Representative Days for Expansion Decisions in Power Systems

摘要

Short-term uncertainty needs to be properly modeled when analyzing a planning problem in a power system. Since the use of all available historical data may lead to problems of computational intractability, clustering algorithms may be applied in order to reduce the computational effort without compromising accurate representation of historical data. In this paper, we propose a modified version of the traditional K-means method, seeking to represent the maximum and minimum values of input data, namely, electricity demand and renewable production in several locations of a power system. Extreme values of these parameters must be represented as they are high-impact decisions that are taken with respect to expansion and operation. The method proposed is based on the K-means algorithm, which represents the correlation between demand and wind-power production. The chronology of historical data, which influences the performance of some technologies, is characterized through representative days, each made up of 24 operating conditions. A realistic case study, applying representative days, analyzes the generation and transmission expansion planning of the IEEE 24-bus Reliability Test System. Results show that the proposed method is preferable to the traditional K-means technique.