A Reasoning-Based Hyper-Heuristic Model for Reliability Optimization in Medium Voltage Power Distribution Systems

In this work a new Hyper-heuristic Model based on reasoning to solve the Reliability Optimization problem in Medium Voltage Electrical Distribution Systems (MV PDS) is presented. The problem focuses on the optimal location of various types of sectioning /protection equipment (s/p), introducing a new economic assessment for the Cost of Failure, considering not only the Expected Non-Supplied Energy (EENS) but other Reliability Indices based on Customers. If only a single type of s/p equipment were considered, the problem can be solved by applying Possibilistic Dynamic Programming (PDP). However, this approach collapses when there are several types of s/p devices. The solution using single Meta-heuristics has convergence drawbacks, prematurely generating sub-optimal solutions. For this reason, a Model based on a Multi-objective Hyper-heuristic Algorithm is proposed as a general solution, applying it to a real MV PDS to analyze and discuss the results obtained. (Abstract)