SilSiiS:';;::' 'iH'':v ^^ ..s^^. '^'(mm > 0-r ^. o A,^ o"""* ^. <> *'7VT* ,0 .-^^ . ^°"^. ./\':=^;:,\. /.^^;^^> .^^\.^%X /.c:^,% .^^ y- 0^ »•"• C- ♦* i'^" . .^" . ip^^. "^#^ .^% ,<^^ "^^, .x^^ • C^MR^ O V .•1°^ •^--0^ .^^^'- "^bv^ f^^«; '-^^0^' :^^^'' -^o-^ » A 4«^ '0- ■<* ..'^ ..G^ ^^ "'TVT* ,G JC ®°^^ Bureau of Mines information Circular/1986 Reliability Prediction for Computerized Mine-Monitoring Systems By Raymond M. Kacmar and Edward F. Fries UNITED STATES DEPARTMENT OF THE INTERIOR \-- '^4U^ iixfer, i^mM'fli^^ Information Circular 9088 Reliability Prediction for Computerized Mine-Monitoring Systems By Raymond M. Kacmar and Edward F. Fries UNITED STATES DEPARTMENT OF THE INTERIOR Donald Paul Model, Secretary BUREAU OF MINES Robert C. Norton, Director T 3^5 c\0 %^ Library of Congress Cataloging in Publication Data: Kacmar, Raymond M Reliability prediction for computerized mine-monitoring systems. (Information circular; 9088) Bibliography: p. 13. Supt. of Docs, no.: I 28.27:9088. 1. Mine safety-Data processing. 2. Mine gases-Data processing. 3. Mine fires-Data processing. 4. Reliability (Engineering). I. Fries, Edward F. II. Title. III. Series: Informa- tion circular (United States. Bureau of Mines); 9088. TN295.U4 622 s [622'.8] 86-600079 CONTENTS Page Abstract ^ 1 Introduction 2 System descriptions 2 Reliability predictions 2 Methodology . 2 Parts-count reliability 4 Prior reliability data 4 Reliability prediction assumptions 4 Reliability models 5 Prediction data and analysis 7 Environmental impact on equipment 8 Conclusion. . . « 8 Bibliography 13 ILLUSTRATIONS 1. Block diagram of system A 3 2. Block diagram of system B 3 3. System A reliability configuration 5 4. System B reliability configuration 6 5. Sample form for assembly-rate summary 7 6. Sample form for raicrocircuit failure-rate summary 8 7. Gp environment for system A 9 8. Gp environment for system B 9 9. Data processing subsystem for systems A and B 10 10. Communications subsystem for system A 10 1 1 . Power subsystem for system A 10 12. Transducer subsystem for system A 10 13. Communications subsystem for system B 11 14. Power subsystem for system B 11 15. Transducer subsystem for system B 11 16. System A reliability — adjusted 12 17. System B reliability — adjusted 12 TABLES 1. Functional subsystems 6 2. Environmental impact on equipment 11 UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT °C degree Celsius MM million h hour V volt Hz hertz RELIABILITY PREDICTION FOR COMPUTERIZED MINE-MONITORING SYSTEMS By Raymond M. Kacmar^ and Edward F. Fries^ ABSTRACT This report presents the Bureau of Mines research on the hardware reliability prediction for two Bureau monitoring projects. The basic concepts of reliability predictions are introduced along with the reli- ability models and assumptions used for this particular evaluation. The results of the reliability predictions are for ground-fixed and naval- sheltered environments. ^tlectrical engineer, Pittsburgh Research Center, Bureau of Mines, Pittsburgh, PA (now with Defense Contracts Administrative Services, Los Angeles, CA) . '^Electrical engineer, Pittsburgh Research Center. INTRODUCTION By using computerized systems that are appearing in the marketplace, mine opera- tors now can monitor and display the sta- tus of underground conditions. However, before they buy, install, and depend on these systems they need to know how reli- able the data from the system will be; also, they need to predict how future ex- pansions will affect system reliability. The Bureau of Mines is conducting re- search to determine the reliability of current monitoring systems and also to demonstrate that standard reliability prediction techniques can be modified to predict the reliability of equipment op- erating in a mine environment. The mine monitoring systems researched for this report are electromechanical systems that remotely sense various