Development of an adaptive control system for automatic regulation of voltage and reactive power for power substation 500/220/110/10 kV
Kalinkin, Sergei (2019)
Diplomityö
Kalinkin, Sergei
2019
School of Energy Systems, Sähkötekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2019101532697
https://urn.fi/URN:NBN:fi-fe2019101532697
Tiivistelmä
In this master’s thesis the development of an adaptive PID controller for the automatic control system of reactive power compensation units, installed at the high voltage power substation of 500 kV is considered. The control system was designed as adaptive to ensure the required quality of a voltage regulation process in various operating modes of power network, some of which can lead to significantly different behavior of voltage transients.
After studying the current trends in designing various methods of adaptive control, an approach based on the application of artificial neural networks was chosen. This work is intended to investigate whether the application of feedforward artificial neural networks is suitable to solve the problem.
There were created a model of power network and a model of adaptive control system in Matlab Simulink. The first one was used to obtain the training data for artificial neural network, which adjusts the coefficients of PID controller, and verify the operation of resulted control system.
In general, the designed adaptive control system shows that feedforward artificial neural network can perform a correct identification and provide correct PID coefficients to ensure the required quality parameters of the transition process. As expected, the qualitative functioning of neural network significantly depends on the training data.
After studying the current trends in designing various methods of adaptive control, an approach based on the application of artificial neural networks was chosen. This work is intended to investigate whether the application of feedforward artificial neural networks is suitable to solve the problem.
There were created a model of power network and a model of adaptive control system in Matlab Simulink. The first one was used to obtain the training data for artificial neural network, which adjusts the coefficients of PID controller, and verify the operation of resulted control system.
In general, the designed adaptive control system shows that feedforward artificial neural network can perform a correct identification and provide correct PID coefficients to ensure the required quality parameters of the transition process. As expected, the qualitative functioning of neural network significantly depends on the training data.