The Bayesian network control method for the safe operation of the flotation process based on time-delay transfer entropy

In complex industrial processes, the current state of a variable is influenced not only by its associated variables but also by its historical states. Abnormal conditions in the industrial process lead to inaccurate identification of causal relationships by the control model. To address this issue, this paper proposes a Bayesian network (BN) control method for the safe operation of the coal slurry flotation process based on time-delay transfer entropy. The method incorporates transfer entropy on the basis of the Bayesian Information Criterion scoring function to eliminate the interference of the historical state of variables on causal identification and describes the causal relationship between variables more accurately by calculating the information interaction between variables. In addition, to eliminate the influence of time delay on information interaction, the time delay parameter is introduced in the calculation process of transfer entropy to effectively improve the accuracy of the BN structure.