论文部分内容阅读
精确故障诊断是预测与健康管理的一个重要部分。它能避免事故的发生,延长设备使用寿命,还能降低设备维修保养费用。本文研究涡轮发动机的故障诊断。由于发动机工作在高温、高压、高转速的严峻环境中,不能安装过多传感器,因此我们无法获得足够多的传感器数据,以至于采用现有算法不能进行精确的潜在故障诊断。本文针对复杂环境下有限传感器数据的发动机故障诊断问题,提出了一种基于信息熵的深度置信网络方法。首先介绍了几种信息熵,并基于单信号熵提出了联合复杂信息熵。其次,分析了深度置信网络的构成,提出了基于信息熵的深度置信网络方法。验证实验表明,与现有的机器学习算法比较,该方法的诊断精度大大提高。
Accurate troubleshooting is an important part of forecasting and health management. It can avoid accidents, extend equipment life, but also reduce equipment maintenance costs. This article studies turbine engine fault diagnosis. Since the engine can not operate with too many sensors in severe conditions of high temperature, high pressure and high rotational speed, we can not get enough sensor data so that the existing algorithms can not accurately diagnose potential faults. In this paper, aiming at the problem of engine fault diagnosis based on finite sensor data in complex environment, a deep confidence network method based on information entropy is proposed. Firstly, several kinds of information entropy are introduced, and the joint complex information entropy is proposed based on single signal entropy. Secondly, the paper analyzes the composition of deep belief networks and proposes a deep belief network approach based on information entropy. Verification experiments show that compared with the existing machine learning algorithms, the diagnostic accuracy of the method is greatly improved.