envi- ronmental and operational parameters and transmit the data to a central location where the data are analyzed and/or dis- played. The equipment normally required to perform these functions consists of transducers, telemetry, one or more com- puters, and the associated peripherals and input-output interfaces. In the ba- sic system, the output from a transducer is converted to a format that enables a signal to be transmitted to a central computer station. There, a processor (or system of processors) tabulates the data, compares it with present alarm condi- tions, displays results, logs the data for future reference, and performs other calculations and data management. SYSTEM DESCRIPTIONS Both systems selected for these reli- ability predictions (to be referred to as system A and system B) have been in- stalled as research projects in conjunc- tion with the Bureau. System A is lo- cated in a research mine, owned by the Bureau, in Allegheny County, PA. System B is located in a commercial mine in In- diana County, PA. Although the system configurations are different, they are both representative of current tech- nology. The computer stations used for these systems were selected by the Bureau under competitive bid. They consist of a microprocessor-based controller with a cathode ray tube (CRT) and floppy disk drive and two printers. It should be noted that both systems are currently being modified to upgrade system capabilities. Therefore, the re- liability predictions in this report will only apply to the original system config- urations. However, these predictions can be easily adjusted once the hardware mod- ifications have been completed. The telemetry system used for sys- tem A employed a four-wire frequency shift-keyed communications link. Two of the wires were used to transmit data and two were used to receive data. On the surface, a telemetry card and local modem connected the computer to the telemetry line. At the sensing loca- tions, an analog-to-digital (A/D) con- verter and a remote modem connected the transducers. Previously, there were five communication "outstations" being used. One station was located on the surface and four stations were located under- ground. A block diagram of this system is shown in figure 1. For system B, two "trunk lines" (four- wire telemetry cables, each consisting of two power lines and two data lines) were used to monitor methane, carbon monoxide, and airflow. A block diagram of this system is shown in figure 2. The transducers sent measurements to the central station through an A/D tele- metry card. The analog voltage from the transducer is converted to an 8-bit digi- tal message and is then telemetered to the computer. RELIABILITY PREDICTIONS METHODOLOGY The standard procedures for performing reliability predictions are described in the military handbook (MIL-HDBK) 217D, "Reliability Prediction of Electronic Equipment." When possible, the methods in 217D, paragraph 5.1, are to be used. Mine power Alarm PR Data PR 1 on\/ 60 Hz Battery Telemetry card and local modem 1 1 1 Surface outstation remote modem and A/D 120 V UPS Computer console RS232 60 Hz Interface FSK line 1 1 1 CO CH4 Air Surface Underground Mine outstation remote modem and A/D Mine outstation remote modem and A/D Mine outstation remote modem and A/D Mine outstation remote modem and A/D i 1 1 1 1 1 1 1 1 1 1 1 CO CH4 Air CO CH4 Air CO CH4 Air CO CH4 Air FIGURE 1.— Block diagram of system A. Alarm PR Data PR RS232 1 Power source and line driver Power source and line driver 120V Computer 60 Hz cor is ole Surface _ Lightning and_ surge protectors _Communication_ trunk lines Underground To other transducers JN CO JN JN CH4 JN Air To other transducers JN JN CO JN CH4 JN Air FIGURE 2.