状态检修专家系统中神经网络的应用研究

Research of Neutral Network in Condition-based Maintenance Expert System

介绍了设备状态检修专家系统的结构，提出了一种基于模块化的神经网络的系统结构和学习算法。它通过用分解判定子模块对输入向量的适当分区域、由合成子网将各区域的结果合成，实现了复杂任务的自动分解判定和模块化训练策略。研究表明该结构和算法是可行的、有效的，它具有并行性高、对新增样本易于学习等特点。

This article introduces the structure of condition-based maintenance expert system, and a structure and algorithm about modular neural network is proposed. By means of decomposition decision sub-modular, it can automatically divide a complex task into a series of sub tasks which are processed by their corresponding subtask nets. The combination strategy can arrive at a solution to a given task. The research shows that the structure and algorithm is practicable and effective. It has higher parallelism and is easy for training new samples.