الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 138 |  |  |
| | | from | 138 | | |
Abstract
In this thesis, an Active Power Filter (APF) controlled by Fuzzy Logic Controller (FLC) is proposed for Power Quality (PQ) improvement in electrical distribution system by reducing the Total Harmonic Distortion (THD) and compensating the reactive power. APF are essentially three-phase Voltage Source Inverter (VSI) which controlled to provide all the reactive power and only the harmonic active power. The D-Q Synchronous Reference Frame (SRF) theory is used to generate the reference compensating currents for APF. An important task in designing APF is to maintain the voltage of the inverter dc bus constant to ensure that the fundamental active power supplied by the source will be equal to the fundamental active power of the load, and no fundamental active power passes through the inverter into the dc capacitor. Generally, conventional PI controller is used to regulate the dc bus voltage. The conventional PI controller fails to perform satisfactorily under load variation. In this thesis a FLC is used to regulate the DC bus voltage. The proposed APF was tested using the MATLAB/SIMULINK toolbox and the results satisfy the harmonic limit specified by IEEE harmonic standard 519. Simulation results obtained are presented and compared with the conventional PI controller. The simulation results shows that the proposed FLC has better performance over the PI controller especially when the load is varied. The nonlinear loads are the main source of harmonics in the electrical distribution system, and they are responsible for PQ disturbances. So, to improve the PQ, the proposed APF is connected to the Point of Common Coupling (PCC) of the nonlinear loads. But before connecting the filter, there is a very important question. Who will pay the cost of that filter? To answer that question, we need to know the true contribution of each load on harmonic pollution. If several loads are connected to the PCC, it is not possible to accurately determine the true contribution of each load, since the voltage of the PCC is nonsinusoidal. To determine the true contribution of each load on harmonic pollution, the load is disconnected from the network and supplied from a pure sinusoidal voltage source. Disconnecting the load from the network is undesirable manner. So, the Artificial Neural Network (ANN) is used to learn the admittance of the nonlinear load and then predicting the true contribution of each load as it where be supplies from pure sinusoidal voltage source without disconnecting the load from the network. The ANN is a Computational model that try to emulate the structure of the human brain. It can be used as a black box approach to model the nonlinear loads. There are certain characteristics in ANN which assist them in identifying complex nonlinear loads. The ANN help use to know the true contribution of each load in harmonic pollution, and hence, the cost of the installed filter can be recovered from these loads as a percentage of their contribution in harmonic pollution.