DIGITAL LOAD -FLOW STUDIES IN POWER SYSTEMS

Abstract

A predetermination •of the distribution of power and voltage in an integrated system of gonoraorsy transmission and distribution circuits, and load is usually an essential requirement in design and operation studies of power systems. Digital commuter techniques of power system analysis have been extensively developed in last two decades. Many methods of load-flow solutions using different algorithms and techniques have boon applied by the power engineers. The important ones among them arc Gauss-Soidel, Norton-Raphson, Modified Newton, Decomposition techniques and non-linear programming methods.. Those have been applied for load flow studios with model voltage analysis, particularly in its admittance form. Now a stage , has boon reached in those applications, T,ncoii tho relative merits nay be closely analysed. Gauss-Seidol method has been very popular Ur_°i th power engineers and has been used accommodating up to 1000 buses 2000 line sections, and. 250 transformers; but this method has the prdblon of convergence. The Novrbon-Raphson technique is faster in eonvorg once, but suffers iAth the drawback of doubtful convorgonce, Modifiod Nowtont s r_Zothod for lord-flow through a atnil_,isaticon techniquo(76) is based on non-linear programming approach and uses a control of convergence process and this method converges for the cases) when the normal. Nowrtont s method diverges. Doconoo si tion technique is a good tool for piece meal solution of large interconnected systems on small computers and is becoming iv popular in the country because of availability of large number of small corrput ers. • The dovolopu.=nt of progranrlin ^acilitics making use of Sr arsi'.y f u J "'..t -ix and compu4,ation .? meth-~ds like trirng u lions ware .rc )t method, pivnta3. condensation and decomposition for solving load flows has helped in faster convergence and in reducing the requirements on the computer storage. In this di ssertati^n the author has attempted to have load-flow studies of a 6-bus sample system by Gauss -Sei:l, No won, Modified Newton & decomposition methods to bring out a comparison of those meta , : with respect to computer storage requirements, computer running timo4conver-gence characteristics, acceleration factors etc* In the conclusion the relative merits and demerits of those methods have been brought out for their >ractic21 application in power system studios, A ,ractical system of 31 buses has boon solved for load-flows in detail b :% Gauss-Seidel method and the results have been discussed in Chapter IV of the dissertation.