Abstract: Aiming at the problem that the load spectrum prediction of key stress point damage deviates greatly from the actual damage due to the loss of the correlation of loads after rain-flow counting, the phase reconstruction method of high-speed train frame load is studied. Taking bogie frame of a high-speed train as the research object, the load time history and the time history of key stress points are obtained through line measurement, and the correlation between different load system components of key stress measurement points and the correlation between stress and load components are analyzed. On this basis, the relationship between load correlation coefficient and load phase is established in combination with the loading control mode of the test bed, and the load phase information lost due to rain flow counting is reconstructed, so that the predicted load damage is closer to the measured stress damage. The results show that the relationship between frame loads is stable, which is the basis of reconstructing the phase matrix of frame loads. The reconstructed load phase matrix makes the load prediction stress closer to the measured stress, and the maximum error decreases from 104% to 6%. The testing time basically has no effect on the phase matrix, and different line conditions and train speed have some influence on the phase matrix, but the average error is not more than 10%. The research work provides an idea for damage prediction by discrete load spectrum and lays a foundation for load spectrum expansion and bench test.

Key words: high speed EMU, frame load, correlation analysis, phase reconstruction