercritical component in a mixture is expressed as: where lim y^ ^- 1 Xj^-O and Hj^ is the Henry's constant of the component in the mixture v-fv''. 147 It may be shown that y^ and H,^ are given by InyJ = InY,^ - InY,^^^^ (o) . z (oX InH, = inyr^ ^w^ln(H,^/y-^) where the index m refers to subcritical components only, the w's are arbitrary weighting factors restricted only by the condition ^km is ^^^ binary Henry's constant of supercritical component k in subcritical component m, and (o) Yk = lim Y,^ m m ^kia = lim Y^ X -». 1 m Two undesirable features of this usual way of writing the unsymmetric convention are that: (1) The quantity Y,^ is required twice, yet it cancels out in the product y^Hu, and so has no effect on fj^. (2) The Henry's constant H,^ depends on the activity coefficient model. 148 By as: the simple expediency of defining a modified Henry's constant InH' = Ilw^lnH K mm ^km and a modified unsymmetric activity coefficient as lnY^'= InT^ "^w^lnT^J • — « ,.* then we have Yj^H,^ = ^ ^.K,^ but are not encumbered by either of the undesirable features just mentioned. 149 7. UNIT CPERATIONS SUBSYSTEM 150 7.1 INTRODUCTION The Executive System functional specifications describe how a process flowsheet is used to develop an ASPEN flowsheet con- sisting of a network of blocks and streams. The blocks of this network are related to process unit operations. The primary purpose of the ASPEN Unit Operations Subsystem (UOS) is to provide a library of mathematical models of the unit operations common to fossil fuel conversion processes. In the context of the overall ASPEN simulation the UOS is used by the Executive to solve the flowsheet heat and material balance problem, generating stream information and unit infor- mation throughout the system. The executive controls the UOS model execution by use of a computational structure developed from the user's description of the flowsheet, the user's speci- fications, and the computational architecture of the system. The information generated by the UOS may be used by the user directly or may be used by the Cost Estimation and Economic Evaluation Subsystem. During its calculations the UOS inter- faces directly with the Physical Property Subsystem. Specifically an ASPEN unit operations model is a mathe- matical formulation of the heat and material balance, rate, and equilibrium relations describing an element of a process flow- sheet. This element is similar to a unit operation in that it may be an actual piece of equipment, part of a piece of equip- ment, or aggregates of equipment. For simplicity any of these flowsheet elements will be called a unit. A unit operations 151 model is used to calculate unit outlet stream conditions given unit inlet stream conditions and a set of user specified unit parameters. Unit information other than outlet stream condi- tions, such as heat duty, is also calculated in most cases. The ASPEN UOS provides models for the unit operations commonly used in fossil fuel conversion processes. This includes the models normally found in existing process simu- lators, such as distillation column models, as well as models for solids handling, coal conversion, etc. that are unique to ASPEN. Since some operations are not standardized, or even developed, to the extent necessary to provide universally applicable built-in models, the UOS also provides for the interfacing of user supplied models. 152 7.2 PRINCIPL3S OF THE UPS Any system that is easy to use, maintain, modify, etc. must have a set of underlying principles common to all segments of the system. This section proposes general principles for the ASPEN UOS. 7.2.1 LEVEL OF DETAIL There are two questions related to the level of detail in a process simulation. The first is "How rigorous should the unit operations models be?" The second is "What energy related stream variables should be computed by the unit operations models?" ASPEN will allow the user flexibility in both areas, and at the individual unit level rather than the flowsheet level. An ASPEN simulation can be tailored to do only the computations necesary to meet the user's needs and converge the flowsheet. With regard to model riqor, process engineers need to be able to model unit operations at different levels of detail and rigor. Furthermore, they need this flexibility independently for each individual unit. UOS models may be categorized as follows. (1) INPUT-OUTPUT MODELS In these models no atteupt is made to model performance from fundamental principles. The user merely provides param- eters for aimplG, unually linear, Input-output relatJonn. An 153 example would be a reactor model in which the stoichiometry and percent conversion of a key component is specified for each reaction. Input-output models are basically material balance rela- tions with enthalpy balance options. For example a split fraction separator can also compute the enthalpy and tempera- ture of outlet streams and a block enthalpy balance when the user specifies that the outlet streams are saturated vapor or saturated liquid at a given pressure. (2) SHORTCUT MODELS These models apply fundamental principles with rather severe simply ing assumptions. An example would be the Fenske- Underwood-Gilliland approach to distillation column models. An enthalpy balance is optional for some short-cut models, such as distillation, and a fundamental part of others, such as heat exchangers. (3) RIGOROUS MODELS These models apply fundamental principles directly to the unit operation. An example would be a plug flow reactor model solved by numerical integra-iion of the describing differential equations. Rigorous models almost always include enthalpy balances, although it can ba an option in some isothermal cases. 154 Why are three levels of detail needed? Presumably a rigorous model will do everything that a short-cut model or input-output model will do. Furthermore, the documentation will be easier for a user to follow if there is only a descrip- tion of one model rather than a description of three. The answer is that rigorous models often require much more data than an engineer might have available. Also rigorous models are generally much more costly to compute than simplified models. Finally, rigorous models are sometimes less reliable than short-cut models since trade-offs of reliability versus accuracy are made in formulating simplified models For enthalpy balance problems, the additional stream vari- ables of interest are temperature, pressure, enthalpy, and phase fraction. For a multicomponent system only two of these quantities are needed to completely specify the system. pressure is almost always a specified variable in process simu- lation so it is almost always available without computation. AS a result it is frequently true that flowsheet convergence can be obtained for the enthalpy balance case (as compared to the mass balance case) with only the consideration of the single extra stream variable of temperature or enthalpy since pressures are usually fixed throughout the system. Of these two variables temperature might seem the most natural choice. However it is inlet stream enthalpy, not temperature, that appears explicitly in an enthalpy balance. For example, a distillation column or adiaoatic flash needs inlet stream 155 enthalpy, but not temperature. In addition, outlet stream enthalpy of an adder is computed directly by addition while temperature must be found by an iterative procedure. As a result of the preceeding considerations UOS models will be based on the following system. ASPEN will have a global enthalpy balance switch with a default of "on". When the switch is "on" all block enthalpy balance options are "on" and all possible energy related stream variables are calculated every pass of the block. Therefore, as a default the user need not be concerned with enthalpy balance options. If the global energy balance switch is turned off by the user, energy related stream attributes are calculated: (a) when necessary as part of the solution of a rigorous model. For example a rigorous distillation model will compute all vapor-iiquid stream variables (flow, temp- erature, pressure, enthalpy, vapor fraction, and com- position) as part of the solution. (b) when no equation solving is involved. For example a stream adder can compute outlet stream enthalpy by simple addition. (c) when the enthalpy balance switch for a block is turned on by the user. A procedure is provided to "fill in