Fig. 5. Stamping feature level of the blackboard for the sample stamped metal part
Then the system opportunistically consults with the perse mapping KSs to transform the stamping features into a set of stamping operations that form the third level of hierarchy on the blackboard, i.e., stamping operation level. The user may also add or revise the stamping operation information interactively by means of interactive tool residing in the blackboard environment.
After further consulting with the staging KS, the stamping operations can be sequenced through a graph-based stamping process plan that forms the fourth level of hierarchy on the blackboard, i.e., stamping process plan level (Fig. 6). In this user interface, the right-hand window shows the graph-based stamping process plan, in which different stamping operations are staged in a same station or sequentially in different station. Figure 6 shows eight piercing operations, two bending operations, three embossing operations, two extruding operations, five notching operations, and two cut-off operations. The left-hand window shows detailed information about a selected stamping operation.
Fig. 6. Stamping process plan level of the blackboard for the sample stamped metal part
Figure 7 shows the corresponding 2-D strip layout solution generated by the computer. Of course, the user can always override the computer-generated strip layout by modifying the default solution with interactive tools residing in the CAD system.
Fig. 7. 2-D strip layout solution for the sample stamped metal part
The prototype system currently holds 120 perse KS rules. The time to design an optimal strip layout for the above sample part was less than 30 minutes running on a personal computer with 2.4 GHz Pentium CPU and 512 MB of memory. This is significantly less than industrial practice, which is around 4–5 days to design strip layouts for parts of equivalent complexity.
5 Conclusions
The paper presented a knowledge-based blackboard framework for stamping process planning. The knowledge architecture consists of specialist KSs controlled by a strategy KS. A prototype system has been implemented in CLIPS interfaced with Solid Edge CAD system through C++. The system contains 120 KS rules. The system has been partially validated. The power of the current system is limited by its KS rule set. Future work is aimed at increasing the rule set for the specialist KSs and including new strategy KS to provide more powerful reasoning strategies. In addition, new graph theoretic strategy rules are being developed to solve specific sub-problems, e.g., a coloring algorithm for clustering various stamping operations.
References
[1] Giarratano J, Riley G (1998). Expert systems: principles and programming, 3rd edn. PWS, Boston．
[2] Fogg B, Jaimeson B (1975).The influencing factors in optimizing press tool die layouts and a solution using computer aids. CIRP Ann 24:429–434．
[3] Nakaham S, Toshio K, Tamura K, Asuke F, Soda C, Nakamura T (1978) .Computer aided progressive die design. Proceedings of the 19th Machine Tool Design and Research Conference. Macmillan, London, 1978, pp 1–12．
[4] Murakami H, Shirai K, Yamada O, Isoda K (1980). A CAD system for progressive dies. Proceedings of the 21th Machine Tool Design and Research Conference, Macmillan, London, 1980, pp 587–591．
[5] Bergstrom K, Kivivuori S, Osenius S, Korhonen A (1988). Computer aided design of progressive die. In: Chersot JL, Ohate E (eds) Modelling of metal forming processes. Kluwer, New York, pp 155–162．
[6] Duffey MR, Sun Q (1991). Knowledge-based design of progressive stamping dies. In: Venkatesh VC, McGeough JA (eds) Computer-aided production engineering. Elsevier, Amsterdam, pp 241–247．
[7] Cheok BT, Foong KY, Nee AYC (1996). An intelligent planning aid for the design of progressive dies. Proc Inst Mech Eng B J Eng Manuf 210(B1):25–35 级进模设计冲压工艺英文文献和中文翻译(7):http://www.751com.cn/fanyi/lunwen_41421.html