— Block diagram of system B. The equations contained in this section of the handbook require the detailed knowledge of the stress and quality lev- els of the components in the equipment under study. However, this detailed in- formation on the parts used for the moni- toring systems was not available for this analysis. Therefore, a combination of the following two methods was used to ar- rive at the predicted system reliability for the two configurations: parts-count and prior experience, Parts-Count Reliability When the methods of paragraph 5.1 can- not be used, MIL-HDBK 217D recommends the use of paragraph 5.2, "Parts-Count Reliability Prediction." This method is based on the type and quantity of the parts used. For this evaluation, it was used for equipment and assemblies whose parts lists were available or could be derived. This prediction procedure has limitations, and past experience has shown it to be conservative. Average stress levels are assumed for parts such that the system reliability can be ex- pressed as 1 MTBF (SYSTEM) = (1) I NiCXe.TTQ.) i=l Where for a given environment MTBF (SYSTEM) = the system average time between failure, h, Xq = generic failure rate for the i^*^ generic part, failures/MM h, Up = quality factor for the i^'^ generic part, Nj = quantity of the i^^ ge- neric part, and n = number of different ge- neric part categories. The following gives a sampling of the assumptions used for the parts-count prediction: 1. Ambient temperature (T^) = 40° C for ground-fixed environment (Gp). 2. Applied stress ratio = 0.5 = actual power dissipated per maximum rated power. 3. Uses average complexities; e.g., any integrated circuit in 501-1,000 gate range uses 875 gates for calculation. 4. A limited number of quality factors. Prior Reliability Data For this method, reliability data for similar parts used in other applications were used. This information was obtained from conversations with reliability ex- perts at the Department of Air Force, Rome Air Development Center (RADC), Reli- ability and Maintainability Engineering Section (RBER), Griff iss Air Force Base, New York. This method was used since several components and assemblies were not of the type that lend themselves to parts-count prediction techniques or could not be broken down into their con- stituent parts. As a result, the reli- ability of these assemblies were based on information obtained from RADC person- nel who have had experience with similar equipment in the past on Air Force sys- tems. Where complexities, environment, and/or duty cycles differed from that of the equipment under study, adjustments were made by RADC. RELIABILITY PREDICTION ASSUMPTIONS As a baseline for the prediction, other assumptions were necessitated as follows: 1. Series model. — A series model was assumed such that a failure of any sub- system or assembly caused a system failure. 2. Duty cycles. — All equipment was as- sumed to be operated at 100% of the sys- tem's operational time with the following exceptions: a. Alarm printer. — The alarm printer was assumed to operate only 10% of the time where the status printer was assumed to operate 100%. This assumption is based on the fact that the alarm printer only operates to output alarm conditions. b. Disk drive and controller. — These assemblies were assumed to be required 5% of the time because they are used only to Initialize the system. 3. Environmental conditions. — The methodology for the prediction was in ac- cordance with MIL-HDBK 217D, which does not have a mine environment. The Envi- ronmental Impact section assesses the system reliability in different environ- ments. For the prediction baseline, a ground-fixed environment was assumed. 4. Parts quality. — As a baseline, com- mercial plastic devices were assumed for integrated circuits and semiconductors (active devices). Passive devices such as resistors and capacitors were assumed to be of commercial quality. 5. Sensing elements. — The reliability of sensing elements themselves was not considered. It was assumed that preven- tive maintenance made their failure rate contribution negligible. Electronics associated with the sensing function are included. 