the blanks", that is, to compute missing stream attributes from known ones. Logic- ally this procedure is executed after flowsheet convergence is obtained, thereby providing the user with desired information 156 out not wasting time doing needless computations during flow- sheet convergence. The other use of this procedure is stream initialization. It can be used to compute enthalpy of a feed or guessed recycle when the user specifies, for example, temperature and pressure. With the above system ASPEN may be used to solve the material balance only problem, the full enthalpy and material balance problem, or problems of mixed type as specified by the user. 7.2.2. MODEL CONVERGENCE CONSIDERATIONS ASPEN algorithms will be as robust as possible. However, as in the case whenever nonlinear equations are involved, con- vergence problems may arise. The following principles relate to model convergence. (1) Procedures contain built-in initialization strat- egies. For example a distillation column model should be able to converge from a minimum of user supplied initialization data, say top and bottom temperature. (2) The user is able to supply initialization information when necessary. In the case of a column the user is able to supply full or partial profiles of tempera- ture, composition and phase flow rate. (3) Convergence parameters are adjustable by the user when necessary. These include damping parameters, step size limitations, unidimensional search on-off switches, maximum number of iterations, etc. 157 (4) All usable block information is saved in the ASPEN data structure for subsequent execution of the block, either in the same run or in a later restart run. (5) Blocks may be reinitialized when the saved information is not valid, say because of algorithm blowup. (6) Diagnostics, at several levels of detail, are avail- able at the user's request. These diagnostics are not considered to be part of the formal ASPEN output and are not saved in the data structure. They are intended only to help solve convergence problems. (7) Convergence tolerances are defaulted by ASPEN but may be overridden by the user. (8) The documentation of an individual model will include a discussion of convergence. 7.2.3. OTHER PRINCIPLES The following is a list of principles that apply generally to the unit operations subsystem. (1) All models, in their basic form, calculate outlet streams from inlet streams and specified parameters. (2) In many cases the user is provided with a choice of specifications for a unit model. These choices may generally be categorized as simulation options or design options. Specifications not provided for may be satisfied by using the control capability. 158 (3) Models generally are restricted to a specific set of stream types. (4) Models not directly related to solids may include solids in the material and energy balance. For example solids are included in the material and enthalpy relations of the vapor-liquid flash algorithm, but not in the phase equilibrium relations. (5) Models provide parameter defaulting whenever poss- ible. The defaults may be constants loaded by the executive or they may be functions of other quanti- ties. For example block pressure may be defaulted -o inlet stream pressure. (6) Models include limits, related to feasible operating conditions or model limitations, for parameters and variables whenever the establishment of such limits is justified. The violation of a limit will result in a warning message either from the Executive (input parameters) or the UOS model itself (calculated variables) . (7) There are no arbitrary limitations, such as number of inlet streams to a block, number of sidedraws from a column, etc. (8) PPS calculational methods can be specified for individual blocks or groups of blocks. In some units multiple PPS methods may be specified, for example different methods may be specified for the shell side and the tube side of a shell and tube heat exchanger. 159 (9) Models are as independent as possible of the methods used by the Physical Property Subsystem to calculate properties. (10) The internal dimensional unit system of the ASPEN UOS is the Basic SI system as defined by NBS Special Publication 330. (11) The documentation for a UOS model will include detailed information on the particular assumptions, methods, solution techniques, etc. employed so that the user can understand exactly what the model is doing. The computer source code for a model will be modularized, structured and commented so that it is easy to understand and modify. 7.2.4. PHASE and CHEMICAL EQUILIBRIUM A phase equilibrium calculation in ASPEN is the determin- ation of system temperature, pressure, enthalpy, number of phases, phase fraction, and phase composition from overall system composition and the specification of any two of temper- ature, presssure, enthalpy, and a given phase fraction. In fact there may result only a single unsaturated phase in which case the term phase equilibrium is somewhat misleading. Also there is the special case of a pure component at its boiling point for which the specification of temperature and pressure is not sufficient. 160 There are several types of phase equilibrium calculations in ASPEN. UOS models that require phase equilibrium calculations will either be specific to a particular type or allow the user to specify the type. The types are (1) Vapor-Liquid (2) Vapor-Liquid-Liquid (3) Liquid-Liquid (4) Vapor-Solid (5) Liquid-Solid (6) Vapor-Liquid-Solid (7) General (vapor -multiple liquid-multiple solid) Types {l)-(3) are conventional process simulation capabilities except that solid phases may be present. These solid phases are assumed not to be in equilibrium. They affect only the system enthalpy balance and undergo only temperature change. In types (4) -(6) the solid phase is assumed to be in equilibrium with the vapor and/or liquid phases. In type (7) multiple liquid and solid phases are considered and the solid phases are assumed to be in equilibrium. The number of phases to consider may or may not be specified in advance. The designer of a process simulator is faced with two questions concerning phase equilibrium calculations. (1) How often should they be done? In other words, is it desirable to completely describe every stream every 161 pass of a simulation? This issue has been discussed in the LEVEL OF DETAIL section. The conclusion was no. (2) Should the user be able to state the phase condition of the system even though the user may be wrong? For example, the simulation of a heat exchanger is much simpler if it can be assumed that no phase change occurs. ASPEN allows the user, in many cases, to specify the phase condition of the system. Phase equilibrium calculations are used in ASPEN both as unit operation models and as a thermodynamic operation available to any function of the simulator. Applications include: (1) stream mixing (2) heat exchanger calculations (4) estimation of physical property parameters (5) solids precipitation (6) bubble point and dew point determination (7) complete thermodynamic state determination of a stream (8) determination of enthalpy from temperature and pressure (9) simulating pressure drop (10) simulating condensers, decanters, etc. (11) simulating flash tanks A natural extension of the phase equilibrium problem posed above is to include chemical equilibrium. The problem state- ment is identical except that overall system atomic composi- tion, rather than molecular composition, is specified. The 162 ASPEN system includes two types of chemical equilibrium calculation. Again they are either specific to particular models or the user may specify type. The two types are: (1) Specified reactions at equilibrium. The type applies when one or more reactions are in equilibrium while the rest of the system is not. Examples are electrolyte systems found in sulfur removal operations and the water-gas shift reaction used to adjust the E2/CO ratio of gasifier products. (2) Entire system is in chemical equilibrium This type can be used to model many combustion and conversion operations. Both types can be applied to single phase or simultaneous phase equilibrium cases 7.2.5. ENTHALPY REFERENCE COMPOUNDS The choice of enthalpy reference state significantly affects UOS models. The reference state most suited to the modeling of fossil fuel conversion processes is based on zero enthalpies for the elemental species (C(S), H2, N2/ 02^ etc.) on which standard heat of formation data is based. Supporting considerations are given below. The choice of reference temperature and pressure does not significantly affect UOS models. The comments apply also to entropy and free energy. 