Accurate prediction of the fatigue life of recycled concrete(RC)is important for the application of RC in pavements and bridges and other projects.Based on the grey Markov model(MGM),the fatigue life data in the original series were continuously updated by using the metabolic theory,and combined with the particle swarm algorithm to optimize the value of the state interval to improve its prediction accuracy for concrete fatigue life prediction.Then the fatigue life test results of RC with different stress levels were used as the original data to establish the fatigue life prediction model of RC based on the improved gray Markov model,and the accuracy of the model and the prediction results before and after improvement were compared and analyzed.The results show that the RC fatigue life N obeys the two-parameter Weibull distribution.The error analysis after converting the predicted values from lg(N)to fatigue life shows that the prediction accuracy of the stress-fatigue life(S-N)curve is low,with maximum relative error of 201.43%.The prediction accuracy of the MGM(1,1)model is improved compared with that of the S-N curve,but the average relative error still reaches 102.20%.The prediction accuracy of the MGM(1,1)model improved by the theoretical improvement of both algorithms has improved considerably,and the average relative error is only 5.62%.The improved MGM(1,1)model is found to have smaller error fluctuations and smaller mean errors than the model in the literature when the experimental data from other literature are introduced for validation and comparative analysis with the model in the literature,and the average relative error is only 1.01%,indicating that the improved MGM(1,1)model has higher accuracy and reliability in predicting the fatigue life of RC.
关键词
再生混凝土/灰色理论/马尔科夫模型/新陈代谢理论/粒子群算法/疲劳寿命
Key words
recycled concrete/grey theory/Markov model/metabolic theory/particle swarm al-gorithm/fatigue life
维普
万方数据