Weak Fault Diagnosis of Ship Fuel Injectors Based on Adaptive CYCBD and RCHBDE

Accurately diagnosing weak faults in marine fuel injectors is essential for minimizing diesel engine testing and maintenance expenses. Consequently, this paper introduces an innovative weak fault diagnosis method for common rail injectors, integrating the Improved Sparrow Search Algorithm (ISSA) optimized maximum second-order cyclicality blind decoding (CYCBD) with Refine Composite Hierarchy Bidirectional Diversity Entropy (RCHBDE). Initially, we propose the ISSA algorithm to find the optimal filter length for CYCBD under different states of fuel injector signals, and apply the optimal CYCBD filter to denoise and filter the fuel injector source signal. To overcome the limitations of Multiscale Bidirectional Diversity Entropy (MBDE) in capturing high-frequency component information, we introduce RCHBDE for extracting fault features from the filtered signals. This proposed method effectively detects weak fault conditions in common rail injectors and, compared to existing methods, demonstrates superior fault recognition rates.