163 (1) The enthalpy balance for any unit, including reactors and coal conversion units is simply "out ^ "in '•' Q where Q is the rate of heat addition to the unit. This simple balance applies to the entire flowsheet as well. In other words all heat effects relevant to chemical process simulation are included in the enthalpy definition. For example there is no need for an explicit heat of reaction term in a reactor enthalpy balance. (2) Enthalpies based on elemental species may be used directly by conversion models not based on explicit reactions. Examples include chemical equilibrium by free energy minimization and specified or correlated yield reactor models. Both of these examples apply to coal conversion modeling. (3) The enthalpy of a material not characterized by chemical compounds, such as coal or petroleum fractions, may be determined simply and directly fropi its ultimate analysis, heat of combustion, and heat capability. Coal and its pure compound conversion procucts have the same zero enthalpy reference and the enthalpy of both is easily obtained from avialable data. (4) Experience has shown that the magnitudes of a pure compound heat of formation ran sometimes overwhelm sensible and latent heat effects and cause some calculations, such as distillation column models, to fail. This problem could be handled from a model convergence point of view for units in 164 which no reactions occur by simply subtracting the ideal gas heat of formation contribution to the feed enthalpy on entry to the model and adding it back into the product enthalpy on exit. Then the internal model calculations would be based on zero enthalpy for the compounds as ideal gases at reference temperature. 165 7.3. UNIT OPERATIONS MODELING CAPABILITIES This section abstracts specific unit operations modeling capabilities as envisioned at the present time. There is no intention of a one-to-one correspondence between the listed capabilities and eventual ASPEN models. The intent is to describe what the ASPEN UOS will be able to do and to indicate what information is required to do it. The present version of this document does not contain abstracts for all of the unit operations that ASPEN will be able to model but the abstracts included here should be considered as complete (not necessarily final) functional specifications. Section 7.5 contains a more thorough list but by unit operation name only. In the unit abstracts that follow, it is assumed that all streams contain constituent and flow information. If any other quantities, such as enthalpy, are required to be known in inlet streams or are calcutlated for outlet streams it is so indi- cated. The section, ADDITIONAL RESULTS lists computed quanti- ties that are not placed in outlet streams but are placed in the block list, for example the temperature profile of a distillation column or plug flow reactor. Table 7.1 gives a prospective list of unit operations models for incorporation in ASPEN. 166 TABLE 7.1 UPS MODEL LIST This table is a prospective list of unit operations models for meeting the requirements of fossil fuel conversion process simulation. TYPE DESCRIPTION STREAM ADDITION STREAM SPLITTING PHASE EQUILIBRIUM SEPARATIONS Single phase streams only V,L or S Vapor-liquid streams Multiple phases, no mass transfer between phases V-L-S streams with mass transfer Vapor-liquid streams Vapor-liquid-solid streams Two-phase flash, vapor-liquid system. Any two of T,P,V or Q specified Liquid-liquid system Three phase vapor-liquid-liquid system Vapor-solid Liquid-solid Vapor-liquid-solid General multiphase including solids Separation factor: specify fraction of each component going to m-1 streams. Two phases in each stream. Specify P,T. or % vaporized for each stream Shortcut distillation - rating case - sizing case Rigorous distillation/absorption/stripping wide boiling feed Narrow boiling and other Packed bed absorber/stripper TYPE DESCRIPTION 167 SEPARATIONS (Cont. ) HEAT EXCHANGERS PRESSURE change: REACTORS Rigorous distillation when two liquid phases are present Liquid-liquid extraction Separation with reaction, particularly electrolyte systems Heat requirement (two phase, three phase) Direct contact heat exchanger (quench) Design Shortcut heat exchanger - no phase change - phase changes (3, no back mixing) - phase change with complete backmixing Thermosyphon reboiler Performance Shortcut heat exchanger - no phase change - one phase change, - one phase change, two liquid phases - phase change with complete backmixing Cold box Slurry Heater Valve Centrifugal pump Positive displacement pump Compressor-centrifugal Positive displacement - rating option using polytropic efficiency Multistage compressor Turbine expander Energy availability Extent, multiple reactions Yield - conversion reactor Equilibrium reactor, single phase TYPE DESCRIPTION 168 REACTORS (Cont.) SOLID HANDLING Multiple phase - equilibrium - rate Continuous stirred tank reactor - design case - rating case General combustor Plug flow reactor - dssign - rating Coal gasification reactors coal liquifaction reactors Coal devolatilizer Coal combustor Radial flow reactor Moving bed reactor Cyclone separator Crusher Grinder Lift tubes Slurry pump Filter/screen/classifier Centrifuge Drier Crystallizer Electrostatic precipitator Hydroclone Kiln drier Venturi scrubber TYPE DESCRIPTION 169 r,Ohll> HANDLING (Cont.) Lock hopper Settler Flotation separator Thickener 170 7.3.1. STREAM ADDITION AND SPLITTING 7.3.1.1. STRSAM ADDER DESCRIPTION Streams are mixed and placed in a single outlet stream. INLET STREAKS Any number and process type OUTLET STREAMS One, most complex inlet type REQUIRED SPECIFICATIONS None ADDITIONAL RESULTS None PPS REQUIREMENTS None OUTLET ATTRIBUTES If all inlet pressures are known, the outlet pressure is the minimum of the inlet pressures. 171 If all Inlot enthalpies are known, the outlet enthalpy is the sum of the inlet enthalpies. In the case of a single inlet stream, all attributes are passed directly to the outlet stream. If it is desired to determine outlet stream temperature and phase fraction the user must specify the phase condition of the system. That is, all vapor, vapor-liquid-solid with or without solid equilibrium, etc. The PPS requirements depend on the specified phase condition. 7.3.1.2 FLOW SPLITTER DESCRIPTION Streams are mixed and the mixture split into any number of outlet streams each having the same attributes, except flow, as the mixture. No component separation is in- volved. Several bases are allowed for the specification of the split. INLET STREAMS Any number and process type OUTLET STREAMS Any number, most complex inlet type 172 REQUIRED SPECIFICATIONS The allowed specifications for the splits are: 1. Mass flow 2. Mole flow 3. Flow fraction 4. Mass flow of e component or group 5. Mole flow of £ comoonent or group In the case where there is a single stream to be split and the stream is single phese with temperature and pressure known the following specification is also allowed. 