A. I. Koldaev, L. B. Kopytkova, I. N. Lavrinenko

The possibility of applying the indistinct logic apparatus for non-fully described systems with unknown dynamics is considered in the article on the example of the problem of defining the condition of management object under the water effect of different nature. The model of microcontroller for managing a reactor installation of butanol hydrogenation is suggested. The five stages of processing the indistinct logic algorithm applied to the investigated reactor installation have been determined. The first stage is the phasefication of inputs. At this stage the initial information is transformed to a convenient kind for application in indistinct logic, the degree of entrance value reference to the indistinct set is determined by the membership function. The second stage is the application of the indistinct operator in the precondition. Two or more values of the membership function received at the phasefication stage are the entrance for the indistinct operator. At the operator exit we get a unique value which is a dimensionless value and determines the condition of the system. At the third stage the implication operator is applied. The exit of each rule is an indistinct set. Before the application of the implication operator it is necessary to choose the weight of all the rules. As the solutions are based on the application of all the rules in the system of fuzzy introduction the rules are to be combined in a certain method of decision making. This process occurs at the fourth stage which is connected with merging (aggregation) all the exits. At the entrance of the aggregation process there should be a set of truncated exit functions returned by the implication process for each rule. At the process exit we receive one indistinct set for each exit variable. The purpose of the fifth stage is dephasefication, i.e. the transformation of indistinct set into an accurate number. The idea of using the Sugeno algorithm for the considered problem is suggested. To increase the accuracy of the exit result it is necessary to adjust the membership function. It is shown, that the application of a neural network allows simplifying the function adjustment process and increasing the accuracy of entrance and exit parameters. The suggested neuro-fuzzy model of the reactor installation management system is proved to be adequate to real data. The designed model can be applied for solving the problem of managing any technological process with non-linear dependence.