Load frequency control of multi source multi area hydro thermal system using soft computing techniques and facts devices

Abstract

Load Frequency Control (LFC) is the predominant control in newlinepower system to match the generation with demand. Power system size, newlinecomplexity and the dynamic nature of load, makes LFC more significant. The newlinespeed governor which acts as the primary control of LFC, does the real power newlinematching using error in frequency. The offset in the frequency is removed by newlineusing secondary controller. The secondary controller is developed using the newlinetransfer function model of the power system. newlineHydro and thermal power plants are considered in this work. Two newlinemodels are developed. Single source multi area hydro thermal system is one newlinein which, thermal plant and hydro plant are considered as separate areas and newlinethey are connected using tie-line. Similarly, considering hydro and thermal newlineplants are connected to each area to form multi source multi area hydro newlinethermal system. When the systems are subjected to unit step load disturbance of newline0.01 p.u., there is steady state error in frequency and tie-line power due to the newlinedrooping characteristics of the governor which is removed by using secondary newlinecontroller. Conventional Proportional Integral (PI) controller is used in this newlinework as secondary controller. In this thesis, the PI controller is tuned using Ziegler Nicholsand#146; (ZN) newlinemethod and Genetic Algorithm (GA). Both the methods remove the steady state error in frequency and tie-line power. GA tuned PI controller provides newlinebetter transient and steady state response when compared to ZN tuned PI newlinecontroller. But in both the methods, the gain values of PI controller are fixed. newlineBased on the system condition, the gain value of the PI controller is newlinechanged using Fuzzy Gain Scheduling (FGS) Controller. The FGS has area newlinecontrol error and change in area control error as the inputs and the outputs are newlinethe gain values of PI controller. Seven linguistic variables for each input and newlineoutput are considered with the total rule of 49. Center of gravity method is newlineused for defuzzification.