6. Volumetric flew In any case, for N outlet streams N-1 specifications must be given. Mixed specifications are allowed. ADDITIONAL RESULTS none PPS REQUIREMENTS Specific volume (for volumetric flow specification only) OUTLET ATTRIBUTES See 7.3.1.1 173 7.3.2. SINGLE STAGE PHASE AND CHEMICAL EQUILIBRIUM 7.3.2.1. TWO PHASE EQUILIBRIUM FLASH DESCRIPTION This flash capability solves the single-stage heat balance, material balance, and phase equilibrium relations for given inlet streams and user specifications. It applies to many process operations. This capability applies to both the vapor-liquid and liquid-liquid cases. However, the text is written in terms of the vapor-liquid case; in the liquid-liquid case merely replace "vapor" by "1st liquid" and "liquid" by "2nd liquid". INLET STREAMS Any number of vapor-liquid. See required specifications concerning inlet stream enthalpy. OUTLET STREAMS One vapor and one liquid stream or one vapor-liquid stream. REQUIRED SPECIFICATIONS There are five flash options: 1. Temperature and pressure specified 2. Pressure and vapor fraction specified 3. Temperature and vapor fraction specified 174 4. Pressure and heat added specified 5. Temperature and heat added specified 6. Specified vapor flow rate In all cases vapor mole fraction and enthalpy, liquid mole fraction and enthalpy, and the unspecified values of temperature pressure or vapor fraction are calculated. In cases 4 and 5 inlet stream enthalpy must be known. In cases 1, 2, and 3 heat duty will be calculated if inlet stream enthalpy is known. Under options 1, 4, and 5 it is possible for the result to be a subcooled liquid or a superheated vapor in which case the vapor fraction is set equal to zero or one as appropriate and the non-ex istant phase composition has no meaning. ADDITIONAL RESULTS Heat duty as discussed above PPS REQUIREMENTS K-Values and phase enthalpies OUTLET ATTRIBUTES T, P, H, V 175 7.3.2.2. MULTIPHASE EQUILIBRIUM FLASH DESCRIPTION This extended flash capability solves the single-stage heat balance, material balance, and phase equilibrium relations for either the specified or unspecified vapor-liquid-solid multiphase problem. An example of a specified problem would be a vapor-liquid-liquid flash where the user knows that two immiscible liquid phases are possible. In the unspecified problem the user does not know a priori how many phases exist. The unspecified phase capability is most likely to apply to systems containing solids where there may be many solid phases present in equilibrium. The specified phase problem is, of course, much easier to solve. INLET STREAMS Any number of vapor-liquid or general. Enthalpies must be known for the specified heat addition option. OUTLET STREAMS One general stream or, in the specified phase case, one stream for each phase. REQUIRED SPECIFICATIONS In addition to the phasa specifications discussed above, the user may specify either (1) temperature and pressure, or (2) pressure and heat added. Other specifications may be achieved by using control procedures. 176 In both cases the compositions, conditions and flow rates of the resulting equilibrium phases are calculated. Under option (2) inlet stream enthalpy must be known and under option (1) heat duty will be calculated if inlet stream enthalpy is known. Under either option a single unsaturated phase may result. For the vapor-liquid-liquid specified pase case only the user may choose between the equilibrium ratio method or the free energy minimization method. ADDITIONAL RESULTS Heat duty as discussed above PPS REQUIREMENTS Phase free energies and enthalpies. OUTLET ATTRIBUTES T, P, Hr phase fractions 7.3.2.3. SIMULTANEOUS PHASE AND CHEMICAL EQUILIBRIUM DESCRIPTION This capability is very similiar to 7.3.2.2 except that chemical equilibrium is assumed to exist in each phase as well as phase equilibrium between phases. As in 7.3.2.2 either the specified phase or unspecified phase vapor-liquid-solid multi- phase problem may be considered. Single phase and vapor-liquid systems are very important special cases. For example the 177 generation of synthesis gas by the partial oxidation of fuel may be modeled as vapor phase chemical equilibrium. INLET STREAMS Any number of vapor-liquid or general. Enthalpies must be known for the specified heat addition option. OUTLET STREAMS One general stream or, in the specified phase case, one stream for each phase. REQUIRED SPECIFICATIONS The free energy minimization method is used so no specifi- cations of reaction stoichiometry or equilibrium constants is required. The required specifications are similar to 7.3,2.2 except that the equilibrium ratio method is not available. In addition the user must specify a set of possible system components to be considered by the pro- cedure. ADDITIONAL RESULTS same as 5.3.2.2 PPS REQUIREMENTS Phase free energies and enthalpies plus atomic constituency of reactant and product components. OUTLET ATTRIBUTES T, P, H phase fractions 178 7.3.3. SEPARATION 7.3.3.1. SPECIFIED COMPONENT SEPARATION DESCRIPTION Streams are mixed and the resulting mixture split into any number of outlet streams each having a different composi- tion and flow. Several bases are allowed for the speci- fication of the split. This capability may be used as an input-output model for virtually any type of separation process. INLET STREAMS Any number and process type OUTLET STRET^S Any number, specified process type for each stream. REQUIRED SPECIFICATIONS The allowed specifications for a component or group to a stream are: 1. Mole fraction 2. Mass fraction 3. Fraction of total inlet component flow recovered (split fraction) 4. Mole flow 5. Mass flow 179 Mixed specifications are allowed. The number of degrees o£ freedom is: {# of streams - 1) x (# of components or groups) . If a set of specifications is not feasible this condition will be flagged. In the case of vapor-liquid outlet streams the user may specify: (1) that the stream is a saturated liquid at a given pressure (2) that the stream is a saturated vapor at a given pressure (3) the temperature and pressure of the stream Under the enthalpy balance option the outlet streams are flashed at these conditions to determine the remaining attributes and overall duty. Options (1) and (2) could apply to distillation simulation and option (3) could apply to a decanter. If the stream contains solids only option (3) is allowed. ADDITIONAL RESULTS Heat duty if energy balance option is requested. 130 PPS REQUIREMENTS None unless energy balance is requested, in which case the requirements are those of the phase equilibrium calculation being performed. OUTLET ATTRIBUTES Depends on specification 7.3.3.2 SHORTCUT DISTILLATION The shortcut distillation capability is designed to model columns with a minimum of numerical computation and calls to the PPS. On the other hand the models are not rigorous and only apply to specific conventional column configur- ations. Shortcut models are particularly useful during the early stages of a simulation project. DESCRIPTION This capability is based on the Fenske or Winn - Underwood - Gilliland approach. The key assumptions are constant relative volatility, constant molal overflow, and optimum feed tray location. The condenser may be specified as either total or partial. The desired column performance is specified as splits for light and heavy key components. The Gilliland correlation is used to compute the actual reflux ratio for a given 181 number of theoretical trays, or the number of theoretical trays for a given reflux ratio. The reflux ratio may be specified as an actual value or as the factor reflux ratio/minimum reflux ratio. The relative volatilities may be specified by the user or may be calculated at a user specified temperature and pressure using the feed composition. Similarly the feed quality may be specified or calculated. INLET STREAMS One vapor-liquid OUTLET STREAMS One overhead, one bottoms REQUIRED SPECIFICATIONS (1) Condenser type (2) Light and heavy key splits and identification (3) Number of trays or reflux ratio or reflux ratio/ minimum reflux ratio (4) Relative volatilities or temperature to calculate them (5) Peed quality (may be calculated if inlet enthalpy is known) (6) Column pressure 182 As an option, the condenser and reboiler temperature and duty may be calculated from an overall column energy balance if the inlet enthalpy is known. ADDITIONAL RESULTS Minimum reflux ratio Minimum number of stages Table of actual theoretical stages vs. reflux ratio or reflux ratio/minimum reflux ratio Relative volatilities if calculated PPS REQUIREMENTS K-values if relative volatilities are calculated K-values and phase enthalpy if energy balance is requested OUTLET ATTRIBUTES PV TH if energy balance is requested 7.3.3.3. RIGOROUS MULTI-STAGE SEPARATION DESCRIPTION This capability is meant to apply to actual staged oper- ations as well as to continuous contacters modeled by equi- librium stages. Liquid-liquid extraction is included. In all cases the results are rigorous in the sense that the heat, material, and equilibrium relations are satisfied for each stage. 183 INLET STREAMS Any number of vapor-liquid to specified stages. Enthalpy must be known. OUTLET STREAMS Any number of saturated vapor or liquid from specified stages. 2. 3. 4. 5. 