Особливості розробки інформаційно-управляючих систем для біотехнологічних об’єктів

Анотація

УДК 681.631

THE FEATURES OF INFORMATION CONTROL SYSTEMS FOR BIOTECHNOLOGICAL OBJECTS

Dudnyk A., PhD,

Lysenko V., Doctor of Science

e-mail: dudnikalla@mail.ua

The agricultural production in the world and Ukraine is filled with modern high-tech enterprises, the hallmark of which is the presence of a biological component. These companies are, first of all, poultry, greenhouses, mushroom production. The part of energy in production costs for these companies reach sometimes 70 % (greenhouses). Under the conditions of the high cost of energy and its actual deficits measures that reduce energy consumption are topical. Our analysis of international experience in the field of automation of control processes in agriculture showed that all of the existing control systems do not take into account possible future changes in the disturbances, in particular air temperature, on the technological object during the entire period of bioobject housing (growing), as well as the dynamics of bioobject states and perform exclusively stabilization mode for technological parameters, given the instantaneous values of the disturbances that is not always effective

With rampant increasing of energy prices is important to use control algorithms of electrotechnical complexes which accompany appropriate technology, taking into account the biological filling state and maximize production profit primarily by reducing energy costs. The intelligent control systems of electrotechnical objects are able to form such algorithms, which are used the theories of stochastic processes, neural networks, game theory and statistical decisions, etc.

Plants states are influenced by solar radiation. Thus, there is a need for analysis and forecasting of outside air temperature and solar radiation intensity for using the forecast results in the formation of electrotechnical complex control strategies with the aim to reduce energy costs in the production of agricultural products.

The solution of this problem is possible in two variants: the identification based on the theory of random processes images (areas) of natural disturbances that characterize the implementation of temperature; neural networks.

Unlike the prediction methods of the theory of random processes, neural network forecasting does not require statistical data on temperature changes in the past, their images developing a system that is running under uncertainty. However, it takes a long time to study and forecast reliability is sensitive to the quality of representation of input information, and cannot enter a priori (expert) information to accelerate learning network.

The research results of intelligent control systems of electrical complexes for biotechnological objects are given in this paper. The study of developed control system shown improving the system performance by 20 % over and reducing of natural gas consumption for heating by 13 % compared with control system that is based on the stabilization algorithm of technological parameters.