6. Software reliability. — The reli- ability of the system software was not evaluated for this report. However, sim- ilar methods exist to determine software reliability but are beyond the scope of this report. RELIABILITY MODELS For this analysis, the two system con- figurations were separated into four functional subsystems: data processing, communications, power, and transducers. This functional designation was made to simplify future comparison between pre- dicted and actual operation. Table 1 lists the assemblies composing the func- tional subsystems at each location. Fig- ures 3 and 4 indicate how the systems were configured for the reliability mod- els with each block (or assembly) being _ UPS JN Chassis CPU PWB Serial I/O PWB Disk and control CRT and control Power supply 57o Console Printer 100% Printer 10% Tel em PWB 5-V power supply 24-V power supply Local modem JN and FSK 15-V power supply 5-V power supply Remote modem A/D PWB CO Airflow CH4 Surface outstation Surface Mine r 15-V power supply (4) 5-V power supply Remote modem A/D PWB (4) CO (4) (4) Airflow CH4 Mine outstation I FIGURE 3.— System A reliability configuration. required for the operational success of the system. Using this reliability mod- el and the four functional subsystems MTBF = previously discussed, the MTBF of system can be expressed as follows: 1 A(DPS) + X(COMM) + X(PS) + X(TRANS) the (2) where X(DPS) = failure rate for the data processing sub- system failures/MM h, X(COMM) = failure rate for the communication subsystem, and X(PS) X(TRANS) failure rate for the power subsystem, failure rate for the transducer subsystem. TABLE 1. - Functional subsystems Site subsystem System A System B Data processing 1 console and 2 line printers. 1 console and 2 line printers. Communications Telemetry with local modem: 1 line driver assembly, 13 A/D 1 surface outstation, 4 mine boards, 13 junction boxes. outstations, junction boxes, connectors. connectors. Power 1 uninterruptible power sup- Transient protection, 2 junc- ply, junction boxes, power tion boxes, connectors. lines, connectors. Transducers .......... 1 surface outstation and 4 mine outstations with 5 CO, 5 airflow, and 3 CH4 transducers. 1 each CO, airflow, and CH4 transducers. 1— — 1 1 Serial I/O PWB Disk and control CRT and control Chassis — -CPU PWB Power supply Printer I0( 1°^ - Printer 10% '— - 57o - — C on sole — 1 r" - — -- — 1 (2) Telenn PWB — 24-V power supply Enclosure 1 1 Transient protect JN 1 1 p n\Mior cniirr e inW lino rlriuo 1. r 1 Surface Mine — (5) 1 i (5) --. r- (3) 1 CO — A/D PWB 1 1 Airflow A/D PWB CH4 - - A/D PWB 1 1 JN 1 1 1 1 U _ _ J L. _J FIGURE 4.— System B reliability configuration. PREDICTION DATA AND ANALYSIS The data needed to perform the parts- count prediction were obtained from the schematics and parts lists of the various assemblies. A sample worksheet for the calculation is shown in figures 5 and 6. However, because of the proprietary na- ture of the monitoring systems, the ac- tual number and type of components used for the calculations are not shown. The failure rate ( ■ J for a partic- \ MTBF / ular group of similar components is then derived by multiplying the number of components by the generic failure rate (obtained from 217D) and the appropri- ate quality factor. As an example, the failure rate for 12 nonmilitary plastic resistors would be as follows: A = (NiXXeXTTQ), = (12) (0.0010(3), = 0.0396. The failure rate for a particular assem- bly is then obtained by adding the fail- ure rates for all of the components used. Likewise, the failure rate for a subsys- tem is obtained by adding the failure rates for each of the assemblies used. Finally, the predicted reliability of the system is derived using equation 2. Quality factors NAIIq Level 1 NAiiq Level 2 Fami 1y lype