A Stacked Autoencoder for Operation Mode Classification of Complicated Industrial Process

In this paper, we propose a novel stacked autoencoder (SAE) based operation mode classification method for the complicated industrial process. In detail, we first add the sparse and regularization constraints into SAE to learn low-dimensional nonlinear representations of high-dimensional data, then the SAE is trained in two steps: unsupervised layer-by-layer pre-training is performed first, followed by supervised fine-tuning. In order to evaluate the efficiency of the proposed method, we conduct extensive experiments on the Tennessee Eastman (TE) process and aluminum electrolysis production process in comparison with several conventional methods, the experimental results validate that the proposed method is more effective than other methods in the mode classification task.