Intelligent control system for technological complexes of aluminum industry enterprises

Relevance. Aluminum production is one of the most important industries all over the world. It has a high environmental load, resource and energy intensity. To reduce the negative impact of aluminum production, efforts are underway to improve its technological processes, aimed at increasing aluminum recovery rates, reducing waste volumes, and lowering energy consumption. The reduction of energy consumption can be achieved, among others, through implementing control systems that provide energy management. Due to the complexity and multiplicity of technological processes in aluminum production, the use of traditional linear approaches for such control systems is ineffective, and adequate mathematical models are required. At the same time, for the production level, the use of mathematical models based on a phenomenological description of the ongoing physical processes overly complicates the control task. As a result, the use of intelligent and adaptive approaches to both the mathematical description of technological processes and the optimal energy consumption management is relevant. Objects. Technological complexes of aluminum production, which have the properties of inertia, nonlinearity, and closedness; and the control system based on artificial intelligence methods. Alumina production is chosen as an illustration. Aim. To develop mathematical models capable of adequately describing the interrelated processes in the technological complexes under consideration, as well as the control system that allows optimal energy consumption management. Methods. For the mathematical model based on balance equations under uncertainty, the fuzzy-set theory was used along with the gradient descent method to identify the model parameters; for the optimization task, the genetic algorithm method was used. Results. The mass balance model and the process conditions changing model have been developed to determine the energy consumption for the technological complexes of aluminum production with continuous inertial nonlinear closed production. Based on these models, the dynamic characteristics of energy consumption and the parameters of technological processes were determined depending on the main controlled parameters, allowing us to predict emergencies. Considering technological parameters and cost factors, the optimization task for energy consumption management was solved.