Reliability improvements and design optimization of mesh distribution networks.

摘要

The dissertation presents a robust and simple tool that can be executed quickly to predict the reliability of a particular network design and different improvement schemes. The proposed tool establishes a process to evaluate different design options and suggest an optimal design based on cost/benefit analysis. The method suggested in this dissertation is based on a multi-parameter cross-sectional-regression. Using regression analysis, a set of equations was established. The equations include number of factors that affects the reliability of the mesh distribution networks. A process of testing and validating each regression is presented, and in some cases, modifications and tuning will be applied to improve the quality of the regression. Finally, an optimization process which includes the reliability equations and a developed simplified cost function is presented with selected examples. The initial results presented in this paper are satisfactory and yield similar results to the one obtained by a simulation process.