| Aiming at the shortcomings of describing the mean stress relaxation in the rotor wheel fatigue life prediction,this paper combined the fatigue resistance design of a turbine integrated rotor,and adopted the local strain method with the eight-level back stress Chaboche model based on the optimization of genetic algorithm to study and establish a rotor wheel fatigue life prediction method considering the dynamic relaxation of the mean stress,and provided a more accurate life prediction method for the integrated rotor fatigue design of turbine.Therefore,it has important theoretical and engineering practical significance.Firstly,a pulsating fatigue tests on smooth specimens of a rotor wheel material(25Cr XXX)were carried out to study the cyclic hardening/softening characteristics,the variation characteristics of internal stresses(back stresses,friction stresses),and the low-cycle fatigue characteristics of the material.The mean stress cyclic relaxation behavior of the wheel material under pulsating fatigue load was also focused on.The results showed that the mean stress of the rotor wheel material showed obvious relaxation under various strain amplitudes,and the larger the strain amplitude was,the smaller the initial average stress was and the larger the relaxation rate was.The mechanical mechanism of the mean stress relaxation was also analyzed from the perspective of internal stresses(back stress,friction stress).Based on the low-cycle fatigue test results of 25 Cr XXX wheel material,the initial values of the kinematic hardening and isotropic hardening for the Chaboche model of eight-level back stress of the material were determined by the experimental method.Furthermore,the parameters of Chaboche model were optimized by genetic algorithm,and the influences of initial strain amplitude,experimental value input quantity and fitness function on the prediction accuracy of Chaboche model were analyzed.The results showed that the first 10 hysteretic loop data with large strain amplitude and the fitness function controlled by peak/valley stress had high precision in simulation.The mean stress relaxation behavior of the rotor wheel material under different strain amplitudes was simulated using the eight-level back stress Chaboche model with optimized parameters,and the simulation results were compared with the three-level back stress Chaboche model and empirical models(Landgraf,Max Well,and Kodama models).The results pointed out that the eight-level back stress Chaboche model can simulate the mean stress relaxation behavior of the rotor wheel material more accurately than the three-level back stress Chaboche model and the empirical model.Based on the local strain method,an eight-level back stress Chaboche model was used to determine the local stress and strain in the notch,and the SWT model was used to calculate the fatigue damage as well as the linear cumulative damage law was used to establish a rotor wheel groove fatigue life prediction method considering the dynamic relaxation of the mean stress in the root of the wheel groove,and the corresponding program was prepared.A low cycle fatigue test was carried out for a certain type of wheel groove component.The fatigue life of the wheel groove component was predicted using the above-mentioned life prediction procedure.The comparison with the test results shows that: the prediction accuracy based on the eight-level back stress Chaboche model was the highest,and the relative error between the predicted life and the test life was within twice the dispersion band,while the predicted life based on the three-level back stress Chaboche model was significantly lower than the test value,and the accuracy of the prediction results based on the empirical model was reduced under small loads.Therefore,the fatigue life prediction method of wheel groove components based on the genetic algorithm optimized eight-level back stress Chaboche model in this paper is worthy of attention and recommended application in the fatigue resistant design of turbine rotors. |
PDF Full Text Request