Optimal Design of Distribution Overhead Powerlines Using Genetic Algorithms

Abstract

As the design of distribution overhead power lines becomes increasingly standardized in the 21st century, the reduced search space of its design parameters allows for increasing opportunities to automate the design process. It involves specifying the distribution utility pole heights and classes, pole-top attachments, and conductor span tensions. A successful algorithm must be able of generating designs that are compliant with applicable codes and utility standards, achieve construction labour and material costs that are commensurate to that of a human-created design, and require a reasonable computation time. A design automation algorithm is created for ATCO Electricity, Alberta, Canada. Based on provided requirements and constraints, it uses the genetic algorithm to optimize the design parameters, including a completed design staking list, material loading file, and relevant calculation reports for use by the design department. Evaluation of the developed tool is based on five samples of real design scenarios. Within a reasonable amount of computation time, the tool produces results free from major design errors, comparable in material and construction costs to those of human-created designs. Furthermore, the design automation tool makes rigorous use of design decisions that result in small cost savings but that are not commonly found in human-created designs.