Transistor Si NPN Transistor SI PNP Diode SI gen purpose Diode zener LED Power transistor Composition RC Film RL Film RN Wi rewound RW Van WW RA Var comm RV Cerami c CX Tanti lum solid CSR Tantilum nonsolid CSR Al dry CE Low power pulse High power pulse/power Audio trans General purpose High current Circular/rack panel Printed wi re board 2-sided board Toggle and push button Rotary Wi rewrap Wave solder Hand solder Crimp Qty. AG Jan Nonmi 1 hermeti( Nonmi 1 plastic NAIlQ Level 3 Semicondo. RES. CAP TFRS Relay Conn PCB Switch conn. Misc. .016 .024 .0031 .012 .033 1.9 .0011 .016 .018 .18 1.10 .14 .018 .014 .28 .29 .019 .14 .038 .33 1.1 .017 .024 .0029 .0029 .96 .000053 .0006 .0055 .001 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 5.0 5.0 5.0 5.0 5.0 5.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 6.0 6.0 3.0 3.0 1.0 20.0 50.0 1.0 10.0 10.0 10.0 10.0 10.0 10.0 3 3 3 3 3 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 6.0 6.0 3.0 3.0 1.0 20.0 50.0 1.0 Microcircuits. From page 2 8-2 0-1/NH Drawing reference Assembly totals Assembly Name Site FIGURE 5.— Sample form for assembly-rate summary. Bits Qty. ROMS RAMS Hermetic Nonhermetic Qty. Hermetic Nonhermetic AGH NAGH AGNH nagnh AGH Nagh ^iW^ \ GNH 1.320 .02 .d3 W 321-576... .02 .03 .04 .06 577-1120.. .03 .05 .06 .11 1121-2240. .05 .06 .08 .16 2241-5000. .06 .10 .13 .28 5501-llK.. .09 .18 .29 .76 11001-17K. .12 .23 .44 1.2 17001-38K. .20 .44 .88 2.5 38001-74K. .33 .84 1.3 3.9 null mill lllllll llllllll Digital Linear Gates Qty. Hermetic Nonhermetic AGH Nagh AGNH Nagnh 1.20 .016 .018 21-50 .020 .023 51-100 .026 .031 101-500 .044 .06 501-1000... .071 .063 .081 .100 1001-2000.. .11 .19 2001-3000.. .14 .30 3001-5000.. .20 .50 5001-7500.. .28 .83 7501-lOK... .43 1.3 10001-15K.. .61 2.2 15001-20K.. .85 3.4 mill mm Assembly Name Transistors Qty. Hermetic Nonhermetic >GH n>gh >GNH NifiNH 1-32 33-100 101-300 .031 .084 .29 .049 .16 .79 mm mm Summary Hermetic Nonhermetic Total total total Digital RAM ROM Linear Total IC's 8-2x6.5 //////////////// 0x17.5 //////////////// 0-1x35 ////////////// Site FIGURE 6.— Sample form for microclrcult failure-rate summary. The results of the reliability predic- tions for the two configurations are shown in figures 7 and 8. The results on the assembly and subsystem levels are shovm in figures 9 through 15. ENVIRONMENTAL IMPACT ON EQUIPMENT The previous predictions for the sys- tems were performed using the MIL-HDBK- 217D ground-fixed environment. However, as noted before, the handbook does not have a specific environmental factor that represents the operation of equipment in a mine atmosphere. It could be argued that this atmosphere is more severe than ground fixed because of the humidity and contamination-corrosion problems. On the other hand, it could be less severe owing to a lower ambient temperature. As a result of this uncertainty, the envi- ronmental impact on the prediction re- sults were reviewed. Table 2 shows how the MTBF varies as a function of environ- ment and the conditions that define the environment in MIL-HDBK 217D. Figures 16 and 17 show the system reliability re- vised when a naval-sheltered environment is used for the equipment in the mine, and when a ground-fixed (Gp) environment is used for the surf ace 'equipment . CONCLUSION The predicted reliability of mine monitoring systems A and B using the methodology presented (on page 11) in hours; in this paper are UPS — JN ?k = 50 A=3 Chassis CPUPWB Serial I/O PWB Disk and control CRT and control Power supply A =1807 57o Console _ Printer 100% A = 400 Printer 10% A = 40 Tel em PWB I I A =200.1 5-V power supply 24-V power supply Local modem L JN and FSK A= I 15-V power supply 5-V power supply Remote modem A/D PWB CO A = 47.2 Airflow A = 29.4 CH4 A =28.3 191.5 Surface outstation Surface Mine r (4) (4) Total failure | rate: 3894 | MTBF= 251 hi X = 191.5 15-V power supply 5-V power supply Remote modem A/D PWB CO \ = 47.4 Mine outstation FIGURE 7.