ADVANCEMENTS IN LOAD FLOW TECHNIQUES FOR RADIAL DISTRIBUTION SYSTEMS

Abstract

This thesis documents a robust, efficient and fast load flow method for both the balanced and unbalanced distribution systems. The balanced load flow method developed is based on modeling a radial distribution system as a series of interconnected laterals whereby each lateral is assigned a level. The radial topology of distribution networks has been fully exploited such that a unique branch and node numbering scheme was adopted. Using Kirchoff's laws, a set of iterative load flow equations was developed to conduct the load flow studies. The algorithm consists of sub-iterations performed on each lateral from the highest level of laterals to the lowest level that is the main feeder. Based on this method, another load flow algorithm was developed for weakly meshed distribution networks. This method converts the meshed network into a radial network by having breakpoints in the loops and then solving the network using the radial load flow method. It uses a multi-port compensation iterative technique to calculate the breakpoint injection currents. The load dependency on the bus voltage has also been incorporated in the load flow method. In addition, a fuzzy expert system containing a set of heuristic rules was developed to estimate the load modeling voltage exponents ? and ? for each bus in the distribution system. The validity of the load flow method was verified by comparing it with the existing methods of load flow. The performance of the new load flow method was also found to be much more superior to that of the Newton Raphson method. An analysis of the convergence characteristic of the load flow technique was carried out. It was found that the convergence of the method is not affected by the number of buses or the RIX ratio of the conductors but is adversely affected by the load sensitivity to bus voltage, the level of loading and the number of levels in the distribution system. In view of these shortcomings, a modification was made to this load flow method such that the sub-iterations are performed from the main lateral to the highest level of laterals. At each iteration on the entire network, a backward sweep is done from the highest level of laterals to the main lateral to calculate the total lateral load drawn by each lateral. It was found that there is up to 42% improvement in the speed of computation. Furthermore, the convergence characteristic of the modified method is less dependent on the voltage sensitivity of the loads. It is also less susceptible to the loading conditions as well as the number of levels that a distribution system has. Due to the high rate of convergence, this modified method is used in several applications, such as, a fast approximate calculation for voltage and losses, constrained load flow and fuzzy load flow. The method was also used to conduct a study on the Conservative Voltage Reduction and it was found that there is a definite demand reduction associated with the voltage reduction whereas the reduction in losses would be dependent on the load composition. Lastly, a three-phase load flow method based on the modified method for the analysis of unbalanced radial distribution systems was developed. This method decouples the three-phase unbalanced system into three single-phase systems. The mutual coupling between the phase conductors in a section is modeled by connecting current-dependent voltage sources in series with the self-impedances. The validity of the method is also verified by comparing it with another three-phase load flow program. A comparison of performance is made and the new load flow method developed is found to be more superior than the latter.