Hydrogen leakage faults classification diagnosis based on data-driven in hydrogen supply system of fuel cell trucks

Abstract

Listen

Translate article

Explosion and deflagration may occur when a fuel cell truck in driving has hydrogen leakage faults, and the complex and changeable operating conditions of the vehicle are incredibly unfavorable for fault diagnosis. Therefore, to provide real-time diagnosis on the hydrogen supply system of driving truck, a diagnosis method of hydrogen leakage faults classification based on data-driven in this paper. The faults are divided into 3 levels using the K-means algorithm, and 4 features are selected as the hydrogen leakage faults diagnosis indicators. The Principal Component Analysis (PCA) method is used to reduce the dimensions of the data, while Support Vector Machine (SVM) is used for hydrogen leakage faults diagnosis. Meanwhile, the kernel parameter σ and the penalty factor C of SVM are optimized by Sparrow Search Algorithm (SSA) to improve diagnostic accuracy. The results show that the diagnostic accuracy of the SSA-PCA-SVM method proposed in this paper is higher than that of the ordinary SVM and PCA-SVM diagnosis methods when σ = 9.79, C = 7.89, which is over 97 %.