Real-time decision support system applied to distribution utility dispatches

Abstract

A distribution utility has to deal with several customer calls regarding grid maintenance or energy issues. Generally, when the proper channels receive a call from the customers, the reported issue pass through a screening phase and, in the end, a maintenance team is sent to the location to solve the problem. However, not all problems are responsibility of the company, generating an unnecessary displacement for the maintenance team, a problem denominated as “improper dispatch”. Improper dispatches generate high costs regarding fuel and logistic. Besides, a high number of improper dispatches can result in heavy penalties to the company since the staff is not available to attend customers that really would need assistance. For tackling this problem, we propose a supervised machine learning solution that uses the customer calls information to classify when a call is improper or not. Our first results indicate that our model achieves up to 80% of assertiveness within a real dataset from the industry. In this work, we show how we built this model, pre-processed the information and, how this solution can be applied to decrease maintenance costs inside an energy company.