Research on life prediction method of rolling bearing based on deep learning and voice interaction technology

Rolling bearing life is an important index to measure the performance of rolling bearing. Therefore, a rolling bearing life prediction method based on deep learning and voice interaction technology is proposed. Bearing vibration signals are extracted from time domain and frequency domain, and PRA data dimension reduction algorithm is used. On the basis of in-depth learning algorithm and voice interaction technology, support vector machine is introduced to generate prediction probability density function, and the residence time of bearing running state is calculated. The degradation state of the bearing is deduced by using voice interaction technology, and the life expectancy of the bearing is calculated to realize the life prediction of the bearing. The experimental results show that when the vibration intensity reaches 12 mm/s, the rolling bearing has been removed. When the vibration intensity reaches 11 mm/s, the rolling bearing life test is regarded as the end of rolling bearing life. The error data of the experimental group was less than that of the control group, and the improvement rate was 19.5%. It is further proved that the designed prediction method can effectively improve the life prediction rate.