روش خودکار استدلال مبتنی بر دانش در برنامه ریزی عملیات سوراخ ماشینکاری

In process planning, how to obtain an optimal process planning is the essential of computer-aided process planning (CAPP) system. The main goal of CAPP system is to derive manufacturing features and machining operations from a design model and sequence the machining operations of the part in a feasible (by some technological constraints) and effective (by some economical standards) order. In this paper, we construct a process-planning model (PP model) for the hole's machining, which consists of three parts: the features framework, the precedent relation net and the sequencing mathematical model. The features framework makes a mapping from manufacturing features of hole into its machining operations. A semantic net named the precedence-relations-net reflects the precedence relationships among hole's machining-operations. Some vectors and matrixes are employed to construct a mathematical sequencing model. Usually, a hole should be machined in several operation directions, v1,v2,…,vMv1,v2,…,vM. In each operation direction, vivi, there are N l basic geometrical units to be operated, namely, View the MathML sourceU1l,U2l,…,UNl. For each operation direction, vivi, a vector and a matrix are defined to memory the process planning and its operation objects. The mathematical sequencing model will generate an optimal process planning in each operation direction by minimizing the number of tool-changes and decreasing the number of operation steps. Therefore, it can shorten processing times and consume less energy. Finally, two hole-machining examples are employed to illustrate our methodology.

Manufacturing process planning is the method to get the necessary manufacturing process and their acceptable sequence in order to produce a given part in an economical and competitive way (a good way by some standards) [1], [3], [4], [5], [6] and [7]. In order to obtain the process planning, the process planner derives some appropriate information from a design model such as the design profile, accuracy, surface roughness, material and so on. All of those information are defined as manufacturing features. Therefore, most of computer-aided process planning (CAPP) systems uses the concept of manufacturing features to describe a part [6], [8], [9], [10], [13] and [14]. There are two main methods of representing manufacturing features: the superficial approach and the volume approach [1], [8], [9], [13] and [14]. In this regard, manufacturing features link the computer-aided design (CAD) and computer-aided manufacturing (CAM) [3]. But the problem is how to derive the manufacturing features from a design model and build the bridge that can translate the manufacturing features into the machining operations sequence. For several years, this problem has received more and more attention from researchers because it is the essential and the biggest problem for generating an optimal process planning. Some researchers construct a knowledge base to solve this problem [1], [11], [13], [14] and [15]. To construct the knowledge-based process planning (or feature-based process planning), artificial intelligence technique such as the expert system [11], rule-based inference [12], the neural network [2] and the genetic algorithms [16], [17] and [18] are always used. Recently, Park [1] employed the knowledge capturing methodology to construct a knowledge base that consists of three sub-models: the object model, the function model, and the dynamic model. Although he gives us a methodology on process planning, it is not a systematic model of process planning. In this paper, we will employ semantic net and a mathematical model to construct a process-planning model. The disadvantage of our method is how to recognize operation directions and basic geometrical units. This paper is organized four sections. In Section 2, we will construct a process-planning model (PP model) that has three parts: the features framework, the precedence-relations-net (PR-net) and the sequencing mathematical model and can be illustrated by Fig. 1. Firstly, a design model should be decomposed into several units in according with its geometrical profile. By features framework, we can get the manufacturing features and the final operation of each unit. Secondly, a semantic net, namely PR-net, will be build to reflect the precedence relations among machining operations. The final operation of each unit is input PR-net, and then a vector, which consists of operations, is output. The counter-order of this vector is an acceptable sequence for each unit. All of those vectors in the same operation direction are input into the sequencing mathematical model to obtain an acceptable and competitive process plan of this operation direction. In Section 3, two hole-machining examples are employed to illustrate our methodology. Finally, discussion and conclusions are given in Section 4. Full-size image (28 K) Fig. 1. Process-planning model.

It is an important issue on feature-based process planning for the computer-aided process planning [1]. Although there have been many studies on feature-based process planning, researches that use framework, semantic net and mathematical model is rare. To construct the process-planning model, firstly, we use framework to do a mapping from features into operations. Using this framework we can obtain the final operation of each unit of a hole. Then, we employ a semantic net to reflect the precedence relations among hole-operations, namely, the precedence-relations-net of hole. If a single operation is input, it can generate a sequence of operations in a feasible order on technological constraints. Finally, and most importantly, we built a sequencing mathematical model to obtain the process plan. In a part, if some unit volumes share the same machining operation and the same cutting tool, they can be aggregated into one machining operation [1]. But aggregating must be limited by some rules, say, the feasibility of the cutting operation and cutting accessibility. It is obviously that if all of machining operations satisfying the relation and those rules can be aggregated, an efficient and competitive process plan will be obtained. The sequencing mathematical model can generate an optimal process plan in each operation direction by minimizing the number of tool-changes and decreasing steps of process. The disadvantage of our method is how to recognize operation directions and basic geometrical units.