Formation of the separation characteristic of ore material thickening based on the model predictive control method

Abstract

The aim of the investigation is to form an optimal separation characteristic of beneficiation processes on the basis on operational information on time-varying of their parameters exemplified by the process of thickening ore raw materials. Methods of research. In the research work, the following methods are used: analysis of scientific research and practical experience; statistics methods and the probability theory for estimation of experiment results; methods of analytical synthesis and numerical simulation; methods of model predictive control for developing control algorithms of the thickening process; numerical simulation methods for synthesizing and analyzing a mathematical model. The scientific novelty of the investigation is in finding optimal values of the control horizon and the prediction horizon in terms of quality control for a single-channel system of model predictive control of ore raw material thickening. Practical significance involves development of methods and software for determining the control horizon and the prediction horizon values of the single-channel system of model predictive control of the process of ore raw material thickening that are optimal from the point of view of quality control, this enabling optimization of separation characteristics of ore raw material thickening. Results To form a separation characteristic of the process of ore raw material thickening based on model predictive control for the single-channel control system of the thickening process, satisfactory control results are achieved by setting the control horizon equal to one interval. For this value, the quadratic control error does not exceed 0.1452-0.1474. A further increase in the prediction horizon is not feasible since it does not allow significant reduction of the quadratic control error. At the same time, the value of 3-5 intervals is sufficient for prediction horizons. These values are determined by an increase in computational complexity of prediction by 10-20 intervals, which causes a slight decrease in the quadratic control error.

4th International Conference on Sustainable Futures: Environmental, Technological, Social and Economic Matters

“…• Analytical advancements. Recognition of mineralogical and technological varieties through ultrasound backscatter spectrograms signifies progress in analytical methods, aligning with the industry's pursuit of precise and sustainable resource characterization [223]. • Predictive models for safer operations.…”