On Some Aspects of Distribution Load Flow

This paper presents a comprehensive analysis and comparisons of the performances of the two commonly used models under different situations for distribution load flow. Investigations reveal that the computational efficiencies of both the models are more or less close over a wide range of loading condition. However, model 2 is less sensitive to model 1 for increase in loading conditions, reduction in load powerfactor and increase in the branch R/X ratios. INTRODUCTION Newton Raphson and Fast Decoupled powerflow solution techniques have efficiently solved well-behaved power systems for more than hvo decades. However, these are in general unsuitable for solving load flow for distribution networks because of high R/X ratios and radial nature of distribution networks. Different techniques have been developed to solve such distribution networks exploiting their radial nature. Literature survey shows that the two most popularly used techniques for distribution load flow (DISTFLOW) are, one based on a Newton like method (model 1) involving formation of jacobians and computation of power mismatches at the end of the feeder and laterals [11 and the other based on backward and forward sweeps (model 2) involving computation of current flows in the branches 121. The said models still date have not been compared to establish the superior of the two for practical application. in view of this a pioneering attempt has been made in this paper to examine both these models M. S. SRINIVAS Engineer-R&D, iTAPS CMC CentrqGachbowIi Hyderabad-500 019(INDIA) more critically. It is worth investigating the performance of each model to changes in system operating conditions like changes in system loading, changes in power factors and changes in R/x ratios of network branches etc., so as to explore the efficiency and effectiveness of each of these models. All these acpects have not been explored till date. In view of the above the following problems have been investigated. a. To study the convergence property of model 1 for different loading conditions, power factors and branch WX ratios. b. To study the convergence property of model 2 for different loading conditions, power factors and branch R/X ratios. c. To examine the computational involvement (number of arithmetic operations) for both the models, compare their performances and hence determine the better of the two models for practical, applications. Model 1 Consider a distribution system consisting of a radial main feeder having n buses, n-l branches and one lateral emerging from the i' bus shown in Figure 1. s.s : Substation VI : Substation voltage S.S vr i+l r"Branch|(i+U l(n-l) In Sn=0 Figure One line diagrain of a distribution feeder,