6. REQUIRED SPECIFICATIONS Many variations of column configuration are possible. The user must specify: 1* # of stages Feed stream location Total, partial, or no condenser Total, partial, or no reboiler Side draw location and phase Interheater/cooler location The remaining specifications depend on the degrees of freedom available. In general the user may specify one quantity for each sidedraw, for each interheater/cooler, for a condenser, and for a reboiler. The possible quantities are: Condenser duty Condenser temperature Reflux ratio Reboiler duty Reboiler temperature 184 Product stream rate Component or group specification in a product stream Mole or mass fraction Mole or mass flow Recovery fraction Interheater/cooler duty Assuming the proper number of specifications are made, quantities not specified will be calculated. It can not be overemphasized that it is very easy to specify a system that Is physically infeasible or that is difficult to solve numerically. A thorough understanding of the process is required to properly specify a multistage separation. ADDITIONAL RESULTS Complete stage-by-stage profiles of temperature, vapor and liquid enthalpy and composition, and phase rates. Any quantity listed under optional specification but not specified PPS REQUIREMENTS K-values and phase enthalpies OUTLET ATTRIBUTES T, P, H, and V 185 7.3.4. SOLIDS PROCESSING 7.3.4.1 CRUSHER DESCRIPTION This capability can simulate normal steady-state, closed- circuit continuous grinding systems. It can be used to predict the size distributions produced by various types of crushers, grinders, mills, or mill-classifiers. It is assumed that nothing happens to the solid, e.g., coal, except for the size reduction. No moisture is lost. Bond Law is used to predict the energy requirement. Defaults supplied are for coal pulverizers. INLET STREAM INFORMATION REQUIRED enthalpy, temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of fluid phase, composition of fluid phase. USER SPECIFICATIONS REQUIRED mill matrix (or in terms of classifer matrix, selection matrix and breakage matrix) OUTLET STREAM INFORMATION CALCULATED enthalpy, temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of fluid phase, composition of fluid phase. 186 ADDITIONAL RESULTS energy requirement PHYSICAL PROPERTIES REQUIRED solid crushing work index 7.3.4,2. VENTURI SCRUBBERS DESCRIPTION This capability can be used to either design or analyze venturi scrubbers. In design option, the maximum size of the venturi scrubber and the maximum gas flow rate are specified by the users. Multiple scrubbers are specified automatically when the maxima are exceeded. INLET STREAM INFORMATION REQUIRED temperature, pressure, laass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas. USER SPECIFICATIONS REQUIRED Analysis Option: diameter of the venturi scrubber throat, gas flow rate, liquid flow rate Design Option: maximum diameter of the venturi scrubber throat, maximum gas flow rate, solid removal efficiency . lo.VjUyir.j^tjtiBttUmrif 187 OUTLET STREAM INFORMATION CALCULATED temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas ADDITIONAL RESUTLS Analysis Option: solid removal efficiency Design Option: diameter of the venturi scrubber throat, gas flow rate, liquid flow rate PHYSICAL PROPERTIES REQUIRED liquid viscosity, liquid partial pressure 7.3.4.3. CYCLONE DESCRIPTION This capability can be used to design or analyze cyclones in which solid particles are being removed from the inlet gas stream. Standard shape of cyclones is used as default for design. In design option, the maximum cyclone size or the maximum pressure drop can be specified by the users. Multiple cyclones in parallel are automatically specified in order to meet the required efficiency. INLET STREAM INFORMATION REQUIRED temperature, pressure, enthalpy, mass flow rate of solid, particle size distribution, mass flow rate of gas, composi- tion of gas 188 USER SPECIFICATIONS REQUIRED Analysis Option: cyclone dimensions, number of cyclones in parallel Design Option: solid removal efficiency, maximum pressure drop, cyclone dimension limits OUTLET STREAM INFORMATION CALCULATED Outlet Solid Stream: temperature, pressure, enthalpy, mass flow rate, particle size distribution. Outlet gas Stream: temperature, pressure, ethalpy, mass flow rate of entrained solid, particle size distribution, mass flow rate of gas, composition of gas. ADDITIONAL RESULTS Analysis Option: solid removal efficiency Design Option: cyclone dimensions, number of cyclones in parallel. PHYSICAL PROPERTIES REQUIRED solid particle density, gas viscosity 7.3.4.4. ELECTROSTATIC PRECIPITATOR DESCRIPTION This capability calculates electrostatic precipitator dimensions required for a given removal efficiency and gas flow rate or calculates removal efficiency from gas flow rate and dimensions. Pressure drop and power consumption can be calculated also. Standard shape of electrostatic precipitator is used as default for design purpose. 189 INLET STREAM INFORMATION REQUIRED temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas. USER SPECIFICATIONS REQUIRED Analysis Option: equipment dimensions, gas flow rate Design Option: solid removal efficiency, gas flow rate OUTLET STREAM INFORMATION CALCULATED temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas ADDITIONAL RESULTS Analysis Option: pressure drop, power consumption, solid removal efficiency Design Option: equipment dimensions. PHYSICAL PROPERTIES REQUIRED gas viscosity, solid dielectric constant, solid specific gravity 190 7.3.4.5 SLURRY PUMP DESCRIPTION This capability can be used to raise the pressure of a slurry stream to a specified value. Correlated curves can be input to estimate slurry pump efficiency. Defaults are supplied. INLET STREAM INFORMATION REQUIED temperature, pressure, enthalpy, mass flow rate of slurry USER SPECIFICATIONS REQUIRED Analysis Option: pump efficiency, work done to the slurry stream Design Option: pump efficiency, outlet pressure OUTLET STREAM INFORMATION CALCULATED temperatue, enthalpy, mass flow rate of slurry Analysis Option: pressure ADDITIONAL RESULTS Design Option: work done to the slurry stream PHYSICAL PROPERTIES REQUIRED molal volume of slurry 191 7.3.4.6 SOLID-ENTRAINED FLOW DRYER DESCRIPTION This capability can be used to simulate the sol id- entrained flow dryer. An example is the drying of coal in a power station mill, in which the pulverized coal from the mill is entrained by a hot gas and carried through a duct and into a burner. The outlet compositions, temperature, pressure and flow rate from the dryer can be calculated from the inlet data and the solid residence time. INLET STREAM INFORMATION REQUIRED Solid Phase: temperature, pressure, enthalpy, mass flow rate, particle size distribution, water content Gas Phase: temperature, pressure, enthalpy, mass flow rate, composition. Hot Gas Stream: temperature, pressure, enthalpy, mass flow rate, compos tion USER SPECIFICATIONS REQUIRED solid residence time in dryer OUTLET STREAM INFORMATION CALCULATED temperature, pressure, enthalpy, mass flow rate of solid, particle size distribution, water content in solid, mass flow rate of gas, composition in gas ADDITIONAL RESULTS none 192 PHYSICAL PROPERTIES REQUIRED saturated water vapor pressure, latent heat of water vapor- ization, diffusion coefficient of water in gas phase, dry solid density, dry solid heat conductivity, wet solid density, solid surface heat transfer coefficient in dryer, solid surface mass transfer coefficient of water in dryer. 7.3.4.7 FABRIC FILTER DESCRIPTION A fabric filter is a fabric through which dust laden gases are forced. An enclosed collection of such filter bags is a baghouse. Figure below shows a three cell baghouse, with the filter bags in two cells filtering while the filter bags in one cell are being cleaned. This capability can be used to analyze and design fabric filters. Users provide the size and shape of bags or use defaults. <=! Gas out Filterin Gas in ■aiMt i wiM w efrii^ ' 19 3 INLET STREAM INFORMATION REQUIRED temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas. USER SPECIFICATIONS REQUIRED maximum pressure drop, clean cloth pressure drop, dust cake permeability, filtering velocity Analysis Option: number of cells in use, number of cells being cleaned, number of bags per cell, filtering area per bag Design Option: solid removal efficiency OUTLET STREAM INFORMATION CALCULATED temperature, pressure, mass flow rate of solid, particle size distribution, mass flow rate of gas, compostiion of gas OTHER CALCULATED QUANTITIES Analysis Option: solid removal efficiency Design Option: number of cells in use, number of cells being cleaned, number of bags per cell, filtering area per bag PHYSICAL PROPERTIES REQUIRED gas viscosity, solid specific gravity 194 7.3.4.8. FLUIDIZED DRYER DESCRIPTION The