Lutska N. Robust optimal control systems for complex technological objects of food industry

Based on the results of the dissertation research, as well as a set of research and applied works on the topic of the dissertation, is proposed a solution of the scientific and technical problem, which consists in the development and implementation of robust control systems for technological objects of food industry (TOFI) in conditions of uncertainty. In addition, the created of scientific and methodological foundations for their application and designed to improve energy, resource efficiency and quality of production. Complex TOFIs characterized by many characteristics, including nonlinearity, delay, multidimensionality, nonstationarity and others. Accurate mathematical models of such TOFI have a high order and as a result complicate the development of the control system (CS). However, you can specify a mathematical model of loworder TOFI and describe the area of uncertainty of the parameters or structure of such a model. Then the synthesis of the control system TOFI must be carried out according to robust criteria. Today, there are various methods of robust-optimal control, but the features of TOFI require a special approach to the development and design of such systems, based on the combination with traditional methods of synthesis and structures of the TOFI CS. The systematization of existing methods of robust stability, robust-optimal methods, as well as methods of CS synthesis for plants with uncertainties carried out. Problems that complicate the application of robust methods for complex TOFI, which include lack of simple methods for identifying mathematical models of TOFI with uncertainties, focused on robust control and misunderstanding of the essence of robust control criterion and its relationship to engineering requirements. For the monitoring task, quantitative estimates of the reproducibility of the TOFI CS proposed, based on which it is possible to assess the degree of aging of the structure and parameters of the CS and to draw a conclusion about the redevelopment of the CS TOFI. The reproducibility estimation performed on a neural network that calculates a series of the following values of adjustable and control variables. Based on the proposed method of effective design and monitoring of the TOFI CS, consisting of the proposed solutions, a decision support system has been created, which automatically performs the development and monitors the reproducibility of the TOFI CS. Several architectures for the implementation of the decision support system and its individual components proposed. The probability of the developed method and the possibility of its implementation in the food industry based on the consistent and correct use of the proposed ontological system and algorithmic software. Adequacy confirmed by the calculations and implementation acts at food enterprises.