Fuzzy systems application to electric short-term load forecasting. I. Problem formulation

Abstract

Economic secure operation of electric power systems requires good prediction for the electric load twenty four hours ahead. This paper presents the application of fuzzy system to electric short-term load forecasting. Two models are used for modeling the electric load. While in the first model, the load power depends only on the hour on question, the second model is a function of the temperature as well as the hour in question. The problem is formulated as a linear fuzzy regression problem. where the linear programming is used to estimate the middle and spread of each load parameter. A triangular membership function is used, which is symmetrical around the middle. Having identified the fuzzy parameters of the load model, the load can easily be predicted for the next 24 hours. Results for a practical system in operation are offered in the text, for different days of the week, working days and weekend day, where the load power is assumed to be crisp in one test and fuzzy in another test.