fluidized dryer capability calculates the particle size distribution of the overhead and overflow streams, the dryer dimensions, the pressure drop and an average moisture content assigned to all particle sizes. The model of quasi-steady-state drying of a sperical particle used in the solid-entrained flow dryer capability is adopted to calculate the moisture content of solid. Figure below shows the fluidized dryer. Solid Feed Solid-Entrained Product ^y Solid Overflow Product Gas Feed 195 INLET STREAM INFORMATION REQUIRED Gas Feed: temperature, pressure, enthalpy, mass flow rate, composition Solid Feed: temperature, pressure, enthalpy, mass flow rate, particle size distribution USER SPECIFICATIONS REQUIRED weight of solid hold-up in dryer, overall heat transfer coefficient of the dryer, gas superficial velocity, ambient temperature OUTLET STREAM INFORMATION CALCULATED Solid Overflow Product: temperature, pressure, enthalpy, mass flow rate, particle size distribution. Overhead Product: temperature, pressure, enthalpy, mass flow rate of solid, particle size distribution, mass flow rate of gas, composition of gas ADDITIONAL RESULTS pressure drop, dryer dimensions. PHYSICAL PROPERTIES REQUIRED gas density, solid density, dry solid density, gas vis- cosity, dry solid thermal conductivity, solid surface heat transfer coefficient in dryer, solid surface mass transfer coefficient of water in dryer, diffusivity of water in gas, latent heat of vaporization of water. w 196 7.3.5 HEAT EXCHANGERS 7.3.5.1 HEATER/COOLER WITH NO PHASE CHANGE DESCRIPTION This capability can be used to model heat addition or removal to or from a stream that does not undergo phase change. Applications include gas streams with entrained solids, superheating, etc. For streams that do undergo phase change the phase equilibrium capability applies, but the computational effort is greater. INLET STREAMS A single vapor, liquid, solid, vapor-solid, or liquid-solid. Enthalpies must be known for the specified heat addition option. OUTLET STREAMS One, same type as inlet REQUIRED SPECIFICATIONS There are two heater/cooler options. 1. Outlet temperature and pressure specified, heat duty is computed if inlet enthalpy is known. 2. Heat added and outlet pressure specified, outlet temperature is computed. 197 ADDITIONAL RESULTS OPTION 1: Heat duty if inlet enthalpy is known. PPS REQUIREMENTS Phase enthalpies. 7.3.5.2 SHORTCUT HEAT EXCHANGER - NO PHASE CHANGE This capability may be used to model heat exchange between two process streams that do not undergo phase change. Several exchanger flow configurations are offered. The capability is shortcut in that overall (lumped parameter) relations are used. INLET STREAMS A single vapor, liquid, solid, vapor-solid, or liquid-solid. Enthalpy must be known. OUTLET STREAMS One, same type as inlet. REQUIRED SPECIFICATIONS There are three specification options: 1. Cold stream outlet temperature is specified, hot stream outlet temperature is calculated. 2. Hot stream outlet temperature is specified, cold stream outlet temperature is calculated. 3. Area is specified, both outlet stream temperatures are specified. 198 In all cases both stream outlet pressures are specified as well as the overall heat transfer coefficient to be used. There are also several configuration options such as co-current, counter-current, multipass shell and tube, and crossflow. This list is open-ended at this time. ADDITIONAL RESULTS Heat duty Effectiveness factor Options (1) and (2); Area PPS REQUIREMENTS Phase enthalpies OUTLET ATTRIBURES Temperature, Pressure, Enthalpy 199 8. COST ESTIMATION AND ECONOMIC EVALUATION SUBSYSTEM 200 8.1 INTRODUCTION Economic evaluation techniques for process plants must meet many requirements. There should be sufficient detail for the capital expenditure decision. Furthermore, the presentation of this detail must be simplified so that the management can use the results effectively in the decision making process. ASPEN should meet the requirements mentioned above so that the cost estimation and economic evaluation subsystem (CES) should have a capability to compute equipment sizes and costs and prepare capital estimates of process plants, which include estimates and analysis of fixed capital investment costs, oper- ating costs, profitability, and sales forecast. It should also be possible to estimate costs of a subsystem or unit operation in the process plants. Moreover the CES should provide the following two levels of accuracy for cost estimations of process plants and the user can choose methods in accordance with his demand: (1) Order-of-Magnitude Estimate for screening and process engineering studies which is based on previous similar cost information; (2) Study Estimate which is based on a knowledge of major items of equipment of the process and whose accuracy is not sufficient for budgeting but a guide to further interest. To meet the requirements, the CES consists of the following functions to provide cost estimates and economic evaluation for chemical processes with options at several levels of detail: 201 (1) Sizing and costing of process equipment (2) Cost data bank (3) Fixed capital investment estimate (4) Operating cost estimate (5) Profitability calculation (6) Sales forecasting (7) Sensitivity analysis The relationships among these functions in the CES are shown in Figure 8.1. In addition, the CES should provide three kinds of usage: the entire system of the CES as a whole, any individual part alone and any sequential series of parts. 8.2. SIZING AND COSTING OF PROCESS EQUIPMENT The usage of this function in the CES is to calculate sizes and costs of each process equipment and may be considered for two engineering stages for process plants. (1) Process Simulation Stage At this stage, a comprehensive capital cost estimate is not necessary but a quick preliminary sizing and costing is required for calculating cost of main plant items which do not include sizing and costing of utility facilities, but calculation of utility consumptions. wmm 202 r\ n^ /f^ [ -J 1 [ (/) E-A r-b ^fe C/3 u ^ -" l-H tU o s t^ < u oo asr 3 M ^ u C a > to 04 w 2 X l-H tsj [I, ro Oi (—1 l-H O r^^ LU. J I CO J E- CO >- w CO W z o l-H 3 i o u o < o H CO Ul E- w O U cu X 00 s 3 00 waHdOMSaficiC*: 203 (2) Cost Analysis and Evaluation Stage At this stage, more precise information and detailed sizing functions are required for cost analysis of fixed capital investment of the plant. Moreover, sizing modules should include routines for calcula- tions of utility consumptions and sizing of utility facilities such as required sizes of steam generation units and water treatment units. If the type of a heat recovery system is not specified, these modules merely calculate utility consumptions for the purpose of cost analysis by summing up heat duties. To analyze this usage, ASPEN should have a capability of comprehensive sizing and costing of main plant items with options at these two levels of detail, so that the user can specify methods and types of equipment, equipment materials and data on intervals which depend on equipment types. If these required values are not specified by the user, the default values should be supplied for costing and evaluation of the process, and most of the simple methods may be assumed as default methods. Equipment sizes or relevant sizing parameters can be estimated by short-cut or empirical methods. Most empirical sizing methods and cost calculations are based on correlation techniques so that, especially, the ranges of their correlations soulc be taken into consideration. In the case out of the range, warning statements should be printed out. In addition, it is still difficult to estimate accurate sizes of newly developed equipment such as coal gasification 204 units or solid handling equipment such as magnetic separator, dust collector or conveyor. In this case, the user can specify easily the sizes of this particular equipment so that the user's input data may be used for the cost estimation. INPUT AND OUTPUT INFORMATION Input data for sizing and costing equipment may be con- sidered as follows: (1) Estimation method. (2) Type of utilities to be used including some parameter. (3) Special data defining individual equipment such as equipment type, nuraber of parallel and stand-by pieces of equipment, and mateials of equipment. (4) Stream and duty data such as follow rates, tempera- ture, pressure, and enthalpy as optional data which can override data from the unit operation subsystem. (5) Cost data of equipment materials and utilities as otpional data which can override data from the cost data bank. (6) Size parameters of equipment as optional data. Output results of this function are as follows: (1) Size parameters of individual equipment. (2) Costs of individual equipment and equipment parts Sizing and costing modules involved in the CES are listed in Table 1. 