Computer-aided dynamic analysis and control design for grinding circuits

"A procedure f or constructing a grinding mill dynamic model is proposed. This model is used in a dynamic simulator of are grinding circuits. The simulator is designed for flexible flow sheets involving different grinding mills and particle classifiers and is able to accept several single variable PID control loops. The dynamic model is validated on a laboratory ball mill, and calibrated to an industrial installation. The simulator is used for characterization of grinding circuit dynamic behaviour and tuning of control strategiesIntroductionThe ore grinding operation is a critical step of the ore preparation process because it is the last comminution stage before the separation of minerals. The properties of the ground minerals to be concentrated, control the quality and efficiency of the separation process. This is why it is important to control properly the operation of the last grinding stage, commonly constituted of a closed circuit ball mill with classifier . To provide an adequate fine product, the grinding operation should compensate for the perturbations of the feed characteristics (size analysis, grindability, etc.). The main objective is to maintain a constant size distribution for providing stable operating conditions to the separation units. Simultaneously another usual objective is to operate the grinding circuit at a maximum capacity.Automatic control applied to grinding circuit s is the appropriate approach to improve its operation. Most frequently average optimal operating conditions are determined from steady -state performances of the circuit , and the automatic control is used for stabilizing the operation at selected values of the process variables . In the traditional approach, one or several single input! single output (SISO) control loops are implemented and empirically tuned. Advanced control practices'!' are based on a quantitative analysis of the circuit dynamics, which could also give a basis for the tuning of conventional SISO 100ps(2)."