On Development of Neural Network Model of Multi-Parametric Objects of Mining and Metallurgical Production Using Mill-Classifier Complex as Example

The paper purpose is to solve an actual problem of modernization of control algorithms used in the production process of the mining and metallurgical branch of industry in order to improve the efficiency of the technology as a whole. Having analyzed the technological processes of the considered industry, it is revealed that, controlling the production flow, the operator does not often have the means to asses quickly the impact of their control actions on the output parameters of the plant, or the assessment is performed by a laboratory method with a significant delay. Therefore, in the course of research and experiments, a method is proposed to develop a model of a technological object using an artificial neural network (ANN) to predict the future states of the technological process. As an example, the mill-classifier complex is considered as one of the most typical in these branches of industry. The ANN applied is due to the multiply loop character of the object, the drift of the parameters of raw materials and plants, the complex functional dependence of the output from the input signals and disturbances influence. Conducting the experiments, various ways of implementing a dynamic neural network model of the technological object are investigated. Further application of the created model and proved methods will be their use in development a predictive system to assist the operator in multiparameter aggregates control to predict the development of the future progress of the technological process by indirectly assessing difficult-to-measure and non-measurable parameters.