PII: S0924-0136(98)00182-4 Journal of Materials Processing Technology 80 – 81 (1998) 131 – 135 Expert system for hot forging design A. ngelo Caporalli a,b, Luciano Antonio Gileno a, Sérgio Tonini Button a,* a School of Mechanical Engineering, State Uni6ersity of Campinas, Campinas, SP, 13083 -970, CP 6122, Brazil b School of Engineering at Guaratinguetá, São Paulo State Uni6ersity, Guaratinguetá, SP, 12500 -000, CP 205, Brazil Abstract Planning hot forging processes is a time-consuming activity with high costs involved because of the trial-and-error iterative methods used to design dies and to choose equipment and process conditions. Some processes demand many months to produce forged parts with controlled shapes, dimensions and microstructure. This paper shows how expert systems can help engineers to reduce the time needed to design precision forged parts and dies from machined parts. The software ADHFD interfacing MS Visual Basic v.5.0 and SolidEdge v.3.0 was used to design flashless hot forged gears, chosen from families of gears. © 1998 Elsevier Science S.A. All rights reserved. Keywords: Expert system; Hot forging; Die design 1. Introduction In recent years, hot forging has shown continuous technological changes, mainly in the geometry of the dies and in the dimensional accuracy of forgings. Since hot forging have to be competitive even for small lot sizes, forgings weight and machining allowances are being reduced and therefore subsequent machining is being reduced with low final costs and investments. With hot precision forging and the concept of near-net- shape-manufacturing (NNS), costs with machining are drastically reduced, so forging operations (and their planning) became an important part of the final cost [1,2]. Most of the time (and costs) involved in planning hot forging processes is related to activities strongly depen- dent on human expertise, intuition and creativity [3] and also to iterative procedures involving extensive experimental work. In order to minimize this lead time, related expert knowledge and criteria must be trans- ferred to automated knowledge-based and rules-based systems. Tisza [4] states that there are a small number of expert systems applied in metal forming, when com- pared with other manufacturing fields. The numerous variables involved in metal forming processes make difficult the automation of decision making and plan- ning of these processes, but some examples can be found particularly for multi-stage processes like cold forging and deep drawing [3 – 7] and for hot forging [8,9]. Biswas and Knight [10] point out that the rationaliza- tion and automation of the procedures by expert sys- tems bring benefits to the hot forging industry like rationalized tooling, improved factory layouts, system- atic design procedures, improved estimating and costing procedures, and standardized work planning and pro- cess design. 2. Planning precision hot forging processes Precision hot forging shows some advantages when compared to conventional hot forging: it minimizes subsequent machining, improves forgings quality avoid- ing overlaps and internal defects, reduces forging loads, and can eliminate heat treatment, since internal stresses are reduced. To achieve these characteristics, process variables like forging temperature and billet volume must be precisely controlled [11]. The planning of a new hot forging process can be divided into three groups of procedures, as shown in Fig. 1. In group A the machined part is modified: * Corresponding author. Fax: + 55 19 2393722; e-mail: ser- gio@fem.unicamp.br 0924-0136/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S 0 9 2 4 - 0 1 3 6 ( 9 8 ) 0 0 1 8 2 - 4 A. . Caporalli et al. / Journal of Materials Processing Technology 80 – 81 (1998) 131 – 135132 Fig. 1. Stages for hot forging planning. machining allowances are applied where necessary, a parting line is properly chosen, appropriate radii are applied to fillet and corners, internal and external drafts are added, holes, ribs and webs are designed to keep forging dies undeformed. The forged part is finally designed considering NNS criteria. The process sequence is designed in group B. Consid- ering part characteristics (material, size and geometry), lot size and available equipment, it is possible to define the number of operations and to design the preform. Using the principles of group technology and defining families of parts, it is possible to find similar forgings and adopt an existing sequence or modify it to the new design. The most difficult and costly stage to plan each operation is to determine the correct preform geometry to assure complete filling of the dies. In industry this stage is still based on ‘trial-and-error’ iterative proce- dures to the production of ‘good parts’, after successive measurements of the forgings and corrections in dies geometry. In recent years, many works based upon numerical methods like the finite element method (FEM) show that computerized analysis can help the forming industry to define the best process sequence, reducing time and costs involved [2,12,13]. Once the process sequence is defined, other process parameters are chosen in group C: forging tempera- tures, lubricants, and equipment. The dies are designed considering preforms geometry, thermal expansion, and the use of standardized tools and inserts. 