论文部分内容阅读
提出了一种以线性递推学习为基础的分类-重构神经网络。网络具有学习算法简单、速度快、学习与分类并行,以及可自动积累知识等基本功能,尤其适用于生产过程的早期故障诊断一类实时系统。给出了化工过程早期故障诊断的应用实例,研究结果证明了网络的有效性。
A classification-reconstruction neural network based on linear recursive learning is proposed. The network has such basic functions as simple learning algorithm, high speed, parallel learning and classification, and automatic accumulation of knowledge. It is especially suitable for a kind of real-time system of early fault diagnosis in the production process. The application examples of early fault diagnosis in chemical process are given. The research results prove the effectiveness of the network.