摘要
智能故障诊断技术为保障工程技术系统的可靠性和安全性开辟了新的途径,随着系统设备和功能的日益复杂化,发生故障的机率以及由此带来的损失越来越大,现有单一、固定的故障诊断方法却难以满足复杂系统诊断的全部要求。该文针对复杂系统故障现象的特点,分析了现有基于规则、基于结构和行为、案例、模糊逻辑、神经网络及其集成知识诊断技术的各自特点和局限性,指出了机器学习对于当前复杂系统智能故障诊断发展的重要性,有利于改变现有单一、固定的故障诊断思维,并对未来的主要发展方向进行了一些探讨。
The technology of intelligence fault diagnosis develops a new way for the reliability and safety of industrial systems, with the gradual complication of the equipment and function of system, which leads to more faults and loss, but , it can no longer meet all requirements of diagnosing fault complex systems by existing solitary and static means. According to the characteristics of complex systems, this paper analyses the characteristics and limitations of the existing fault diagnosis and integrated approaches based on rule, structure and behavior, case, fuzzy logic and neural network, points out the importance of machine learning to intelligent diagnosis system ,which is beneficial to change the mind in fault diagnosis, and discusses the main research directions of intelligent diagnosis system in the future.
出处
《计算机仿真》
CSCD
2004年第10期4-7,11,共5页
Computer Simulation
