The Principal contribution of this dissertation lies in developing an efficient optimization framework for distribution system operational and planning studies. The first part of this dissertation introduces a novel power flow model, which is equally appropriate for use at both distribution and transmission levels and can be extremely useful whenever fast, robust, and repetitive power flow solutions are required. We develop the proposed linearized AC power flow model (LACPF) based on linearization of the full set of conventional power flow equations, and therefore includes voltage magnitude solutions and reactive power flows, unlike traditional linearized power flow methods. Further, the model presented in this dissertation is non-iterative, direct, and involves no convergence issues even with ill-condition systems. We test the proposed model on several distribution systems and has found to perform with speed and accuracy appropriate for repetitive solutions.The second part of this dissertation develops an efficient optimization framework to handle several distribution system operational and planning problems. The proposed framework uses linear programming, because linear programming based formulations tend to be flexible, reliable, and faster than their nonlinear counterparts. We consider voltage bounds, reactive power limits, and all shunt elements in the proposed optimization model. We use the proposed optimization framework to solve the problem of optimal sizing and placement of distributed generation. For this particular case, we use loss sensitivity factors and sensitivity analysis to estimate the optimal size and power factor of the candidate distributed generation units. We also perform exhaustive power flow studies to verify the sizes obtained by the proposed method. We demonstrate the effectiveness of the method on several benchmark systems and prove that the method could lead to optimal or near- optimal global solution, which makes the proposed method very suitable to use in several optimal distribution system planning studies.We solve the problem of optimal economic power dispatch of active distribution systems. We propose a piecewise linear model to approximate the current carrying capacities of distribution feeders. The degree of approximation in this model can be improved to the desired level by increasing the number of line segments used, without substantial affect on the main routine and the computational speed. We further develop linear models for cost functions of generating units, loads, and total power losses. We apply methods, which are developed based on nonlinear programming and conventional linear programming to evaluate the effectiveness of the proposed method. We show that the results obtained by the proposed framework correspond closely with those obtained by nonlinear means, while requiring lower computational effort. We describe a method for solving the distribution system reconfiguration problem with an objective of reliability improvement. From practical perspective, distribution systems are reconfigured radially for best control and coordination of their protective devices. Therefore, we develop a graph theoretic method to preserve the spanning tree structure of the distribution system. We further develop an intelligent search method based on binary particle swarm optimization technique, to seek for the best combinations of sectionalizing and tie-switches that minimize the amount of total power curtailment. Since the time and computational effort spent in evaluating reliability indices are of great concern in both planning and operational stages, we propose a probabilistic reliability assessment method based on event tree analysis with higher-order contingency approximation. We demonstrate the effectiveness of the proposed method on numerous radial distribution systems and show that the amount of annual power curtailment of in-service consumers can be tremendously reduced using the proposed method.
Read
