Prediction of Combustion State of Industrial Kiln based on CNN-LSTM-Attention Model

This paper constructs a CNN-LSTM-Attention linear fusion kiln state prediction method, aiming to overcome the current industrial kiln combustion state monitoring problems such as low accuracy and slow response. Firstly, a 30-dimensional kiln combustion parameter dataset is established by multi-sensor fusion; then, the samples are balanced and the data features are standardised by various data processing operations; finally, an integrated system of prediction and warning is constructed through model training and optimisation. The experimental results show that the model accuracy reaches 0.97; at the same time, the superiority of the CLA model is verified through ablation experiments and comparison tests.