Outlier Detection of Cement Rotary Kiln Parameters Based on CNN-LSTM-Attention Model Prediction

The working conditions of clinker calcination process in cement rotary kiln are complex and changeable, and the change of various parameters is an important factor affecting the quality of cement. Aiming at the problem of abnormal detection of parameter data in the production process of cement rotary kiln, this paper proposes an abnormal value detection method based on CNN-LSTM-Attention model prediction. Three parameters of kiln main current, kiln head pressure and kiln head temperature related to common abnormal working conditions are selected and introduced. Firstly, the CNN-LSTM-Attention model is trained using normal historical data to learn the time series characteristics of the data. Then, the trained model is used to predict the test set part, and the abnormal data points are identified by comparing the difference between the predicted value and the actual value of the test set part. The experimental results show that this method can effectively detect the abnormality of various parameters of cement rotary kiln, and provides a more effective means for fault diagnosis and prevention in the production process. This method has good practicability and reliability, and has certain engineering application value.