A safe control scheme for the dense medium coal separation process based on Bayesian network and Active Learning

The paper presents a safe control method for the operation of dense medium coal preparation process based on Bayesian network (BN) and active learning. The dense medium coal preparation process fluctuates significantly in operating conditions and abnormal operating conditions occur frequently due to the uncertainty in the environment. To address this issue, we establish the BN model of safety control by integrating qualitative expert knowledge and quantitative data information based on the analysis of causes of abnormal conditions and corresponding operation schemes, which acts as the basis for the safety control. Because the previous BN structure learning method based on observation data information can only get Markov equivalence class and needs a large amount of data, it is difficult to establish an accurate BN. To address this issue, we propose a BN structure learning method based on active learning to improve its efficiency and accuracy. Simulation results show that the proposed method can effectively provide safety control decisions for abnormal conditions in the dense medium coal preparation process.