THE USE OF NEURAL NETWORK REGULATORS TO IMPROVE THE ENERGY
EFFICIENCY AND QUALITY OF THE VENTILATION AND AIR CONDITIONING SYSTEM OF A BURIED STRUCTURE 

Igor I. Zvenigorodsky, Candidate of technical sciences, associate professor, head of the sub-department of protective structures, Military Educational and Scientific Center of the Air Force" Air Force Academy named after Professor N.E. Zhukovsky, and Yu.A. Gagarin" (54а Starykh Bolshevikov street, Voronezh, Russia), zvendocent@mail.ru
Yury T. Zyryanov, Doctor of technical sciences, professor, professor of the sub-department of design of radioelectronic and microprocessor systems, Tambov State Technical University (106 Sovetskaya street, Tambov, Russia), zut-tmb@mail.ru
Dmitry I. Ulshin, Candidate of technical sciences, researcher, Military Educational and Scientific Center of the Air Force" Air Force Academy named after Professor N.E. Zhukovsky and Yu.A. Gagarin" (54а Starykh Bolshevikov street, Voronezh, Russia), wm_d@mail.ru

Background. The relevance of the work is due to the possibility of increasing energy efficiency with the help of neural network automatic control systems, since the existing models of ventilation and air conditioning systems and their automatic control systems consider individual processes occurring in the ventilation system and do not take into account all control channels and disturbances, the relationship of adjustable parameters, spatial distribution points of application of influences and variability of the structure of the control object. The aim of the work is to reduce the cost of electricity by improving the efficiency of energy use, as well as improving the quality of management. Materials and methods. To achieve the goals set, the methods of computer neural network modeling were used. Results and conclusions. A neural network model of an automatic control system for the process of air conditioning of a buried structure was built, which operates under the influence of stepwise disturbing influences using neural network controllers that control by the method of "detection of discord". Preliminary calculations of the energy efficiency of the proposed neural network control in real systems show that the saving of electrical energy in comparison with traditional PID control reaches 7–10 % depending on the mode of operation of the system, which in large-scale systems is expedient and in demand from an economic point of view.

automatic control systems, neural network controllers, buried structures

Zvenigorodsky I.I., Zyryanov Yu.T., Ulshin D.I. The use of neural network regulators to improve the energy efficiency and quality of the ventilation and air conditioning system of a buried structure. Nadezhnost' i kachestvo slozhnykh sistem = Reliability and quality of complex systems. 2023;(1):99–105. (In Russ.). doi:10.21685/2307-4205-2023-1-12