EFFECT OF PARALLELISM IN CALCULATING THE EXECUTION TIME DURING FORECASTING ELECTRICAL LOAD

Abstract

The electrical load is of great economic importance to electric power industry. Load information is of a great technical importance for a stable electricity system as a secure supply could be guaranteed. The present paper aims at finding the effect of parallel computing in terms of execution time of the scripts used for the mathematical models of load profiles to address the forecast of electrical load. A parallel system based on the sequential and parallel execution of the scripts by splitting input data was used to investigate the effect of parallelism. The mathematical pattern was derived from the regression model with several independent variables and represent the dependence of electrical load upon time, temperature and humidity. Regression analysis involved the programming environment R. In addition, the data have been analyzed for a clearer picture of data distribution.