Fatigue Life Prediction For Excavator Working Device And Research On Reliability Assessment Method Under Extreme Small Sample

Excavator is an important mechanical equipment in engineering construction,as a large structural component,it is difficult to have enough samples to implement the full-scale life test on its working device due to high cost for its fatigue life prediction and reliability analysis.The extreme small sample size can hardly meet the classical statistical requirements of traditional methods.In this paper,based on the fatigue test data of excavator working device,the fatigue life of excavator working device is predicted by nominal stress method;three reliability evaluation methods for the extreme small sample case are proposed.The specific contents are as follows:Firstly,based on the test load spectrum,the life of the non-welded and welded parts of the excavator working device is predicted by using the nominal stress method.Then,the fatigue life of excavator working device is predicted by nCode Design-life.The results show the correctness of the work.Secondly,The virtual augmented sample&modified Bootstrap method,GM model&Bootstrap method and BP neural network&Bootstrap method are proposed to evaluate the reliability of excavator working device under extreme small sample.For the GM model&Bootstrap method,the GM(1,1,λ~*)model based on genetic algorithm and the GM(1,1,λ~*)model based on particle swarm optimization are proposed to improve the accuracy of reliability evaluation.Based on the extreme small sample data of excavator working device fatigue test,the reliability of excavator working device is evaluated by using the above three methods,and the evaluation results are compared with that from semi-empirical evaluation,which verifies the correctness and validity of the three evaluation methods proposed.At the same time,the characteristics and applicability of the three evaluation methods are discussed.The proposed methods provide new way to evaluate the reliability of expensive large-scale structural parts.Finally,based on the MATLAB GUI,the“Reliability analysis system of excavator working device based on extreme small sample”is developed.The system includes three modules:virtual augmented sample&modified Bootstrap module,GM model&Bootstrap module and BP neural network&Bootstrap module.The system provides users with a practical reliability analysis system for excavator working device under extreme small sample for engineering application.