Intelligent prediction method and system for residual life of main speed reducer, equipment and medium

Abstract

The invention relates to an intelligent prediction method and system for the residual life of a main speed reducer, equipment and a medium. The method comprises the steps of obtaining vibration data of multiple monitoring data channels of target equipment; performing local feature extraction on the vibration data of each monitoring data channel by using a preset multi-size convolutional neural network to obtain corresponding equipment failure feature information; extracting the equipment failure feature information of each monitoring data channel by using a preset deep feature extraction network to obtain corresponding deep time sequence feature information; inputting the deep time sequence characteristic information of the plurality of monitoring data channels into a preset residual life prediction initial model for training to obtain a residual life prediction target model; and based on a linear regression smoothing strategy, predicting the residual life of the target equipment at the current moment by using the residual life prediction target model. The method and system can effectively detect the fault characteristics of the automobile main speed reducer, and can accurately predict the residual life of the automobile main speed reducer.

Description

technical field [0001] The present application relates to the technical field of mechanical equipment life prediction, in particular to an intelligent prediction method, system, equipment and medium for the remaining life of a final reducer. Background technique [0002] As the key component of reducing speed and increasing torque in the rear axle of the automobile transmission system, the final drive is the main source of automobile failure. has an important impact. Therefore, in order to ensure the safe and reliable operation of the vehicle, it is necessary to carry out real-time state detection and remaining life prediction for the main reducer of the vehicle. [0003] At present, the real-time condition monitoring and remaining life prediction of the main reducer of the automobile are mainly realized in two ways. One is to simulate the decline trend of the equipment by building a model of the mechanical equipment. This method requires a full understanding of the working...

AI Technical Summary

Problems solved by technology

[0004]First, due to the change of the load during the driving of the car, the vibration interference of other components in the transmission system, and the sudden Noise, creating a strong noise environment, it is difficult to obtain effective vibration signals

The wear and degradation process of the main reducer in a strong noise environment has the characteristics of strong nonlinearity and instability. It is difficult to directly predict the vibration acceleration signal collected by the sensor to achieve ideal results.

[0005]Second, since a single sensor is limited by its location and direction, using multiple sensors to collect multi-channel time-series data can effectively reduce the uncertainty of single-source data. However, the heterogeneity of multi-sensing data is inevitably introduced, which increases the difficulty of deep temporal feature extraction.

Traditional technology is not capable of extracting timing features from sensing signals, which easily leads to loss of important features

[0006]Third, in practical applications, due to the complex operating environment and the randomness of load conditions, the remaining service life is a random variable, and the final drive There is a complex nonlinear relationship between the remaining service life and the running time, and the discontinuous output of the traditional prediction model will lead to low prediction accuracy

Images