Research On The Establishment And Application Of Engine Reliability Sample Set

Reliability design is an important step in modern engine reliability research.However,in reliability design,due to the lack of systematic data,the results are often far from the actual engineering.Aiming at this problem,the reliability sample set of engine and subsystem based on field data is firstly proposed and created,which provides data support for reliability design of engine.By sorting out and analyzing nearly 4000 engine field test data,mining them deeply,analyzing the failure situation of different types of engines in different engineering backgrounds and the utilization rate of different types of engines in different engineering backgrounds systematically,the analysis and evaluation of failure situation and utilization rate analysis are regarded as the important content of establishing reliability sample set,and preliminary construction is made.The basic structure of the sample set of engine reliability is established.The three modules of fault analysis,reliability evaluation index and reliability model of the sample set are established.Taking reliability design as an example,the data support and reference function of reliability sample set are elaborated in detail.In order to improve and supplement the content of reliability index module of reliability sample set,the mileage distribution of engine is identified and analyzed by quality management software.By comparing and analyzing the goodness of fit of various distributions,the Weibull distribution with good fitting results and convenient parameter estimation is selected.Based on this,the probability model based on mileage distribution is established,and the model A and B engines in traction locomotives and traction locomotives are calculated.The average fault-free working mileage in the application of road transport vehicle and construction machinery is analyzed and studied.The changing trend of maintenance-free working period with mileage in the application of A and B engines in traction locomotive,road transport vehicle and construction machinery is analyzed and studied.The maintenance-free working period in different mileage stages of engine is obtained,and the average fault-free working period and maintenance-free working period are regarded as the content of reliability index module.Based on the sample set of engine reliability,a multi-factor model of engine reliability is constructed,and the effects of engine type,engineering background,mileage,service time and usage rate on the cumulative failure rate of engine are analyzed.Through manual data sorting and machine data cleaning,the linear relationship between continuous variables(mileage,time and usage rate)in influencing factors is recognized.Based on this,a suitable machine learning algorithm,ridge regression,is selected to establish engine reliability model and optimize parameters.Then,the high fitting accuracy of the reliability model after parameter optimization in various engine models and engineering background and in each time interval and mileage interval are verified.This model can be used to analyze the change of cumulative failure rate of engine,and it can be used as an important part of the reliability model module of sample set as well as an important reference for future reliability design and upgrading work.