Design and development of intelligent control strategies and algorithms for automated control of biotechnical objects under uncertainty

Abstract

Uncertainty of environmental conditions is typical for most biological and agricultural systems. This uncertainty has recently become a serious challenge because of global climate changes. Due to delayed response to unpredictable environmental changes, the functioning of these systems is not always optimal. Another source of uncertainty is the adaptive response of a biological system, which usually does not coincide with the chosen objective function. This study focuses on developing intelligent control strategies to improve the performance of biotechnical systems under uncertainty. We use cause–effect diagrams to analyze the effect of uncertainties on the key performance indicators (KPIs). Our concept of intelligent control is based on the statistical evaluation of environmental uncertainties, while the decision-making is based on game theory. This intelligent control concept is universal and could be applied to different classes of biotechnical systems, from small animal farms to large industrial plants. Sugar production is considered as one of the objects of research, which requires minimizing sugar losses in four main unit operations: diffusion, purification, evaporation and crystallization. Each unit operation was modeled with the artificial neural networks (ANN) trained on the laboratory datasets. Based on simulated or real inputs from the technological process, ANNs predict outputs in terms of quality, losses, and energy efficiency. In stationary mode, ANNs are used to predict optimal technological regimes. They are used to correct technological regimes to maximize the objective function in real time. The application of intelligent algorithms improved process performance and energy efficiency.