— Gp environment for system A. (4) (4) Airflow CH4 X=29.4 A =28.3 Chassis A = 1807 CPUPWB Serial I/O PWB Disk and control 5% CRT and control Power supply Console Printer 100% A =400 Printer 10% A=40 (2) Telem PWB 24-V power supply Enclosure Transient protect A= 1.84 JN 7^=0.495 A = 37.49 Power source end line driver Surface Mine (5) (5) (3) CO A = 31.59 A/D PWB A = 23.18 Airflow A=41 A/D PWB A = 23.18 CH4 A =26.89 A/D PWB A= 23.18 JN A = 1.380 I ! I I I ! Total failure rote = 3035.1 MTBF= 329 h FIGURE 8.— Gp environment for system B. 10 ■ " ■ ■ 1 Serial I/O PWB Disk and control CRT and control Chassis — CPU PWB Power supply Printer 100% Printer 10% l^M.ll A = 1512.31 A = 93.16 >=25 A = 90.41 A=68.7I A =400 )k = 40 A_=I807 4 5% Console Total failure rate = A DPS " 2247 MTBF = 445 h FIGURE 9.— Data processing subsystem for systems A and B. Telem PWB A= 13.07 5-V power supply A=26.39 24-V power supply X=63.I5 Local modem A = 9753 A =20014 15-V power supply A = 6.29 5-V power supply A = 26.39 (3) Remote modem A= 8926 A/D PWB A = 23.18 _) L X= 191.48 Surface outstction 15-V power supply A=6.29 5-V power supply (3) Remote modem A/D PWB A =26.39 A = 89.26 A = 23.I8 n A= 191.48 Mine outstation Total failure rate = AroMM "58.4 MTBF= 863 h FIGURE 10.— Communications subsystem for system A. JN and conn A=I.O UPS Junctions and connectors A =50 X = 3 Total failure rate = ^pS = 53 MTBF = 19,000 h FIGURE 11.— Power subsystem for system A. CO Airflow CH4 >=47.4 > = 29.4 A = 28.3 Surface Mine (4) (4) (4) CO Airflow CH4 > = 47.4 A = 29.4 X = 28.3 Total failure rate = ^ TRANS " 526.0 MTBF= 1901 h FIGURE 12.— Transducer subsystem for system A. 11 TABLE 2. - Environmental impact on equipment Example MTBF MIL-HDBK 217 environment Environment definition Ta, °C part E (IC) adjust factor Ground-benign (Gb) Nearly zero environmental stress with optimum engineering oper- ation and maintenance. 30 0.38 0.21 Ground-fixed (Gp) Conditions less than ideal to include installation in per- manent racks with adequate cooling air, maintenance by military personnel, and possi- ble installation in unheated buildings. 40 2.50 1.00 Naval-sheltered (N s) Surface ship conditions similar to Gp but subject to occasional high shock and vibration. 40 4.00 1.67 Naval-unsheltered (Nj).... Nominal surface-ship-borne con- ditions but with repetitive high levels of shock and 75 5.70 3.15 vibration. (2) JN Telem PWB \= 1.38 A= 12.18 1 > = 37.49 (2) 24-V power supply I — X=6.4 Enclosure A =0.33 Power source and line driver (13) A/D PWB > = 23.18 Total failure rate = ^cqivIM MTBF = 2939 h 340.21 FIGURE 13.— Communications subsystems for system B. (2) Transient protect ^=1.84 JN X = 0.495 Total failure rate = "ApS - 4.175 MTBF = 239,521 h FIGURE 14.— Power subsystem for system B. Environment Ground-fixed ^ . . . , Naval-sheltered . System A System B 251 209 329 283 For equipment on surface. For equipment in mine. The actual reliability of monitoring sys- tems in the field will be compared with the predicted reliability once a suffi- cient data base has been established. Areas of potential improvements in reliability can be categorized as (5) (5) (3) CO Airflow CH4 >= 31.59 1 __j >=4I.I A = 26.89 Total failure rate = ^JRANS ' 444.2 MTBF = 2251.7 h FIGURE 15.— Transducer subsystem for system B. 12 UPS 7^ = 50 JN A = 3 Chassis CPU PWB Serial I/O PWB Disk and control CRT and control Power supply > = 1 807 Console Printer I007o > = 400 Printer I07o A = 40 Tel em PWB 5-V power supply 24-V power supply Local modem X =200.1 JN and FSK ■K= I 15-V power supply 5-V power supply Remote modem A/D PWB CO I = 47.4 Airflow CH4 A = 29.4 A = 28.3 I = 191.5 Surface outstation Surface Mine r Total failure rate: 4780 MTBF= 209 hi > = 319.8 15-V power supply (4) (4) 5-V power supply Remote modem A/D PWB CO A = 79.2 14) 14; Airflow CH4 A=49.2 A =47.3 Mine outstation FIGURE 16.