| With the acceleration of urbanization in China and the rapid increase of permanent population in urban areas,urban rail transit,as a fast and convenient way of public travel,has been welcomed by many cities.With the promotion of national policies,urban rail transit in China has now ushered in an unprecedented development.However,with the continuous growth of passenger flow and operating mileage,the working environment of rail vehicles is becoming more and more complex.As the most critical part of the rail vehicle structure,the requirements for reliability are also increasing and how to ensure the bogie has enough reliability of fatigue life in operation has become an urgent problem to be solved in relevant industries.At present,the reliability test analysis of the frame in China is still in its infancy,and a standardized system has not been formed.Therefore,based on the finite element simulation and test data analysis,the following research is carried out in this paper for the reliability testing technology and fatigue life evaluation of the frame:(1)Based on the finite element theory,according to the relationship between displacement mode and strain mode,the expression of strain mode is derived,and the orthogonality of strain mode is verified.Based on bayesian system identification theory,the optimal sensor placement based on minimizing information entropy theory is studied,and the influence of prediction error on the results is analyzed.An optimization method based on improved prediction error is proposed.The reciprocal of the sum of normalized modal strain energy is used as the principal diagonal element of the prediction error covariance matrix.The non-diagonal element is expressed by an exponential correlation equation combining the distance of the measuring point and the response level.Aiming at minimizing the information entropy,the forward sequential sensor placement method is used to obtain the optimal location of measurement points.Taking the bogie frame as the research object,the sensor optimization is carried out based on the finite element model.The results show that the optimization method proposed in this paper can effectively reduce the aggregation effect of measuring points,improve the uniformity of configuration,and maintain a high amount of measuring point information and good modal independence.(2)The two kinds of interference characteristics,power frequency interference and pulse interference,which are often found in the measured strain signals of the frame,are analyzed.According to its statistical independence with strain signal and its morphological independence with strain signal,the method of noise reduction based on blind source separation and compressive sensing theory is studied.The algorithms of removing power frequency interference based on independent component analysis and removing pulse interference in narrowband signals based on morphological component analysis are proposed.Based on the analysis of existing algorithms,the factors affecting the convergence speed and accuracy of the two algorithms are studied.An independent component analysis method based on improved particle swarm optimization(PSO),and morphological component analysis method based on fast adaptive shrinkage threshold algorithm(FASTA)and p-exponential threshold denoising method are proposed.The simulation signal and the measured signal are used to verify the performance of the algorithm.The results show that the improved method has a significant improvement on the noise reduction effect compared with the traditional method.(3)Based on the frequency domain fatigue theory,the mutual coupling relationship between loads is studied,and the frequency spectrum calibration method of load spectrum is proposed.Based on the correlation between the measured data of the load,the crossspectral density parameter between loads is used as the characterization of the coupling between loads.Based on the basic theory of multi-axis frequency domain fatigue,the expression of equivalent stress in the frequency domain is derived.The 0-order spectral moment of the self-power spectral density of the stress signal is used as the parameter to characterize the damage,and the contribution ratio of the constrained load to the damage of the measuring point is determined.According to the principle of damage consistency,the NSGA-II multi-objective optimization algorithm is used to calibrate each load system.Based on the "load stress" transfer coefficients calibrated in the laboratory and the measured load and stress data,the load system is calibrated based on the optimization method considering the coupling effect as well as without considering the coupling effect respectively.The results show that the error band of the damage calibration by calculated with considering the coupling effect is lower than that by calculated without considering the coupling effect,and the load adjustment range is smaller than that of the latter.(4)The factors affecting the randomness of cumulative damage were studied.Based on Miner’s linear rule,cumulative damage was considered as a random variable.The new heuristic algorithm is used to fit the probability density function of the stress spectrum,and combined with the S-N curve equation of the material,the probability density function and statistic of the single damage increment are derived.The influence of load sequence on cumulative damage is studied.Based on the idea of damage transformation,a probability cumulative damage model considering load interaction effects is established.A Wiener process with fixed effects and random effects is established.The Monte CarloMarkov chain method is used to estimate the reliability of the bogie frame related to the operating mileage based on the cumulative damage data.Finally,based on the long-term tracking measured data,the reliability of the gearbox suspension is calculated.The results show that using heuristic algorithm to fit probability function is more consistent with the actual data.The Wiener process with random effects can reveal the reliability variation more accurately than the Wiener process with fixed effects.(5)Based on the spectral decomposition method in the frequency domain fatigue theory,the measured wideband random signal is converted into a constant frequency narrowband constant amplitude load signal for bench test.The relationship between the load correlation coefficient and the phase is derived.Considering the coupling between the measured loads of the line,the correlation coefficient and phase difference between the program loads are determined according to the requirements of the test bed control system,and the calculation method of the expected rate of damage is given.According to the principle of damage consistency,the minimum mean square error between the calculated cumulative damage rate and the measured cumulative damage rate is used as the objective function to calibrate the program load amplitude.The calculated damage of the bench test spectrum based on the measured data is linearly extended to 2.4 million kilometers,and compared with the damage obtained according to the bench test method specified by the standard.The results show that the the bench test spectrum based on the measured data improves the test efficiency,and the life evaluation results are close to the measured values. |
PDF Full Text Request