Application of Artificial Neural Network and Fuzzy Logic Controller in Power Quality Improvement

Bishi, Monika (2017) Application of Artificial Neural Network and Fuzzy Logic Controller in Power Quality Improvement. MTech thesis.

[img]PDF (Fulltext is restricted upto 23.01.2020)
Restricted to Repository staff only

1659Kb

Abstract

The main reason of the occurrences of many undesirable phenomena in the operation of power system is the extensive use of the nonlinear loads for example arc furnace and static power converters. Among all power quality issues, Harmonic contamination can be taken as the most important. Due to the introduction of nonlinear load, harmonic components are getting injected in the source side and hence affects the power distribution system, and the sensitive loads. Hence it has become the major concern. Harmonic current components can cause excessive heating in rotating machinery, hence increases the power system losses, also causes significant interference with the communication circuit which share a common path with the AC power lines, and also results in the erroneous operation of the control and regulation circuits due to the noise generation.
Previously passive filters are used to mitigate the harmonic components but they have certain limitations. Hence to solve these power quality issues active power filters are developed over the years. From different types of APF configuration, Shunt active power filter is used to compensate the reactive power demand by the nonlinear load and also eliminates the current harmonics. A control techniques which uses the artificial neural network based controller is discussed in this thesis. ANN based controller makes the system adaptive. Hence reactive power and harmonic components can be compensated by the adaptive shunt active power filter. An unity power factor can also be maintained by the proposed ANN based controller. The shunt active power filter control strategy also requires a self charging circuit, which is used to maintain a desired DC link voltage. It can be made possible with a PI controller or a fuzzy logic based controller. Here a fuzzy proportional controller with crisp integral controller is discussed to regulate the capacitor voltage of the inverter. Integral control is used to eliminate the steady state error. Hence the proposed system consists of two soft computing techniques: ANN used for fundamental component extraction, another controller used fuzzy logic to regulate the DC voltage.
The performance of the adaptive shunt active filter can be discussed by using SIMULINK or MATLAB model. The operation and design concept of the proposed system can be verified from the simulations results. Detail description of each section of the proposed system is also given.

Item Type:Thesis (MTech)
Uncontrolled Keywords:Active Power Filter; Neural Networks; Power Quality; Weight Updating Technique; Adaptive Linear Neuron (ADALINE); feed-forward multilayer neural network (MNN); HFP+I controller; PI controller; Fuzzy-PD controller; Hysteresis controller
Subjects:Engineering and Technology > Electrical Engineering > Power Systems
Engineering and Technology > Electrical Engineering > Power Electronics
Divisions: Engineering and Technology > Department of Electrical Engineering
ID Code:8896
Deposited By:Mr. Kshirod Das
Deposited On:04 Apr 2018 11:15
Last Modified:04 Apr 2018 11:15
Supervisor(s):Panda, Anup Kumar

Repository Staff Only: item control page