生产过程轻微和早期故障智能诊断的一种新方法

A NEW INTELLIGENT APPROACH TO DIAGNOSE THE SLIGHT AND INCIPIENT FAULT DURING THE PRODUCTION PROCESS

为实现生产过程轻微故障和早期故障的正确识别与及时诊断,提出了一种结合征兆缩放技术和模糊知识表达的故障综合智能诊断的新方法。在模糊模式识别、神经网络等故障智能诊断方法的基础上,利用征兆缩放技术可有效降低故障知识库的复杂程度,实现生产过程微弱故障的正确检测、诊断。为提高智能诊断模型对早期故障的识别能力并保证诊断结果的可靠性和稳定性,综合运用了趋势型和语义型两种故障征兆,探讨了两种征兆的模糊表达及合成计算方法。以火电机组高加系统轻微和早期故障诊断为例,验证了本文方法的有效性。

In order to realize slight and incipient faultdiagnosis correctly and timely for industrial process, anintelligent fault diagnosis method is put forward by jointapplication of symptom zoom technology and fuzzy symptom representation. The complexity of the fault knowledge librarycan be effectively reduced and the slight fault can be correctly diagnosed with the symptom zoom technology, together withan intelligent fault diagnosis method, such as fault fuzzy pattern recognition or ANN method. Two types of fault symptoms,trend symptom and semantic symptom, are adopted to increase the ability to diagnose the incipient fault timely and reliably. Both symptoms are represented with fuzzy math method andtheir integrated calculations are discussed. Some examples are given to verify the effectiveness of the method.