基于知识和语义网络的加工流程自动推理算法

Auto-reasoning methodology based on knowledge and semantic net in process planning

为了将被加工零件的一系列几何特征转化为加工操作和加工流程,并实现技术上可行和经济上节约的目的,提出一种由设计模型特征框架、加工操作优先关系语义网络和加工流程排序数学模型3部分组成的加工流程设计模型。零件的几何特征被记为U1,U2,…,UN等单位,对于每一个几何特征单位Ui,用两个向量Fi和Pi表示这一几何单位的特征和加工操作,用矩阵PP存储加工流程,矩阵PO存储每一步加工操作的操作对象。设计模型特征框架建立了一个从特征集合到加工操作的映射;然后用一个语义网络来表示加工操作之间的优先关系。通过向量和矩阵在加工操作语义网下的运算得到一个加工流程。由于充分定义了加工操作的属性并有加工语义网络的指导,保证了解的优化和可行。

In order to convert a series of geometric features into machining operations and sequences in a feasible and money-saving order,a Process-planning （PP） model was proposed. The model consisted of three parts： the features framework of design model, the precedent relationship semantic net and the sequenced mathematical model. A part was decomposed into several basic geometrical units, namely, U1 , U2,…,UN. For each unit -Ui, two vectors, namely F, and Pi, were used to represent the features and machining operations of Ui. Matrix PP was used to memorize the process plan sequences. Matrix PO was used to store the objects of machining operations. The features framework of design model established a mapping between manufacturing features set with machining operations. A semantic net called the precedence-relations-net was used to reflect the precedence relationships among the machining operations. Finally a machining process plan sequences were calculated by vectors and the matrixes under the guidance of precedent relation semantic net. For fully definition of the machining process operation and the guidance to precedent relation semantic net, the solution was feasible and optimized.