Research Of Speciyfing The Extreme Load And Its Application In Compiling Load Spectra

Load spectrum is the foundation of fatigue test on components, its authenticityaffects the accuracy of fatigue life prediction directly. At present, the development ofload spectrum compiling methods in some areas, such as cars, agricultural machineryfield and wind power generation and so on, has been quite mature, and thecorresponding industry standard and special processing software have been formed.However, in engineering machinery field, the load spectrum compiling work startslate and still takes imitation and reference at present. To research the predictiontheories of safety durable service performance of the engineering vehicles, combinedby the National Natural Science Foundation of China (NO:50805065) and on thebasis of referencing the domestic and foreign relevant research, the key technologiesof load spectrum compiling including the selection of distribution functions and thedetermination method of extreme loads have been researched in present paper, themain works and research contents are listed as follows:(1) Exploring general methods of establishing the mixture distribution model.According to the operating characteristics and load distribution characteristics of eachoperating section of wheel loader, the method to establish the best statistical functionsis put forward based on the mixture distribution theory. The evaluate matrix ofmulti-alternative and multi-objective properties is established, the weight coefficientsof different criteria are obtained by the MTOPSIS optimization method. Then themulti-criteria decision technology is used to deal with the multi-objective propertiesand number of the basic function of mixed distribution model is determined with theoptimization idea.(2) Establishing a statistical model of extreme loads and giving out theapproximate expression of extrapolated extreme loads. Considering the randomness ofthe load samples and connecting with the statistical theory, the statistical model ofextreme loads is established, the comparative analysis of the extrapolation risk valueunder different distribution function is conducted, and the best fitting function isselected. Based on the POT theory, the experience calculation formula of extreme loads is determined using the reverse educing method on condition that the thresholdis known.(3) Proposing a new method to determine the threshold and parameters. Thetheoretical basis and application scope of the method to determine the threshold underdifferent fields are summarized, the influence law among high low threshold,deviation and variance is analyzed, and solving process of the threshold is deducedbased on the self-help sampling method (Bootstrap). In addition, in view of therandomness, unrepeatable and the difficulty to predict for engineering vehicle load,parameters of the statistical distribution function are regarded as random variablesusing Bayesian theory, and original parameters are updated with measured data.Finally, through the comparison of fitting effect before and after correction, therationality of the calculation results considering the Bayesian theory is verified.(4) Taking the gearbox output flange of wheel loader as an example to analyzethe distribution characteristic of extreme loads and verifying the validity and accuracyof theory and methods proposed in the present paper. The comparative analysis ofdiscrete degree of extreme distribution, threshold differences and the size of extremeloads under three different working conditions is conducted, and the cause of thisphenomenon is analyzed carefully.A complete process of specifying the extreme load of engineering vehicle incompiling the load psectra was proposed in this paper, including the fundamentaltheory and the detailed realization procedure.This project will provide basic theory toaccomplish the work on establishing the standard of fast detection to the engineeringvehicles, which will present the important academic significance and actualapplication value.