Optimal Congestion Management in Deregulated Power System

Abstract

The structural arrangement and the operation of electric power industry have been completely changed since last few decades. This restructuring of the power industry has been done to initialize market competition among the various participants to enhance market profits. In a deregulated utility system the generation, transmission and distribution systems work as independent entities and it becomes hard for a system operator to maintain the synchronism among all the systems especially when congestion or line outage occurs. If the congestion persists for a long time, it may harm system security and reliability. Optimal congestion management must be done to alleviate the congestion so that market efficiency, stability and reliability do not alter. Numerous methods are available in art of literature to relieve the congestion. These are broadly classified as convention methods and optimization methods. The conventional optimization methods of congestion management perform perfectly if the objective function has good continuity and differentiability. As the congestion management is a nonlinear complex problem, the conventional optimization methods are not well suited as tuning of assessment weighing factors is a laborious task. The solution of nonlinear optimization problems completely confine on the weights, so the optimization methods are more appropriate for congestion management.

The main objective of the thesis is to develop an optimized congestion management methodology in restructured transmission system in order to optimize the congestion cost in deregulated environment. Thesis also aims to determine the optimized location of Unified Power Flow Controller (UPFC) device under severe line outages conditions.

The main purpose of the thesis is to develop the hybrid optimization algorithms, which can yield only the feasible solutions of congestion management cost over the entire searching area. To achieve this, two evolutionary chaotic particle swarm optimization algorithms namely; Improved Tent Map Adaptive Chaotic Particle Swarm Optimization (ITM-CPSO) and Twin Extremity Chaotic Mapped Particle Swarm Optimization (TECM-PSO) have been framed. The proposed algorithms; ITM-CPSO and TECM-PSO have been implemented to evaluate congestion management cost and rescheduled generation cost respectively. Required load management has also been incorporated while evaluating overall congestion management cost in order to make the approach more robust, secure and reliable. Furthermore, ITM-CPSO has also been explored for finding suitable locations and parameter settings of Unified Power Flow Controller (UPFC) device for the most severe line outage cases of two test systems.

Till date sensitivity factor based approach has been widely used to decide the number of participating generators for the rescheduling process, but this sensitivity factor based approach requires a rigorous system operator efforts and computational time. To get rid of this, a modified Upstream Real Capacity Tracing (URCT) algorithm has been formulated and implemented in the thesis for deciding the number of participating generators for the rescheduling process. Use of URCT algorithm has an advantage of handling less and specific generator information resulting in less computational efforts and time. Furthermore, implementation of TECM-PSO algorithm to achieve near global optimal solution of rescheduled generation cost function has remarkably enhanced the profile of congestion management in terms of reduced amount of rescheduled generation with high computational efficiency rate and decreased rescheduled generation cost.

The main attribute of proposed algorithms are that these have evolved near global optimum solutions in every independent trial run which remain consistent even for large system as well. Rapid and consistent convergence, improved voltage profile, increased computational proficiency and best quality solutions are the major achievements of the research. Decrease in Net Generation Rescheduled (NGR) using proposed approach lies in the broad scale range of 1.15 to 1.61 and the percentage decrease in Rescheduled Generation Cost (RGC) has been achieved up to 32.4 % for most severe line outage cases. Also the percentage Convergence Mobility Rate (CMR) is found to be the highest among all comparative algorithms for each line outage case.