【正文】 opriate operations features have been assigned to stamping features of a product based on featureoperation criteria, parameters of the stamping feature and their correlativity. This assignment is a decisionmaking activity using a set of rules with a decisionmaking tree and modelbased reasoning methods. With knowledge between operations, such as operations order constraint (doafter) and operations bination constraint, process paths have been improved based on relevant intelligent reasoning methods. Based on the relationships (preferredto) between processes and machines/dies, the structure of die and machine for each process can be identified, since the process route has been determined. In this stamping process planning, the procedure and information have been controlled by a process control structure that is associative and integrated. 1 Introduction Recently, research on the puteraided process planning (CAPP) system for sheet metal has been widely reported. Park et al. [1] constructed an automated process planning system for ax symmetric deep drawing products. Tessa [2] and Kang and Park [3] presented a group technology and modularity to construct a CAPP system for process sequence design in an expert system for nonax symmetric deep drawing products with elliptical shape. Gao et al. [4] developed an advanced software toolset used for the automation of sheet metal fabrication planning for aircraft ponents. Zussman and Horsch [5] proposed a motion planning approach for robotassisted multiplebent parts based on Cspace and a potential field. Wang and Bourne [6] proposed an automatic process planning system with the features well investigated and the production plans researched with nearminimum manufacturing costs. De Vin et al. [7 , 8] developed a sheetmetal CAPP system called PARTS, which integrates cutting, nesting, bending and welding processes for bending sequences. Streppel et al. [9] showed the ambiguity of conventional tolerances and presented a method which replaces conventional tolerances with geometrical tolerances for process planning in small batch sheet metal part manufacturing. Amoral et al. [10] proposed a method which generated feasible bending sequences of a sheet metal part handled by a robot, and discussed the determination of the best grasping positions and repositions. Aomura and Koguchi [11] pro posed a method to generate bending sequences of a sheet metal part handled by a robot. Liao and Wang [12] proposed an evolutionary pathplanning approach for robotassisted handling of sheet metal parts in bending. Lutters et al. [13] developed a generic architecture for puter aided process planning based on information management for sheet metal manufacturing in a small batch part environment. Kumar and Rajotia [14] had proposed a method of scheduling and its integration with CAPP, so that online process plans can be generated taking into account the availability of machines and alternative routes. The contents above are mainly for process parameter calculating, pathplanning and some sketch map of workpieces for specific types of sheet metal, such as axismetric and nonax symmetric deep drawings, plex bandings and sheerings, and so on. The automobile body panel is one kind of sheet metal part, which is plicated in shape, with groups of free form surfaces, a large figure in size and is always manufactured by stamping processes. Automobile panels can be considered as a bination of some mon stamping, such as irregular drawing, flanging/bending, trimming and piercing, etc. The process planning of these panels is more plicated than mon sheet metal stamping, which is generally dependent on engineers experience to plete. It is believed that the process path plan for automobile panels is requisite and acquirable. In essence, the stamping process path for automobile panels is to determine the necessary forming processes and their sequences in order to produce a particular part economically and petitively. Process paths generation is a decisionmaking process. Decisions on stamping operations for a particular feature have to be formed on various independent conditions such as which operation should be performed with which die and tools and under what forming parameters. A CAPP system for these should be an integrated environment to deal with knowledge to reduce the dependence on engineers or experts, and realize the process planning with scientism. Thus, knowledge based engineering (KBE) is applied to advance the stamping CAPP system for automobile panels, and even to improve the petitiveness for the automobile industry. This paper is particularly concerned with the construction involved with developing a CAPP system based on KBE. 2 KBE in CAPP system for stamping KBE Knowledge based engineering (KBE) is one innovative method of artificial intelligence for engineering design developed in the 1980s. So far, there is no generally accepted and mature definition for KBE. However, it is recognized that KBE is an intelligent method to resolve engineering problems, which can realize inheritance, integration, innovation and management of domain expert knowledge through the drive, multiplication and application of knowledge. A knowledgebased system (KBS) is one that captures the expertise of individuals within a particular field (the ―domain‖), and incorporates it and makes it available within a puterized application [15]. The level of plexity of the tasks performed by such a system can vary greatly. However, it can generally be said that while a domain expert would find them routine, they would be outside the capabilities of a person unfamiliar with the domain [16]. KBE provides an open architecture and reuse ability of experience and knowledge, which can deal with multi domain and multiexpression of knowledge, and can form an integrat