Research On Diesel Engine Health Management System Based On Rough Set And Fuzzy Neural Network

Diesel engine,one of the most widely applied power machines in modern society,has played an important role for some fields such as aerospace,large-scale water and hydropower systems as well as shipping.Diesel engine is vital to power takeoff resources,especially in the shipping power system,and whose situation directly determines the system and personal security of staff involved.With the rapid growth of modern science and technology,shipping diesel engine unceasingly develops towards the direction of being tremendous,intelligent integration and complexity,so that the capability of diesel engine is improved and enhanced greatly.But meanwhile the failure rate is rising as well,which makes the classification and characteristics of its failure mechanism more complicated.In consequence,people pay increasing attention on the safe operation and maintenance support as well as other problems of shipping diesel engine system.Aimed at these problems and for promoting the operation efficiency,this paper is designed and researched to analyze typical failure diagnosis and health management system of the diesel engine mainly focusing on their failure features extraction,failure diagnosis efficiency and accuracy.The contents include:(1)To simplify vibration signal data acquisition system,this paper analyzes the stability and repeatability of the acquisition system and proves the practicability of unsteady vibration signal acquisition system by calculating relative standard deviation of multiple test data’s six time-domain indicators.(2)To propose a methodology based on combining EEMD with SVD that can be applied to feature extraction of the initial vibration signal of diesel engine by EEMD-SVD method.(3)To combine the rough set algorithm with the fuzzy neural network,a rough set and fuzzy neural network structure system is established to apply to the fault diagnosis of diesel engine,which achieves the complementary advantages of the two,and effectively improves the efficiency and accuracy of fault diagnosis.(4)Based on the aforementioned research methods,they are embedded in the health management analysis system as sub-modules to realize the function of fault diagnosis and health management,and to verify the effectiveness of the methods.