205 8.3. COST DATA BANK The cost data bank will supply several kinds of cost data which are required for costing of process equipment and cost estimation and economic evaluation functions for process plants. The cost data bank for ASPEN should include the following items: 1) Process equipment costs 2) Labor categories and wage rates 3) Chemical material and utility costs Raw material, product material and utilities 4) Fixed capital investment factors and costs Onsite facilities and offsite facilities 5) Unit operating costs Process unit conversion costs and plant overhead expenses 6) Currency conversion factors 7) Escalation rates Table 2 shows details of cost data items involved in the cost data bank. Some of the cost data are stored in the data bank as the original data, but most of them are stored in the form of correlations which may be linear or non-linear equations with one or multi-parameters. These correlations associate with standard deviations, % errors and ranges of applicability, etc. The stored cost data can be replaced temporarily by the 206 user's input data. If the cost correlations are used for points out of the ranges, the accuracy of the estimates will be decreased, and extraordinary values may often be obtained. In this case, warning statements should be printed out so that the user can find the results having lack of reliability. Moreover, the cost data bank should be updated periodically so that most current data may be available for the cost estima- tion system in ASPEN. 8.4. FIXED CAPITAL INVESTMENT COST ESTIMATE This function in the CES will be used to calculate fixed capital investment which provides the physical facilities required to put a project in operation. Fixed capital investment can be divided into various cate- gories: design and engineering, land purchase and improvement, manufacturing, receiving and shipping, and start-up. The esti- mation logic of fixed capital investment is shown in Figure 8.2. The CES should provide the following two levels of accuracy for cost estimations: (1) Order-of-magnitude estimate: This method was developed primarily for preliminary screening of alternatives of process plants and based on power-law correlations with previous similar information so that it can be called "short-cut" method. In this method, no equipment sizing is necessary for fixed capital investment estimates, but its accuracy is reduced. E-t iJ 2 < M ei S EH H E-i W CU W O < W U u > z M 207 u Q Ui •J z U M H fl4 Im X » c •H +J g •H ■P (0 0) +J G s (Q 0) > c •H cn rH 4J (d CO •P •H u P4 Id •a V4 (d •0 >i 0) X '0 •H c (M (d U 0^ c <4-l •H -0 '0 rH •H Xi 3 4J JH 0) s s, D» C •pH •H 4J U (d •H M > 4J k c D^ 0) C g •H 04 cu •H •H 04 tr 0) k c m •H 4J (0 id u iH •H rH D^ (d +J t^ (0 c (N •H • 0) 00 t3 0) 3 U iH P D^ c •H H h 208 (2) Study estimate: This method uses plant cost factors to estimate tota.. investment costs from main plant equipment. These cost factors are applied to account for materials, labor, etc. required for installation of process plants. The basis of the estimation procedure is th?j major equipment costs based on carbon steel. Cost ratios are used to generate the direct investment. For this estimate, several kinds of estimation methods are available as follows: a. Factorial methods which are based on the total cost of major plant equipment of each process and factors developed from statistical analysis for several project types. b. Modular methods which are based on the total costs for each plant equipment type and a set of factors for materials, piping, installation, etc. c. Analytical methods which are based on the total costs for each plant equipment type and corre- lations including labor, materials and miscel- laneous, and piping material factors in terms of the total costs of heat exchangers, vessels, piping plus drivers, and tower shells. d. Equipment Ratio methods which are based on the costs of each major plant equipment and cost ratios. The costs are based on carbon steel. Figure 2 shows calculation logics of the Equipment R?.tio methods. 209 The CES should supply these methods for fixed capital investment costs at several levels of detail and reliability. The user can choose one of these method and specify some factors Which depend on the method. INPUT AND OUTPUT INFORMATION Input data for fixed capital investment estimates may be considered as follows: (1) Estimation method. (2) Factors as optional data which depend on the estimation method. (3) Major equipment cos^t to be used if sizing and costing program has not applied. (4) Process utilities and general facilities, if not calculated based on requirements. (5) Cost data and factors as optional data which can override data from the cost data bank. Output results of this function are as follows: (1) Fixed capital investment cost of the plant including total equipment and labor costs. (2) Costs of utility facilities and general facilities. (3) Design costs and field expense of the plant. 210 8.5 OPERATING COST ESTIMATE This function will be used to calculate operating cost, or manufacturing cost, involved in keeping a project, operation or piece of equipment running and producihg. Operating costs, which can be broken down into the vari- able, semi-variable, fixed portions and distribution costs, are related to production rates, investment cost, operating labor require- ments, and raw material and utility costs. Operating labor requirements are estimated on the basis of men per equipment unit for each equipment class. Distribution costs are esti- mated from other operation data modified to the situations under study. Each cost item is calculated by the use of cost data stored in the data bank and ratios or factors including scalling ones for power-law methods. INPUT AND OUTPUT INFORMATION Input data for operating capital estimates may be optional data as follows; (1) Estimation method. (2) Number of equipment and process conditions. (3) Fixed capital investment (4) Operating data (e.g., number of operators, specification of utilities, etc.) %^isss 211 (5) Cost data (e.g., material cost, labor cost, etc.) (6) Economic data (e.g., distribution costs, adminis- tration, selling and general expense, interest rate, etc.) Output results on operating costs £or process plants include the following items: (1) Direct production costs These costs may be considered as variable costs which are proportional to the production rate. A. Raw Materials and Utilities a. b. c. d. Raw material Processing material Utilities Others (maintenance material, operating supplies, and royalties and rental) B. Labor a. b. c. Direct operating and maintenance labor Operating and maintenance supervision Payroll burden on all labor charge (2) Indirect production cost These costs may be considered as semi-variable costs which have some dependence on both £ixed and variable costs. 