Fuzzy Comprehensive Evaluation for Operating Load and Efficiency of Semiautogenous Grinding Process

Semi-autogenous grinding (SAG) processes are widely used in mineral processing. In industrial production, operators often lack a comprehensive and objective assessment of the operating conditions of an SAG process and usually make different control decisions based on experience to meet production demands. This paper explains a two-level fuzzy comprehensive evaluation method for the load and efficiency of an SAG process. First, the operation mechanism of an SAG process is analyzed, and the evaluation indicators of load and efficiency are determined. Then the levels and weights of each indicator are determined by combining expert experience and actual production data. Finally, the operating load level is first evaluated by combining the information entropy weight and fuzzy computation. When the load level is within an acceptable range, a fuzzy comprehensive evaluation of the efficiency level is then performed. The modeling method of efficiency indicator, circulating load ratio (CLR), is integrated with random forest (RF) and Bayesian optimization (BO) to carry out the efficiency evaluation. Experiments using actual production data demonstrate CLR modeling performance with a mean absolute error (MAE) of 1.084% and a mean absolute percentage error (MAPE) of 9.415%. The accuracy of the operating efficiency evaluation is 81.69%. This method provides an objective and uniform criterion for operators to evaluate the operating load and efficiency of an SAG process, thus guiding the regulation of actual production.