SMALL SIGNAL STABILITY CONSTRAINED DISTRIBUTION SYSTEM RECONFIGURATION

Abstract

In modern age, the demand of electricity is ever-increasing due to rapid urbanization and industrialization. To satisfy this ever increasing demand for electricity, power systems are operating close to their operational limits. This situation has led to higher system losses and poor voltage regulation. Studies indicate that approximately 1013 % of the total power generated is lost as I2R losses at the distribution level, which in turn, causes increase in the cost of energy and poor voltage profile along the distribution feeder. Therefore, the requirement for reliable and sufficient power supply is becoming more and more intensive. In order to improve the reliability of a radial distribution system, the configuration of the network should be optimal to maximize the operational benefits. For improving the performance of the distribution system (DS), reconfiguration of DS has been studied and implemented throughout the world for more than two decades. In the face of varying loads, power loss in a distributed network will not be minimum for a fixed network configuration. In addition, due to the rapid expansion of distribution networks, the voltage stability of distribution systems has become an important issue. Hence, there is a need for reconfiguration of the network from time to time. In distribution system reconfiguration (DSR), the topology of the DS is altered by the operation of tie-line and sectionalising switches to achieve operational improvements. Further, distributed generation (DG) will play an important and crucial role in emerging DS. Increasing DG deployment in DS causes a significant impact on the power flow, voltage profile, power losses and stability. However, due to increasing deployment of DG units in a DS, the power flow pattern in a DS can change significantly, which can cause operational problem in an otherwise optimally reconfigured (without any DG) system. Therefore, it is necessary to undertake reconfiguration of a DS in the presence of DGs. The literature published in last one decade has proposed several optimization methods in the presence of DGs for achieving new optimized system configuration. However, in the literature, only steady state behaviour of DGs are included so far. When the system configuration changes (with the DGs installed at some fixed buses), the impedance between the DGs may change which may lead a stable system into unstable mode or vice versa. Therefore, while undertaking DSR in the presence of DGs, stability aspect of DS should also be considered. DSR without stability consideration may lead to an unstable configuration of DS in the presence of renewable energy i sources (RES) based DG such as wind power and solar photovoltaics based DGs. In view of this, the first contribution of this thesis is multiobjective DSR with stability consideration in which the objectives are to reduce the system real power losses and to maximize the voltage stability with minimum number of switching number operations while simultaneously maintaining the dynamic stability of the system. Further, different operating limits of the DS have also been considered in the formulation. Initially, the main emphasis is on the small signal stability of the synchronous machine based DGs under deterministic environment of the distribution system. However, the deterministic nature of the DS may lead to inadequate configuration of DS due to varying load demands and intermittencies associated with DGs. Therefore, the formulated DSR problem has been further solved taking into account the uncertain load demands and generation from renewable energy sources. These uncertain quantities have been represented by appropriate probability density functions (PDF). Further, impact of integration of PV-DGs, small hydro power plant (SHPP) DGs and DFIGs has been considered on reconfiguration. To accurately evaluate the small signal stability of RES-DGs, detailed models of DGs (SHPP-DGs, PV-DGs and DFIGs) have been included in this formulated DSR problem. Further, the generations from SHPP-DGs, PV-DGs and DFIGs are strongly correlated to the generation from the adjacent SHPP-DGs, PV-DGs and DFIGs respectively due to similar water availability, solar irradiance and wind speed at that area. As a result, correlations among RES-DGs have been taken into account in the formulated DSR. To investigate the change in stability with the increased penetration of the DGs, the eigen value analysis of the DS has been carried out. Further, it is a commonly recognized fact that load modeling has a major impact on the DSR. Therefore, proper and adequate load modelling is also required for DSR. In this thesis, voltage dependent loads in the form of load combinations (constant power, constant impedance and constant current loads) have been considered during DSR. Furthermore, a methodology for determining the proper switching sequence has been presented in this thesis. In the literature, different techniques such as mathematical programming based algorithm, heuristic rule-base approach, meta heuristic approach and hybrid methods have been presented to solve the DSR problem. However, in many meta-heuristic approaches, multiobjective DSR problem is converted into single objective optimization problem by using weighting factors. The values of the weighting factors generally depend on the relative importance of the objective functions (as perceived by the system operator). Therefore, tuning of weighting factors are requried for ii every network. To mitigate this problem, a large number of multiobjective evolutionary algorithms (MOEAs) have been proposed in the literature. These techniques find a set of Pareto optimal solutions. Most widely used Pareto optimal solution technique is non dominated genetic algorithm-II (NSGA-II). However, crowding distance calculation as a diversity measure in NSGA-II can be highly time-consuming for more than three objectives. To overcome these difficulties, Knee point driven point algorithm (KnEA) has been utilized for DSR. The main difference between KnEA and other multi-objective EAs (MOEAs) such as NSGA-II is that KnEA uses knee points as a secondary selection criterion additionally with the principle of non-dominated sorting to enhance the search ability of MOEAs. Knee points of non-dominated fronts in the current population are preferred for selection to maximise the hypervolume that helps in maintaining better balance between the convergence of the method and the diversity in the population. In this methodology, an adaptive strategy is utilized for identifying the knee points in the present population without prior knowledge about the number of knee points in the true pareto front to accelerate the convergence and promote diversity.