212 A. General work expenses B. Indirect payroll costs C. Ccntingencies (3) Fixed costs These costs are independent on production rate. A. Factory overhead including taxes and insurance- general work expense B. Depreciation (4) Distribution Costs These costs include costs of loading, packing and shipping. A. Materials B. Labor C. Overhead 8.6. PROFITABILITY CALCULATION The comipon criteria used in economic evaluation of process plants may be prof itability „ The measure of the profitability is usually some of the following economic indices which can be calculated in the CES by the use of cash-flow techniques; (1) Net present value (NPV) which converts all future cash flow dollars to an equivalent amount at time zero by the use of discounting techniques. (2) Present value (PV) which indicates the value of all 213 monies paid or received in the future to evaluate large investment projects as economic consequences of time. The present value is established from the following relationships: PV » Future Transaction x Discount Factor (3) Discounted cash flow rate of return (DCFRR) which measures the profitability of the project in terms of relative discount rate by a series of trial and compu- tations using various rates of interest. (4) Return on investment (ROI) which appeals to management in the formats such as the profit and loss state- ments. The return on investment is expressed by the following relationships: Benefits - Costs - Expenses - Income Taxes ROI % « 100 x Capital Investment " (5) Payout time (PT) which measures exposure and elapsed time between initial expenditures and full recovery of the investment costs. FT (YEARS) = |§Y These items of the profitability may indicate the financial effect of the operational parameters on the investment decision. Indices ROI and PT are oversimplified representa- tions of the overall venture and in general they lay emphasis on the short term prospects of the project. Indices NPP, PV and DCFRR enable a more sophisticated representation and allow for the longer term aspects of the project. ' "m^m 214 8.7. SALES FORECASTING Selling prices and sales volumes in economic evaluations are very important because any change in price levels are directly reflected in profits. But in spite of greater signif- icance of these items on profitability, accuracy of estimates are usually of the order of V- 50%, Market research data form the basis for estimation but it appears that relatively little improvement in estimation techniques has been incorporated in economic evaluations. Sales forecasting function is composed of three phases: (1) The relationships between selling prices of products and production rates are correlated on the basis of the "learning curve" principle. (2) Variations of prices with time are also incorporated in this function. (3) The prediction of sales volume is based on common trend curves such as polynomials, exponentials and modified exponentials. The assumption which usually prevails during preliminary analysis is that production volume equals actual sales volume. During a typical venture period for process plants, volume builds up following start-up, levels off and then falls towards the end of the period. A combination of the common trend curves can be used to represent the sales volume profile. The initial and final periods can be estimated using exponential growth curves and the mid-venture period can be fitted to a polynomial curve. 215 But it is still difficult to predict sales forecasting items accurately. Therefore, the CES should accept the user supplied data such as selling price profile and sales volume history. 8.8 SENSITIVITY ANALYSIS This function will be used to do sensitivity analysis, which studies effects on fixed capital investment, operating cost and profitability, by changing some parameters such as the plant conditions, the cost data of materials, tax rate, infla- tion, etc. from maximums through average to minimum. It should be possible to choose any variable as an independent parameter for the sensitivity analysis. Some examples of independent variables are listed as follows: Equipment parameters (e.g., heat duty, number of trays), stream parameters (e.g., flow rate, etc.) Raw material cost Product material cost (selling price) Utility cost Labor cost Maintenance cost Chemical and catalyst cost Byproduct cost Operating cost Research expense Investment cost 216 Sales volume Byproduct rate Tax rate Plant lifetime Plant capacity Royalties Output on the results of sensitivity analysis may be in a form of graphic charts or tables which will be readily under- standable for cost engineers and managements. 217 Table 1. SIZING AND COSTING BLOCKS FOR MAJOR PROCESS EQUIPMENT UNIT OPERATION S/C BLOCK TYPE ADDITIONAL TYPE SEPARATION TWO PHASE FLASH MULTIPHASE FLASH DISTILLATION VESSEL 1 EVAPORATOR VESSEL 2 DISTILLATION TOWER (TOWER WITH REBOILER AND CONDENSER) TOWER TOWER TOWER WITH REBOILER TOWER MAGNETIC SEPARATOR CENTRIFUGE CLARIFIER VESSEL 3 ABSORPTION STRIPPING EXTRACTION SOLID SEPARATION ADSORPTION ION EXCHANGE HEAT TRANSFER NO PHASE CHANGE H.E. HEAT EXCHANGER PHASE CHANGE W/0 BACKMIXING H.E. PHASE CHANGE WITH BACKMIXING H.E. HORI ZONTAL VERTICAL CONICAL EXTERNAL CIRCULATION INTERNAL CIRCULATION LONG TUBE WIPED FILM JACKET TYPE HORIZONTAL VERTICAL CONICAL TRAY TOWER - SEIVE TRAY - VALVE TRAY - BUBBLE CAP TRAY PACKED TOWER same as DISTILLATION TOWER HORIZONTAL VERTICAL SHEEL AND TUBE DOUBLE TUBE JACKET TYPE 218 9 Table 1 (Continued) UNIT OPERATION S/C BLOCK TYPE ADDITIONAL TYPE THERMOS YPON REBOILER HEAT REQUIREMENT HEAT EXCHANGER AIR FIN COOLER SLURRY HEATER SLURRY HEATER PRESSURE CHANGE PUMP PUMP CENTRIFUGAL RECIPROCATING DIAPHRAM ROTARY POSITIVE DISPLACEMENT SLURRY PUMP SLURRY PUMP COMPRESSOR '11' COMPRESSOR/BLOWER BLOVffiR FAN CENTRI FUGAL/AX I AL - NO DRIVER - STEAM TURBINE - ELECTRIC MOTOR - GAS TURBINE RECIPROCATING (same as CENTRIFUGAL) ROTARY TURBINE TURBINE STEAM GAS ENGINE EJECTOR EJECTOR REACTOR EXTENT REACTOR VESSEL TUBULAR YIELD GIVEN FIXED BED EQUILIBRIUM FLUIDIZED BED STIRRED TANK MIXING VESSEL PLUG FLOW MOVING BED 1 219 Table 1 (Continued) UNIT OPERATION S/C BLOCK TYPE ADDITIONAL TYPE COAL GASIFICATION COAL REACTOR GASIFIER COAL LIQUEFACTION LIQUEFACTION COAL OEVOLATILIZER DEVOLATILIZER COAL COMBUSTER COMBUSTER COMBUSTER FURNACE PROCESS HEATER PYROLYSIS REACTOR FURNACE CRYSTALLIZER CRYSTALLIZER SOLID HANDLING CYCLONE DUST COLLECTORS SETTING CHAMBER ELECTROSTATIC CYCLONE PRECIPITATOR ELECTROSTATIC HYDROCLONE PRECIPITATOR FRBRIC COLLECTOR GRINDER/CRUSHER GRINDER/CRUSHER BALL MILL ROD MILL HAMMER MILL FILTER/SCREEN FILTER ROTARY VACUUM CLASSIFIER VERTICAL DISK HORIZONTAL PAN PLATE AND PRAME LEAF CENTRIFUGAL DRIER DRIER TRAY PNEUMATIC CONVEYOR ROTARY SPRAY BATCH SCRUBBER GAS SCRUBBER SPRAY TOWER PACKED TOWER VENTURI ORIFICE FLOOD DISH Table 1 (Continued) 220 UNIT OPERATION S/C BLOCK TYPE ADDITIONAL TYPE SOLID TRANSFER CONVEYOR BELT BUCKET ROLLER SCREW VIBRATING PNEUMATIC ZIPPER SLURRY OFFSITE FACILITY TANK/FLASH STORAGE TANK CYLINDRICAL - GAS - LIQUID - CRYOGENIC SPHERICAL - GAS - LIQUID - CRYOGENIC COOLING TOWER COOLING TOWER MECHANICAL- DRAFT ATMOSPHERIC TOWER NATURAL- DRAFT IP^ 221 Table 2. COST DATA BANK ITEMS 1. Process Equipment Costs for Each Type of Equipment a. Equipment Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Base Costs f. Adjustment Factors Design Types and Sizes Equipment Materials Design Pressures g. Base Module Cost Factors h. (Material Cost)/(Labor Cost) Ratios i. Direct Cost Factors j. Indirect Cost Factors 2. Labor Categories and Wage Rates a. Location Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Direct Operating Labor Costs f. Maintenance Labor Costs g. Operating Supervision Costs h. Maintenance Supervision Costs i. Labor Efficiencies 222 Table 2 (Continued) 3. Chemical Material and Utility Costs a. Material Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Material Costs Product Materials Raw Materials Operating Supplies Maintenance Materials Utilities 4. Fixed capital Investment costs for Each Type of Process A. Onsite Facilities a. Process Names and Codes b. Exponential Facotrs c. Time Bases d. Cost Indices e. Reference Codes f. Dollar Magnitudes g. Direct Field Labor Mainhours h. Material and Labor Factors 223 B. Offsite Facilities a. Process Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Percents Process Unit Costs 5. Operating Capital Costs for Each Type of Process A. Process Unit Coversion Costs a. Process Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Ptocess Unit Conversion Costs B. Plant Overhead Expenses a. Plant Names and Codes b. Time Bases c. Cost Indices d. Reference Codes e. Percent Capital Investments 6. Currency Conversion Factors a. Currency Unit Names and Codes b. Time Bases c. Exchange Rates 224 Table 2 (Continued) 7. Escalation Cost Rates a. Location Names and Codes b. Time Bases c. Reference Codes d. Cost Indices TABLE 3, LIST OF COST ESTIMATION AND ECONOMIC EVALUATION BLOCKS 225 BLOCK TYPE DESCRIPTION CAPITAL FIXED CAPITAL INVESTMENT Two levels of accuracy are prepared; - Order of magnitude estimate Study estimate OPERATING OPERATING COST - Direct cost Indirect cost - Fixed cost PROFITABILITY PROFITABILITY INDICES - Net present value Present value - Discounted cash flow rate of return Return on investment Payout time FORECASTING SALES FORECASTING Sales volume Selling price SENSITIVITY SENSITIVITY ANALYSIS Sensitivity on: - Fixed capital investment Operating cost - Profitability indices t>U.S. GOVERNMENT PRINTING OFFICE:! 978 -7itO -306/ 1+266 REGION N0.4 UNIVERSITY OF ILLINOIS-URBANA 3 12101916754