— System A reliability— adjusted. Chassis A = I807 CPU PWB Serial I/O PWB Disk and control 5% CRT and control Power supply Console Printer 100% A =400 Printer 10% A = 40 Telem PWB 24-V power supply Enclosure A =37.49 Power source and line driver (2) Transient protect 1= 1.84 JN A =0.495 (each) Surface Mine (5) (5) (3) CO A = 52.75 A/D PWB A =38.7 Airflow A = 68.64 A/D PWB A = 38.7 CH4 A = 44.91 A/D PWB A= 38.7 I. Total failure rate = 3535.8 MTBF = 283 h JN A=2.3I FIGURE 17.— System B reliability— adjusted. part-quality upgrade and elimination of reliability design deficiencies. As pre- viously defined in the prediction as- sumption, the baseline prediction was performed using standard-quality parts. Sensitivity to part-quality as a poten- tial means of reliability is a proven fact based on field experience. Reli- ability of the system can be improved by doing the following: Use hermetic devices in an atmos- phere as corrosive and humid as that of a mine. Use MIL-grade connectors and solder connections instead of integrated-circuit sockets. 13 o Establish a well-defined inspection and preventive maintenance schedule. All equipment should be burned-in at an elevated temperature for at least 100 h with the last 24 h failure free. All printed circuit boards exposed to the mine environment be conformally coated to prevent moisture and corrosion problems . System developments should involve front-end design consistent with the cri- ticality of operation and safety in a se- vere environment. BIBLIOGRAPHY Donald, L. B., and R. M. Baker. An Annotated Bibliography of Coal Mine Fire Reports (contract J0275008, Allen Corp. of Am.). V. 1, BuMines OFR 7(l)-80, 1979, 98 pp., NTIS PB 80-140205; v. 2, BuMines OFR 7(2)-80, 1979, 400 pp., NTIS PB 80-140213; v. 3, BuMines OFR 7(3)-80, 1979, 654 pp., NTIS PB 80-140221. Henley, E. J., and H. Kumamoto. Proba- bilistic Parameters of Whole Processes. Sec. 4.2.5 in Reliability Engineering and Risk Assessment. Prentice-Hall, 1981, p. 180. Kacmar, R. M. Reliability of Comput- erized Mine-Monitoring Systems. BuMines IC 8882, 1982, 10 pp. U.S. Air Force, Reliability Analysis Center. Rome Air Development Center Re- liability Design Handbook. RDH375, Mar. 1976, pp. 19-21, 292. U.S. Army Aviation Systems Command. Introduction to Reliability. Ch. 1.0 in Pocket Handbook on Reliability. Sept. 1975, p. 11. U.S. Code of Federal Regulations. Ti- tle 30 — Mineral Resources; Chapter I — Mine Safety and Health Administration, Department of Labor; Subchapter — Coal Mine Safety and Health; Part 75 — Manda- atory Safety Standards — Underground Coal Mines; July 1, 1983. U.S. Code of Federal Regulations. Ti- tle 30 — Mineral Resources; Chapter I — Mine Safety and Health Administration, Department of Labor; Subchapter D — Elec- trical Equipment, Lamps, Methane, De- tectors; Tests for Permissibility; Fees; Part 18 — Electric Motor-Driven Mine Equipment and Accessories; July 1, 1983. U.S. Mine Safety and Health Administra- tion (Dep. Labor). Injury Experience in Coal Mining. Annu. Inf. Reps. 1973-80. Watson, R. A. An Intrinsically Safe Environmental Monitoring System for Coal Mines. Proc, 6th Conf . on Coal Mine Electrotechnol. , WV Univ., Morgantown, WV, July 1982, pp. 345-360. f- U S. GOVERNMENT PRINTING OFFICE: 1986-605-017/40,059 INT.-BU.OF MINES,PGH.,PA. 28317 4386 405 U.S. D«partm«nt of the Interior Bureau of Mines— Prod, and Distr. Cochrans Mill Road P.O. Box 18070 Pittsburgh. Pa. 15236 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE. $300 I I Do not wish to receive this material, please remove from your mailing list* I I Address change* Please correct as indicated* AN EQUAL OPPORTUNITY EMPLOYER o .V -^ ".^^W* ^'^'' '^^^ °.^«^/ ^^'^'^^^ ^ .<' *bv^ ^^-V'. V f/ \/W^^/ v^*/ \/^\/' v^-/ V V, ^^ '""* A <> *?X ' * .G^ '-"Siii ^:'S:i||||ii LIBRARY OF CONGRESS 002 955 935 1 111 % ,.,::iiJ i't^^;^' ': ■'■■■■■ : *. K ;'ti, y^;~