时域分析的多数据融合故障诊断方法

Multi-data Fusion Faulty Diagnosis Method of Time Domain Analysis

介绍了时域分析中的有纲量参数和无纲量参数,构建了多数据融合故障诊断系统。系统包含数据级、特征级和决策级融合诊断模块,并行BP神经网络作为特征级诊断模块,既克服了BP网络的缺点,又解决了D-S证据理论难以获得BPA的弊端。对液压泵进行应用试验,结果表明,该诊断系统充分利用了多数据的冗余信息,提高了诊断的可靠性和准确性。

This paper introduced dimensional parameters and nondimensional parameters in time domain analysis, built multi-data fusion fault diagnosis system, which included data level fusion module, feature level fusion module and decision level data fusion module, adopted the neuronal networks to conquer the shortcoming of neural network and to solve the abuse of evidence theory which is difficult to gain BPA. The fault diagnosis system was tested on the hydraulic pump, and it was shown from the result of the experiment： the system used fully redundant complementary information of multi-dam and improved the reliability and exact ratio of fault diagnosis.