Polupan V. Automated control of the juice purification station based on intellectual decision support system and methods of coordination.

In the dissertation the new scientific and technical task solutions of increasing the defecosaturation station efficiency on the sugar plant are proposed at the expense of the intellectual decision support system and coordination methods. As an analysis result of technological processes and defecosaturation station functioning features in general, the need to supplement the control system with the decision support subsystem in order to coordinate the control subsystems work within the technological process of sugar juice cleaning was identified. To solve the tasks, methods of modern theory of automatic control, statistical analysis, multicriterion optimization of simulation model and modern information technologies in production automated control systems were used. As an analysis result of the defecosaturation station experimental data as a complex control object, the main properties and behavior, the interconnection between the adjacent sections and the subsystems of the defecosaturation station were revealed. The formulation and solution of the defecosaturation station control subsystem coordination task is distinguished from the known by use of the modified genetic algorithm. The algorithms research for solving the defectosaturation station subsystems coordination problem as a Pareto compromise between the quality juice indicators provision and the controlled object productivity, which in turn leads to improved juice quality and the plant productivity. The possibility of using the classical hybrid algorithm for solving problems has been checked. Selected settings for the genetic algorithm. Also added to the classic genetic algorithm of the hybrid function of the PSO. The algorithmic and software of the defecosaturation station subsystem coordination procedures, based on the hybrid genetic algorithm, has been developed, which makes it possible to solve the coordination set task in an efficient and effective manner. Static models were developed for the defecosaturation station subsystems quality estimation and productivity performance by using regression decision trees, the use of which made it possible to calculate defecosaturation station productivity and quality performance. The automated control system structure of the defecosaturation station has been improved by applying an intelligent decision support subsystem. As a theoretical and experimental researches result, the structure, algorithmic and software of the defecosaturation station automated control system was developed on the basis of intelligent decision support systems and coordination methods. The developed algorithmic and software was implemented at the Shamraev Sugar Plant and is used in the educational process at the Department of Integrated Automated Control Systems of the National University of Food Technologies, which is certified by appropriate acts and certificates.