基于STEP和改进神经网络的STEP-NC制造特征识别方法

特征识别是实施STEP-NC重要的一步，也是实现开放式、智能化和网络化STEP-NC数控系统的关键。本文提出了一种基于STEP和改进神经网络的STEP-NC制造特征识别方法。该方法首先在对STEP AP203中性文件进行几何拓扑信息提取后，基于边的凹凸性判断构建了零件最小子图。然后，将混沌算法、遗传算法与BP神经网络算法有机相结合提出了改进的BP神经网络。最后，通过将获得的零件模型最小子图信息数据输入到改进的BP神经网络，实现了对STEP-NC制造特征高效精准地识别。通过实例验证了该方法的有效性和可行性。

Method for STEP-NC manufacturing feature recognition based on STEP and improved neural network

Feature recognition is an important step to implement STEP-NC theory and a key to realize the open, intelligent and networked STEP-NC CNC system. A feature recognition method based on STEP and improved neural network for STEP-NC manufacturing features is presented in this paper. This method first extracts the geometric and topological information from the STEP AP203 file of a part, and builds the minimum subgraph of the part based on the judgment of the convexity of edges. Then, an improved BP neural network is proposed by combining the chaos algorithm, the genetic algorithm, and the BP neural network algorithm. Finally, by inputting the information data from the minimum subgraph of the part to the improved BP neural network, efficient and accurate feature recognition for STEP-NC manufacturing features in the part is achieved. The validity and feasibility of the proposed method are verified by two case studies.