Fault detection and diagnosis of belt weigher using improved DBSCAN and Bayesian Regularized Neural Network
Various faults occurred in the continuous bulk materials weighing equipment (CBMWE) usually lead to more economic loss and waste of human resources inevitably. A new approach based on the improved DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering and Bayesian regularization neural network (BRNN) is proposed for online fault detection and diagnosis of CBMWE--electronic belt weigher (BW). Firstly, in view of the fault data varying with equipment flow, an improved DBSCAN clustering algorithm is developed to realize the online fault detection by extracting the fault data with the clustering analysis of the real-time data. Secondly, BRNN is proposed as a classifier to identify the fault pattern with the extracted fault data. Both the models of online fault detection and diagnosis are realized using MATLAB. Finally, the test result shows that the proposed online fault detection and diagnosis model is able to cope with the online fault detection and diagnosis of BW and also yields great diagnostic accuracy. In general, this approach for online fault detection and diagnosis of BW has a great significance to bulk weighing equipment.