基于数据驱动的风机组网系统故障诊断

Data Driven Fault Diagnosis of a Fan Network System

风机组网系统的大型化、复杂化、综合化特性增加了其故障诊断的难度。基于主元分析、KullbackLeibler散度、多分辨率奇异值分解的数据驱动方法,通过故障检测、故障定位、故障分析三个步骤,实现了对风机组网系统中增益性故障以及周期性冲击故障的诊断与分析。利用主元分析与Kullback-Leibler散度结合,实现系统故障检测;利用Z分解和Kullback-Leibler散度来故障定位,将故障定位到具体的风机;利用多分辨率奇异值分解来分析故障种类。仿真结果证明了该方法的有效性。

The fault diagnosis of large- scale, complex and integrated fan network systems is performed by three data driven methods based on principal component analysis, Kullback- Leibler divergence and multi- resolution singular value decomposition. Fault detection, fault location and fault analysis has been realized in three steps, i . e., gain fault, periodic impact fault diagnosis and analysis of fan network system. The fault detection system uses the combination of principal component analysis and Kullback-Leibler divergence, the fault location, i . e., the fault of a specific fan is determined using Z- decomposition and Kullback-Leibler divergence, and multi-resolution singular value decomposition analyses the fault type. The simulation results prove the effectiveness of these methods.