Zinkevych P. O. Automated control of energy supply systems of objects with active consumers. – Qualifying scientific work on manuscript rights.
Dissertation for obtaining the scientific degree of Doctor of Philosophy in specialty 141 "Electroenergetics, electrical engineering and electromechanics". - National University of Food Technologies of the Ministry of Education and Science of Ukraine, Kyiv, 2023.
In the dissertation, the urgent task of increasing the efficiency of electricity consumption and electricity supply of industrial and civil facilities with active energy consumers is set and solved, by creating an automated power supply management system based on the use of intelligent control methods: intelligent methods of load forecasting and electricity generation, algorithms for determining the load profile of active consumers and demand management of active consumers, the algorithm for managing electric energy storage devices, methods of synthesis of fuzzy regulators.
In the first section, an analysis of electricity supply with active consumers was carried out: a complex of issues related to the development of automated control systems for electricity supply with active consumers using intelligent control methods and information technologies, including active consumer demand management subsystems, modes of energy storage during peak compensation, was analyzed load, ensuring energy-efficient voltage levels in systems with active consumers (own generation).
The characteristics of the "active" consumer as an element of the energy supply system, in which the consumer's behavior can be controlled, are given: integration of own generation, a mechanism for stimulating unloading, a counter-planning mechanism, and demand management mechanisms. The formulation of the research task is presented.
In the second chapter, a systematic analysis of the process of managing the generation, transmission and consumption of electric energy at industrial and civil facilities with active consumers is carried out. Proposed approaches to building an automated power supply management system for industrial facilities with active consumers and organizational and technical means of managing the power supply system of an industrial facility.
According to the SADT methodology, based on the IDEF0 notation, a system model of power supply control of the central office with active consumers was built.
In the third chapter, the methods of short-term forecasting of generation of electric power plants and load of consumers are considered, in particular, statistical methods and intelligent methods used for control tasks. A comparative analysis of short-term one-step PEN methods was performed, based on ARIMA models, "naive forecast" and ANFIS methods. The forecasting results showed that using the method based on ARIMA (2,1,2) is more effective compared to the "naive forecast" and ANFIS methods. At the same time, for the training and test samples, the specified model provides an RMSE of 0.0317% and 0.0359%, respectively.
A comparative analysis of short-term multi-step PEN methods was performed, based on ARIMA models, "naive forecast" and ANFIS methods. The forecasting results showed that the use of the method based on ARIMA (4,1,2) is more effective compared to the "naive forecast" and ANFIS methods. At the same time, the specified model provides an RMSE for the test sample of 0.052%.
The analysis of the method of short-term multi-step forecasting of FES generation, based on the artificial neural network NARX for the prediction for the next 24 hours, was performed. The conducted studies showed that multi-step forecasting based on the open loop (openloop) provides better results than the closed loop closeloop. At the same time, the specified model provides an MSE for the test sample of 0.0257%.
As a result of the study, it was found that ARIMA and NARX models are effective forecasting tools in automated power supply management systems of industrial and civil facilities with active consumers.
In the fourth chapter, an approach to the construction of an automated power management subsystem of active consumers (APKPAS) is proposed, which is designed to reduce the payment for electric energy and reduce the energy intensity of production due to the use of predicted values of load and generation capacity of generators of own generation (PPS, DH), ensuring effective use generators of own generation, SNE and load management taking into account the peculiarities of tariffs for electric energy (EE).