3. The ADHFD system This work shows a semi-automatic system developed to design dies and plan precision hot forging using near-net-shape criteria. The ADHFD (automated de- sign of hot forging dies) is being developed to work with families of automotive gears (longitudinal sections shown in Fig. 2), classified by designs: with or without web, one-sided or two-sided hub, and by characteristic dimensions (external diameter, hub diameter, rim length, hub length). These gears are forged within some near-net-shape principles: small machining allowances, flashless, minimal draft angles and fillet radii [14]. The expert system ADHFD is being developed with four codes. The rule-based system is written in MS Visual Basic v.5.0 and linked to MS Access v.7.0 data-Fig. 2. Families of automotive gears. A. . Caporalli et al. / Journal of Materials Processing Technology 80 – 81 (1998) 131 – 135 133 Fig. 3. Automotive gear. Fig. 4. Gear forging. bases. Metal flow is simulated with Ansys v.5.3 to prevent filling defects and to optimize process sequence. With SolidEdge v.3.0 the forgings and the dies are modeled in 3-D and assembled. The volumes are then calculated and respective technical drawings generated. The system is considered semi-automatic since the user can choose similar forgings, analyse machined part details and decide how to modify the forging and the process sequence. The databases contain detailed information about variables of processes already planned: raw materials properties (stock size and tolerances, flow stress, formability), forging temperatures, lubricants, presses (capacity, table dimensions, stroke and speed), about NNS criteria and about typical process sequences (billet cut-off, heating, preform, finish forging and trimming). 4. An example of application The forging process of the gear shown in Fig. 3 was planned using the expert system developed. This gear is similar to the family A (Fig. 2) with web and one-sided hub. The teeth were previously removed since they will be cut by hobbing. In the first stage (group A — Fig. 1) after the analysis of the machined part, the machining allowance was defined equal to 2.0 mm, based on the biggest dimension (external diameter 158.36/158.11). This allowance was added to the surfaces with neces- sary subsequent machining (the hole, the external di- ameter, the faces of the hub and the faces of the rim). Then the fillet and corner radii were chosen. The 1° draft angle was defined to help removal of the forging from the die. The height and the wall thickness of the hub define the thickness of the internal web equal to 12.0 mm, large enough to prevent the damage of work- piece and dies. Finally, forging tolerances were defined to all dimensions. Fig. 4 shows the final forging. In group B, after searching for a similar forging, it was chosen to modify an existing forming sequence with four operations: preforming, blocking, finishing and trimming (Fig. 5).Fig. 5. Process sequence. A. . Caporalli et al. / Journal of Materials Processing Technology 80 – 81 (1998) 131 – 135134 Fig. 6. Simulated evolution of the blocking operation. The dimensions of the billet were calculated from the volume of the final forging. The best shape and dimen- sions of the preform were defined from the results of simulations with FEM program ANSYS. As shown in Fig. 6 an incorrect billet shape causes the incomplete filling of the dies in the blocking operation and internal defects in the forged part. Finally (group C) the dies were designed considering the dimensions of the parts in each operation, added with 1% because of the thermal expansion verified at forging temperatures. Table 1 shows process parame- ters chosen from databases and related to material properties, the number of operations and the results of simulation. 5. Conclusions The expert system ADHFD showed good perfor- mance when aiding the design of flashless hot forged gears. The design time is considerably reduced when compared with conventional techniques. Some modifi- cations have to be done in the program to allow the design of asymmetric parts with complex geometric details. Acknowledgements FAPESP (Fundação de Amparo à Pesquisa do Es- tado de São Paulo) and CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nı́vel Superior) spon- sored this project and are gratefully acknowledged. References [1] H. Kudo, Towards net-shape forming, J. Mater. Process. Tech- nol. 22 (1990) 307 – 342. [2] A.G. Mamalis, D.E. Manolakos, A.K. Baldoukas, Simulation of the precision forging of bevel gears using implicit and explicit FE techniques, J. Mater. Proces. Technol. 57 (1996) 164 – 171. [3] H. Kim, T. Altan, K. Sevenler, Computer-aided part and pro- cessing-sequence design in cold forging, J. Mater. Process. Tech- nol. 33 (1992) 57 – 74. [4] M. Tisza, Expert systems for metal forming, J. Mater. Process. Technol. 53 (1995) 423 – 432. Table 1 Process parameters Billet dimensions ¥25×35 mm Initial forging temperature 1100°C 950°CFinal forging temperature Press Eccentric forging press 2000×104 NCapacity Stroke 280 mm Stroking rate 80 min−1 1090×1350 mmTable dimensions Colloidal graphite in oilLubricant A. . Caporalli et al. / Journal of Materials Processing Technology 80 – 81 (1998) 131 – 135 135 [5] G. Yang, K. Osakada, A review of expert systems for process planning of cold forging, Manufact. Rev. 6 (2) (1993) 101 – 113. [6] A. Azushima, M. Kim, A PC-based expert system for forming sequence design, Ann. CIRP 39 (1) (1990) 245 – 248. [7] P. Bariani, E. Benuzzi, W.A. Knight, Computer aided design of multi-stage cold forging processes: load peaks and strain distribution evaluation, Ann. CIRP 36 (1) (1987) 145 – 148. [8] T. Penchev, Development of an expert system for hot die by breakdown of the parts into groups, J. Mater. Process. Tech- nol. 30 (1992) 285 – 287. [9] D. Glynn, G. Lyons, J. Monaghan, Forging sequence design using an expert system, J. Mater. Process. Technol. 55 (1995) 95 – 102. [10] S.K. Biswas, W.A. Knight, Towards an integrated design and production system for hot forging dies, Int. J. Prod. Res. 14 (1) (1976) 23 – 49. [11] E. Doedge, J. Thalemann, F. Weber, Hot forging of precision parts, J. Mater. Process. Technol. 35 (1992) 469 – 481. [12] J. Kusiak, A technique of tool-shape optimization in large scale problems of metal forming, J. Mater. Process. Technol. 57 (1996) 79 – 84. [13] L. Fourment, T. Balan, J.-L. Chenot, Shape optimal design in forging, in: S.-F. Shen and P. Dawson (Eds.), Simulation of Materials Processing: Theory, Methods and Applications, NU- MIFORM’95, Balkema, Rotterdam, 1995, 557 – 562. [14] R.J. Shipley, Precision forging, in: Metals Handbook, ASM Intl., vol. 14, 1988, 150 – 157. .