The fuzzy correlation approach in operation of electrical distribution systems

Abstract

Purpose

The purpose of this paper is to introduce a method based on the fuzzy correlation for modelling of active and reactive powers from the substations of the electrical distribution systems, at the peak load.

Design/methodology/approach

Based on the correlation theory, the fuzzy models of the loads can be obtained using a new algorithm. If in the case of the principal/connection station there is sufficient database information for a good forecasting of the load, then for those substations where data are missing (there is no continuous monitoring or the measuring system can be broken for a while) the forecasting of the load can be performed using the correlation studies. The starting point of the algorithm is statistical analysis of the active and reactive curves of the substations and utilization of a fuzzy linear regression model. This can be made for different time windows (window 24 h, window 7 h, etc). The window 24 h can be used successfully to estimate the hourly load on any substation. The other time window (7 h) can be used in the peak load estimation of the substations, using the maximum value of the active power recorded in a reference substation.

Findings

The numerical data show that the fuzzy correlation models can be used with very good results for determination of the peak load corresponding distribution substations, and further with the state estimation of the system. In this study, the influence of the time window size is presented in detail, and the fuzzy correlation models for the peak loads from the distribution substations are obtained.

Originality/value

Starting from the correlation theory, a method of fuzzy modelling of active and reactive powers from the substations of an electrical distribution system is proposed.

Citation

Grigoras, G. and Cartina, G. (2013), "The fuzzy correlation approach in operation of electrical distribution systems", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 32 No. 3, pp. 1044-1066. https://doi.org/10.1108/03321641311306060