Evaluation model for mill load based on PSO-SVM algorithm

Mill operating load is an important indicator to ensure that the mill is operating in an optimal condition and to avoid downtime. This paper proposes a mill load evaluation method based on industrial big data and soft measurement models. To solve the problem that the mill load is difficult to measure and characterise in real time, a large amount of data generated by the mill operation under the industrial field connection is collected and a soft measurement prediction model is established. Some features are selected from the operating mechanism of the mill and the statistical features of the data to construct the SVM prediction model. The model is optimised by the PSO algorithm to make it more accurate. The root mean square error of the test model is obtained to be only 0.0502. Combining the feed amount, mill power and predicted ore return amount according to the empirical model, it performs the real-time evaluation of the load state of the mill, and real-time early warning of the operation of the mill based on the evaluation results.