Optimal Load Shedding for Voltage Stability Enhancement by Genetic Algorithm

Outage of a heavily loaded transmission line or tripping of large generating unit may lead the system toward collapse. Under such circumstances, compared to other measures, load shedding procedure is an efficient method to make the power system voltage stable when many system variables are out of their normal ranges and the system is driven toward collapse. This paper introduces the concept of the continuation power flow analysis to be used for load shedding using power world simulator.The appropriate load buses for the shedding are identified by sensitivities of voltage stability margin using P-V curves at different buses. Then, the amount of load shedding at each bus is determined by applying GA to solve a nonlinear optimization problem formulated in the optimal power flow framework. It uses the P-V curves to find the knee point of a certain bus. The proposed approach has been tested and examined on IEEE-30 bus test system. Keywordsload shedding; voltage stability; genetic algorithm; P-V curves. Introduction In case of emergency, which may occur in an electric power system as a result of a sudden increase in system load demand or unexpected outage of a generator or other equipment, the system frequency will change. Certain control actions are performed in order to prevent the deterioration of the system and to restore it back to normal state. Load shedding is defined as the set of controls, which results in a decrease of load in the power system in order to reach a new equilibrium state. Different techniques have been proposed to solve the load shedding problem in either the dynamic or steady state cases. Generally, there are two ways to provide voltage stability, which are classified as preventive and corrective actions. In the first approach, the security margin is estimated with respect to credible contingencies with a reasonable probability of occurrence, and then appropriate preventive actions are taken by re-adjusting the most effective controls to provide a sufficient margin when needed. Corrective control actions, on the other hand, are usually used for correction of security, acceptable only in the presence of severe disturbances. Load shedding may be needed if operating conditions isolate some constraints and no control action is available. According to the classification of power system states (i.e. normal, alert, emergency, extreme emergency and restorative), load shedding would be allowed under the emergency and extreme emergency states, when many system variables are out of their normal ranges, and hence the system is driven toward collapse [5]. The load-shedding schemes proposed so far can be classified into three categories. In the first group, the amount of load to be shed is fixed a priori . This scheme is similar to the under-frequency load shedding scheme. Here, the minimum amount of load to be shed is determined using time simulation analysis, incorporating dynamic aspects of the instability phenomenon [8]. Obviously, dynamic simulation is time-consuming and is suitable for special cases such as transient voltage-instability analysis. In addition, it is more difficult to incorporate a time simulation study into an optimization model. The second group tries to determine a minimum load for shedding by estimating dynamic load parameters. In this approach, results are very sensitive to dynamic load model parameters. Finally, in the third group, minimum load shedding is determined using optimal powerflow equations based on a static model of the power system. The dynamics associated with voltage stability are often slow, and hence static approaches may represent a good approximation. The basic idea behind this approach is to identify a feasible solution to the power-flow equations [9-12]. THE PROPOSED METHODOLOGY Voltage instability is generally triggered by either of two types of system disturbances: component outage and load increase. Such disturbances increase the reactive power demand of the transmission network. Outage of a heavily loaded transmission line or tripping of a large generating unit may lead the system toward collapse. Under such circumstances, load shedding is usually initiated after exhausting all other countermeasures in an attempt to arrest a voltage collapse condition. Usually, computation of a minimum load to be shed is carried out through an OPF framework. In this approach, the main objective is “interruption cost minimization”, while voltage stability refers to voltage and transfer limits. However, such an approach cannot guarantee sufficient margin to the collapse point. Here, we attempt to develop a structure to cover these flaws. The main objective is modified to consider both the technical and economic aspects of each load. A loading margin is used to ensure voltage stability. For a particular operating point, the amount of additional load in a specific pattern of load increase that would cause a voltage collapse is called the loading margin. To ensure selection of the most effective loads, we incorporate first-order sensitivity factors of the load margin with respect to active and reactive loads into the objective function. These factors are calculated at the saddle node bifurcation point [14]. To ensure voltage stability, the loading margin is considered as a soft constraint into the model. Using this indicator, the operator ensures that reactive power is provided locally. The overall aim of load shedding is depicted in Fig. 1. Suppose that the system is normally operated at point 1. Following the occurrence of a contingency, the P–V curve changes in such a way that the new margin becomes unsafe, although both voltage and transfer limits are allowable. The aim of load shedding is to International Journal of Applied Engineering Research, ISSN 0973-4562 Vol.7 No.11 (2012) © Research India Publications; http://www.ripublication.com/ijaer.htm readjust the initial operating point to provide a sufficient margin (i.e. moving back P0 (point 2) to P0 −_Pshed (point 3)). Fig. 1 The load shedding scheme. A flow chart of the proposed optimal load shedding is shown in Fig. 2. According to this procedure, after occurrence of a contingency, the loading margin and its sensitivities are calculated by a continuation power-flow method. Under such a condition, when this margin is less than a predefined level (λmin), the power system is voltage-unstable. Fig. 2 Flow chart of the load-shedding procedure. In this situation, load shedding is triggered if the other controls are exhausted. To identify a more sensitive area, the sensitivity of the loading margin with respect to active and reactive power is calculated at each bus (∂λ/∂P). GENETIC ALGORITHM Genetic Algorithms have recently received much attention as robust stochastic search algorithms for optimization problems. GAs are blind search technique using stochastic operations based on the mechanics of the survival of the fittest. It also works with a population of individuals rather than single point. Operation, involve random number generation (mutation), string copying (reproduction), and partial string exchange (crossover). Each string represents a possible solution. It starts with the formulation of the “fitness function”, which represents the objective function for the problem. Based on the fitness of the population strings, two parent strings are selected probabilistically in the reproduction process. Two child strings are then generated from the parent strings in the process of crossover by complementing the child strings at selected bit positions. Mutation is then applied on Occurrence of Severe Contingency Calculating of loading margin (λ) and its sensitivities at “SNB” Voltage stability met? 1) Operational constraints are met? 2) λ > λmin? Run Load Shedding as follow: