Fault Detection Method via k-Nearest Neighbor Normalization and Weight Local Outlier Factor for Circulating Fluidized Bed Boiler with Multimode Process

In modern complex industrial processes, mode changes cause unplanned shutdowns, potentially shortening the lifespan of key equipment and incurring significant maintenance costs. To avoid this problem, a method that can detect fault operating in various modes is required. Therefore, we propose novel detection uses k-nearest neighbor normalization-based weight local outlier factor (WLOF). The proposed performs normalization using neighbors to consider possible normal data WLOF used for detection. contrast statistical methods, such as principal component analysis (PCA) independent (ICA), (LOF) density neighbors. However, because LOF significantly affected by distance between its neighbors, multiplied proportionally each improve performance LOF. efficiency was evaluated multimode numerical case circulating fluidized bed boiler. experimental results show outperforms conventional PCA, kernel PCA (KPCA), (kNN), particular, improved accuracy 20% compared with methods. be applied real process multiple modes.