电站锅炉补给水处理流程全工况故障检测方法研究

Study of the Method for Detecting Faults in the Makeup Water and Feedwater Treatment Flow Path of a Utility Boiler Under All Operating Conditions

复杂工业流程存在多稳态工况切换及切换的过渡过程,导致传统PCA(principal component analysis)故障检测方法易于误报故障。本研究提出了基于稳态因子的过渡过程判别方法及基于相似因子的工况自适应匹配方法,将其融入PCA构建了新的故障检测方法,且将该方法用于电站锅炉补给水处理流程的故障检测。以该流程的全工况运行数据对算法进行了验证,结果表明:此方法能有效消除过渡过程的影响,并能通过工况匹配提高故障检测性能且减少故障误报,可以实现水处理流程的全工况故障检测。

There exists a multi-steady-state operating condition switching-over and their transient process in a complex industrial flow path,causing the traditional principal component analysis and fault detection method easily to mistakenly alarm a fault. As a result,the authors proposed a transient process identification method based on the steady-state factors and an operating condition self-adaptive matching method based on the similarity factors and incorporated them into the principal component analytic method to form a new fault detection method. The method in question was used for detecting faults in the makeup and feedwater treatment flow path of a utility boiler and verified by using the operating data of the flow path under all operating conditions. It has been found that the method under discussion can effectively eliminate the influence of the transient process and enhance the fault detection performance and reduce the number of faults mistakenly alarmed through a matching of the operating conditions,thus accomplishingthe fault detection of the water treatment flow path under all operating conditions.