Research On Assessment Method Of Diesel Enginr Health Condition Based On Multi-parameter Fusion

Ensuring the healthy and stable development of nuclear power is a major strategy for China’s energy industry.As a 1E-class backup power supply for nuclear power plants,emergency diesel generators should be quickly started and output rated voltage and power within 10 seconds after an emergency.This puts extremely high demands on the performance of the emergency diesel engine.However,at present,there is still a lack of research on the health condition assessment method for the start-up process of emergency diesel engines at home and abroad.Took an emergency diesel engine of nuclear power plant as an example,this paper summarizes the processing method of non-stationary signal of emergency diesel engine,puts forward the principle of health class division of emergency diesel engine and the principle of selecting health assessment indicators.The vibration and speed characteristics and related parameters of the emergency diesel engine are analyzed.Seven health assessment indicators such as high-pressure gas shock vibration acceleration were selected.Finally,a method for selecting the assessment indicators of emergency diesel engine health condition is proposed,ditto the health condition of the emergency diesel engine single cylinder and the overall machine was proposed.The assessment method was carried out on-site verification of the three methods.The specific research contents are as follows:(1)Using the Sigmoid function to improve the whitening weight function of gray clustering,calculate the weight ratio of the indicators by entropy weight method,and finally calculate the gray clustering coefficient of each cylinder health condition of the emergency diesel engine.The research on the assessment method of single cylinder health condition of emergency diesel engine based on improved grey clustering method and entropy weight method was completed,and the field verification was carried out.(2)The cloud generator is built to generate the diesel engine normal cloud model,and the weight ratio calculated by the entropy weight method and the improved analytic hierarchy process is combined to calculate the integrated cloud center of gravity vector of the emergency diesel engine in the ideal state and the state to be assessed.The weighted deviation degree is obtained,and the deviation degree is input into the cloud generator to obtain the cloud model assessment result.The research on the health condition assessment method of the overall emergency diesel engine based on the cloud center of gravity assessment method and the combined weight method is completed,and the field verification was carried out.(3)Constructing the emergency condition indicators of the emergency diesel engine and the membership function of the health condition level,using the idea of cooperative game theory to combine the fuzzy set value method,the improved analytic hierarchy process and the entropy weight method.Using D-S evidence theory to fuse the basic credibility distribution results of emergency diesel engine health condition,according to the principle of maximum membership degree,the health assessment result of the overall machine was obtained.The research on the health assessment method of the overall emergency diesel engine based on D-S evidence theory and cooperative game theory was completed,and the field verification was carried out.After field verification,the final results of the three assessment methods are consistent,which are consistent with the actual state of the emergency diesel engine site.