The Condition Monitoring And Fault Diagnosis Of ORC System

Organic rankine cycle(ORC) system is waste heat recovery power generation equipment, a large number of domestic and foreign scholars have done a lot meaningful research on it, but the research of condition monitoring and fault diagnosis has not attracted enough attention yet, this paper has completed exploratory research in this aspect. ORC system is a nonlinear, complex coupling dynamic system, so make it running under normal conditions can assure its maximum efficiency, thereby maximizing energy efficiency. Failure will cause decline of the system performance, so condition monitoring and fault diagnosis research has great’significance.This paper first presents a variety of typical faults on organic Rankine cycle, and then uses dynamic kernel principal component analysis (DKPCA) which completes fault diagnosis and fault isolation on six typical faults. The comparison of the corresponding fault diagnosis and detection (FDD) results between DKPCA and dynamic principal component analysis(DPCA) indicates that the condition monitoring and fault diagnosis strategy for nonlinear ORC system based on DKPCA has better performance. Finally, the dynamic maximum entropy principal component analysis (DMaxEnt-PCA) method is proposed to complete the FDD of linear ORC system with non-Gaussian disturbance and the results are compared with DKPCA and DPCA. The comparison results reflect that the method which combines PCA and entropy can process non-Gaussian data effectively and point out that research on combining KPCA and entropy method can process non-linear and non-Gaussian data for further study.