基于人工智能方法的复杂过程故障诊断技术  

Fault Diagnosis Technologies Based on Artificial Intelligence for Complex Process

作　　者：桂卫华[1] 刘晓颖[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《控制工程》2002年第4期1-6,共6页Control Engineering of China

摘　　要：由于复杂过程因素多 ,波动大 ,反应机理复杂 ,无法建立精确的数学模型 ,传统的故障诊断方法很难取得令人满意的结果。针对复杂过程的特点 ,利用智能技术无需建立对象精确模型的优势 ,研究适合复杂过程实现的基于人工智能方法的故障诊断技术。并对构造智能诊断系统所需要解决的机器学习技术从知识获取、深浅知识表示方法和规则更新方面进行了分析。Owing to many factors, great fluctuations and intricate reactive mechanism of complex process, it is hard to get the satisfied results by the traditional fault diagnosis methods. In allusion to the characteristics of complex process,without object precise model, this paper studies the fault diagnosis technologies based on artificial intelligence, and analyses the pivotal technologies for construction of intelligent diagnosis system from knowledge acquisition, integrated model of shallow and deep knowledge and updating rules. In the end, future research direction is predicted about fault diagnosis technologies for complex process based on artificial intelligence, and problems that should be met are pointed out and discussed.

关 键 词：人工智能 复杂过程 故障诊断 机器学习 神经网络 数学模型 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TP18[自动化与计算机技术—控制科学与工程]