Ball mill bearing life prediction method based on ST-BiLSTM

The invention provides a ball mill bearing life prediction method based on ST-BiLSTM, and the method comprises the following steps: collecting an original time domain vibration signal of a rolling bearing in a continuous time period, carrying out the noise reduction smoothing processing of the original time domain vibration signal, obtaining a smooth time domain vibration signal after noise reduction, and carrying out the prediction of the life of the rolling bearing. Performing normalization processing on the smooth time-domain vibration signal to obtain normalized data, dividing the normalized data into a training set and a test set, inputting the data of the training set into a TCN-BiLSTM-ATTENTION model for training, then performing output by adopting a full connection layer to obtain a trained TCN-BiLSTM-ATTENTION model, inputting the test set into the trained TCN-BiLSTM-ATTENTION model, and finally, outputting the trained TCN-BiLSTM-ATTENTION model to obtain a trained TCN-BiLSTM-ATTENTION model. And obt