Selection of solutions for the operational neurocontrol of the mixture grinding process in cement production

The article proposes a method of neurocontrol by the technological process of grinding the mixture in cement production in order to increase its energy efficiency. The need to use neural control is caused by the fact that the quality of grinding and the consumption of resources depend on many factors that present great difficulties in their measurement and prediction of performance indicators. Reliable measurement of influencing factors is necessary to solve the problem of determining the best combination of the volume of the ball load of grinding and the required amount of solids to optimize the rate of reduction of the particle size of the charge with a minimum specific energy consumption. Neurocontrol is based on the training of a neural network with a teacher, which is played by an experienced mill operator, who realizes the effective control of the grinding process. The controller, built on the basis of the neural network, should work in real time and reflect the current state of the grinding process. The choice of solutions for solving operational control problems using a neural network belongs to the class of multi-criteria tasks. The paper proposes a decision-making method based on the set of permissible technical conditions imposed on the grinding process. Such a formulation of the problem is generally contradictory. The paper proposes an approach to solving this problem on the basis of determining the maximum number of joint weighted conditions imposed on the process. This approach allows you to organize an interactive procedure for selecting a feasible solution for the operational control of the grinding process. An operative computer model of the clinker grinding process in cement production is proposed.