Evaluation of Energy Distribution Using Network Data Envelopment Analysis and Kohonen Self Organizing Maps

Abstract

This article presents an alternative way of evaluating the efficiency of the electric distribution companies in Brazil. This assessment is currently performed and designed by the National Electric Energy Agency (ANEEL), a Brazilian regulatory agency, to regulate energy prices. This involves calculating the X-factor, which represents the efficiency evolution in the price-cap regulation model. The proposed model aims to use a network Data Envelopment Analysis (DEA) model with the network dimension as an intermediate variable and to use Kohonen Self-Organizing Maps (SOM) to correct the difficulties presented by environmental variables. In order to find which environmental variables influence the efficiency, factor analysis was used to reduce the dimensionality of the model. The analysis still uses multiple regression with the previous efficiency as the dependent variable and the four factors extracted from factor analysis as independent variables. The SOM generated four clusters based on the environment and the efficiency for each distributor in each group. This allows for a better evaluation of the correction in the X-factor, since it can be conducted inside each cluster with a maintained margin for comparison. It is expected that the use of this model will reduce the margin of questioning by distributors about the evaluation.

Open Data and Energy Analytics

“…Electricity data can also support a better evaluation of the distributors' performance, as described by Ganhadeiro et al [5] in a case study in Brazil. The authors propose an improved methodology to better assess how environmental variables affect the energy efficiency of electricity distribution companies.…”

An Improved SOM-Based Method for Multi-Robot Task Assignment and Cooperative Search in Unknown Dynamic Environments

“…The self-organizing map 2 of 18 (SOM) is an unsupervised neural network for data visualization that was first proposed by Kohonen [9]. Some SOM-based methods have been used for energy management, such as battery sorting [10], evaluation of energy distribution [11], efficient utilization of resources [12], etc. Some SOM-based approaches have been proposed to resolve the task assignment for multi-robot systems [13][14][15].…”

Enhancement of a Short-Term Forecasting Method Based on Clustering and kNN: Application to an Industrial Facility Powered by a Cogenerator

“…In another study, a methodology based on energy performance certification is defined to estimate building energy demand using machine learning (decision tree, SVM, random forest, and ANN) [14]. Ganhadeiro et al evaluates the efficiency of the electric distribution companies using self-organizing maps [15]. Machine learning methods are implemented in different environments: MATLAB [16][17][18] and R [19,20] are the most famous ones for research.…”