基于子PCA模型的故障分离方法及其应用

在应用传统PCA法对火电厂生产过程实施故障检测与分离时,火电厂生产过程中的变量表现出在一定的稳定工况范围内会出现小范围的波动振荡性以及部分变量间存在强相关性,这就导致了传统的贡献图法容易得出错误的故障分离信息。针对贡献图的缺点和火电厂的生产特点,提出一种子PCA模型的故障分离法,此法利用Q统计量的故障检测能力和统计规律来实现故障变量的精确定位。最后使用此法仿真了火电厂主汽压系统的故障分离过程,仿真结果表明使用子PCA模型法进行故障分离的优越性。

Application of fault isolation in thermal power plants based on subPCA

To define a new principal component analysis (PCA) fault isolation that adapts to the control process of thermal power plant, the abilities of two basic detection statistics, Hotelling T2 and Q are analyzed. According to the limits of the conventional fault isolation method, a new strategy for sensor fault isolation is proposed. Based on long-term data sampling from boiler process in a thermal power plant, the proposed method is applied to sensor fault detection and isolation. Simulation result shows the effectiveness of this method.