Key Technology Research For Mechanism Fatigue Life Prediction And Reliability Design

For many years, Life prediction involving fatigue and reliability assessment ofkey structure parts have been the research hot spot for experts and scholars, in manyfields concerning national welfare and the people’s daily life such as automobile,engineering machinery, mining machinery, special equipment, etc. In the research offatigue damage life, computational prediction is usually applied with Miner andCoten-Dolan criteria. In the meanwhile, traditional mechanical models based on thestress-strength principle are used in current researches concerning reliabilitydesign.Nevertheless, due to the applications of experience factors in damage modelsand uniqueness of traditional calculation method, the deviation between the fatiguereliability calculation results and the real value has been resulted. Therefore, how toimprove the accuracy and feasibility of the research about the fatigue prediction andreliability calculation then carry out of the scientific and effective research is veryurgent and necessary.Based on conclusions of related research at home and abroad, also Miner andManson principle, an improved fatigue life prediction model combined with reliabilitycalculation of high accuracy were proposed and improved concerning robustoptimization design in this paper, through fuzzy mathematics and PSO inconsideration of robust optimization design.The main innovation points of the work include：(1) A new fatigue life calculation method has been put forward. Based on theMiner standards, modified fatigue damage model is built by the virtualization toworking stress loading/interval/frequency and the convergence is confirmed throughtheoretical derivations. Compared with simulation analysis results of fatigue life, themain work gives clear evidences to the dependability about the equal classification ofvirtual interval length model. It overcomes the inaccuracy of the calculation due toexperience factors in Miner principle and embodies advances and rationality of themodified model.(2) The research of the improved Manson-Halford fatigue damage model hasbeen made. To improve the accuracy of fatigue life prediction, the research about theexperience data such as inflection point coefficient and power exponent in the modelis carried out. As a result, the influence of the experience data on the fatigue damagecalculation results is obtained and the value span is dicussed. The research shows that the accuracy of the improved Manson-Halford fatigue damage model is advanced butinferior to which of the equal classification of virtual interval length model.(3) The research for reliability and high precision mathematic calculation modeland algorithm has been proposed. According to the existing fatigue damage life model,based on reliability-stress-life principle, interference interval integration method,trapezoidal law and the rules of the Simpson three reliability calculation mathematicalmodel are established based on the random parameter probability density distributionconsidering the fatigue loading work stress and the uncertainty of the strengthparameter, then related algorithm about reliability with high accuracy is studied. Theresearch results manifest that the accuracy and efficiency of the new method has beenimproved effectively in comparation with existing methods.(4) The research of the reliability design based on the fatigue load has been made.With the high accuracy calculation methods of reliability, considering the randomnessof parameters about the fatigue work stress, strength of materials, the fuzzy multipleobjective optimization design mathematical model based on reliability is built and theadaptive values of fuzzy multiple objective function are handled through weightingcoefficients with further effective research. The example of the research resultsembodied in the reliability design elements of multiple objective constraintoptimizations shows that it is simple and practicable to obtain the system designinformation through the research method.(5) The research of reliability robust design based on improved Particle SwarmOptimization algorithm (PSO) has been made. Particle swarm algorithm belongs tothe advanced evolution algorithm. In the foundation based on the reliabilityoptimization design, two improved PSO algorithms about the dynamic accelerateconstant factor and the speed of the adaptive factor are proposed to build the multipleobjective reliability robust optimization design mathematical model for the numericalsolution. The main research work lies in the realization of the robust design aboutmechanical system through the effective combination of modified PSO and reliabilityrobust design.The research achievements of this topic could be well applied in many fieldsconcerning national welfare and the people’s livelihood such as vehicles, engineeringmachinery, mining equipment, etc. Experimental results show the strong engineeringbackground of the research.