MODELLING OF THE PROCESS OF CONTROLLING THE STATE OF A CLOSED AQUATIC ENVIRONMENT USING FUZZY LOGIC

In the state of significant development of aquaculture, the problem of managing the parameters of the aquatic environment becomes more urgent. Existing classic systems for controlling parameters of closed aquatic environments use strict models of dynamic control and are used for monitoring, control and regulation of those parameters. These models have a number of disadvantages related to the specifics of strict algorithmic control and the complexity of adequate models for growing living organisms. Such models require significant calculations related to solving differential equations and are bad in adaptation to uncommon changes of parameters due to external factors. In addition, such models are prone to fluctuations in parameter value changes in transient processes. The study emphasizes the importance of an adaptive control approach in aquatic biotechnical systems operating in unpredictable environments. The proposed model accounts for the uncertainty and incompleteness of information that arises during the monitoring of parameters such as temperature, oxygen level and other critical indicators. The implementation of such a system can reduce the risk of biological material loss and improve the efficiency of technological processes. The article also outlines the potential for using the model in conjunction with modern real-time data acquisition systems to improve control accuracy. To improve the control process, the use of a fuzzy logic apparatus is proposed. An improved model for controlling the state of closed aquatic environment based on the Mamdani method was created. A controller structure with a fuzzy logic inference module based on the transformation of input signal data into linguistic variables has been developed, which allows to avoid the solution of differential equations and transfer the solution of the problem to the system of logical rules activation. The work of the created model was tested and compared with the classic control system. The results of experimental testing confirm the effectiveness of the approach: the control error has been reduced, and the stabilization time after external disturbances has been shortened. The created model allows to adequately respond to changes in environmental indicators and avoids fluctuations in the value of the controlled parameter, which makes the system work more predictable and reliable. The prospects of using the developed model for combining the values of several input parameters for the formation of a logical conclusion are given.