Vibration signal characteristic value selection method and elevator health state evaluation or fault diagnosis method

Abstract

The invention discloses a vibration signal characteristic value selection method, which comprises the following steps: performing variance analysis on characteristic values of m * n groups of vibration signals, analyzing the relevance between the characteristic values and label values of vibration signal state labels, and performing primary coarse screening by deleting the characteristic values which are in low relevance with the label values; calculating a distance correlation coefficient of each characteristic value and other characteristic values to obtain a comprehensive distance correlation coefficient, performing characteristic value correlation analysis, and performing secondary coarse screening by deleting the characteristic value with the high comprehensive distance correlation coefficient; and performing eigenvalue matrix combination on the eigenvalues reserved by the second coarse screening, and performing recursive feature elimination based on the combined eigenvalue matrixto obtain an optimal eigenvalue combination. According to the vibration signal characteristic value selection method, the characteristic values of the vibration signal data can be effectively screened, the optimal characteristic value combination forming the mechanical equipment operation state vibration model is obtained, the dimensionality of the model is effectively reduced, calculation resources are saved, and the accuracy of the model is improved.

Description

technical field [0001] The invention relates to the technical field of machine learning and data mining, in particular to a method for selecting vibration signal eigenvalues ​​and an elevator health status assessment or fault diagnosis method. Background technique [0002] In recent years, with the rapid development of sensor technology, miniaturized and low-cost vibration sensors are used to obtain vibration signals during the operation of mechanical equipment by installing them on mechanical equipment, such as figure 1 shown. A large number of normal, abnormal and faulty vibration signals of mechanical equipment obtained by sensors are used to extract the characteristic values ​​of the corresponding vibration signals after data cleaning and signal processing. Based on machine learning technology and data training, a diagnostic model of mechanical equipment operating status is established. According to the above-mentioned diagnosis model of mechanical equipment operation s...

AI Technical Summary

Problems solved by technology

[0007] In order to obtain the best eigenvalue combination, the method of recursive feature elimination is usually used for eigenvalue selection, but its disadvantage is that when the number of eigenvalues ​​is too large, the calculation scale is huge, and the entire recursive process is too long, which affects the efficiency of feature selection; In addition, due to the randomness of the selection of the training data set and the test data set, the effect feature value deleted during the recursive process may be affected by the data distribution, thereby affecting the accuracy of the diagnostic model of the mechanical equipment operating state

Images