| During the vehicle design, the load analysis based on the users is a very crucial step. Therefore, when the fatigue cumulative damage, fatigue limit, and sudden fracture failure are considered for designing mechanical components, the statistical analysis of load in amplitude domain and time domain should be conducted to achieve the correct balance between safety design and over design. In order to achieve the indoor load simulation test, proving ground durability test and the fatigue life prediction, the extreme load determination and load extrapolation for quickly getting a certain life cycle load spectrum is the current research focus in the field of engineering vehicles. However, the maximum load contained in the whole life cycle may not appear in the measured load, so how to effectively use extreme information of load sample and determine a reliable method of extreme load determination is one of the research contents. In load extrapolation, to consider the influence of extreme load cycles on fatigue life, an improved extreme extrapolation method is presented to make up the disadvantages of existing methods. The research in this paper is supported by the national natural fund project(No: 51375202), combining Markov chains Monte Carlo simulation(MCMC) with Multi-criteria Decision Making(MCDM) technology, the Extreme Value Theory(EVT) is applied to the extreme load determination and load extrapolation, the main research contents and conclusions analysis are as follows:1. After summarizing the current status of load spectrum compiling and extreme value theory, etc., the Generalized Extreme Value(GEV) distribution based on BMM and Generalized Pareto Distribution(GPD) based on POT were compared from the extreme sample selection, fitting distribution function selection, parameter estimation method, the key variables determining method, and the application. It can provide reference for subsequent researches.2. Due to the importance of load reconstruction in the reliability test, indoor load simulation test, and fatigue life analysis, MCMC theory was presented and the length of load simulation was discussed through statistical properties. Then the simulated load was extrapolated by POT. In this method, the load randomness between the thresholds and the numbers of the exceedances increased after the simulation, and the extreme distribution model of the extreme loads was more accurate, eventually the extrapolation loading-time and circulating cumulative frequency curve were obtained.3. Combining the load characteristics of engineering vehicles with the advantages of BMM model for seasonal data, an improved extreme sample method based on the comprehensive inspection index was proposed to avoid wasting extreme information. Meanwhile, the reliability and stability of the proposed method was validated by comparing block division results of the same length load at the same point.4. In order to verify the effectiveness of the proposed method, the measured load at the axle shaft under the condition of spading primary soil was illustrated as an example. Taking the spading section as data sample of BMM to fit GEV distribution, and analyzing qualitative the fitting effect by graphic method, the rationality of the extreme load determination was evaluated. Based on MATLAB, a graphical interface software was developed to provide convenience for engineers in load spectrum compiling and realize the standardization of load processing.To provide basic data for the static strength design, indoor load simulation test and fatigue life analysis, a certain life cycle load spectrum should be obtained quickly. This article summarized the key technologies of load spectrum compiling, emphatically discussed the extreme load determination and load extrapolation methods, and finally can realize arbitrary length of load simulation, extrapolation and the determination of extreme load. |
PDF Full Text Request