M.A. Thesis:

Short- Term Load Forecasting Using optimized fuzzy systems with Firefly algoritm

Load forecasting has been a central and comprehensive process in the planning and operation of the electricity industry. Many methods of dealing with this problem have been researched and applied in the last two decades. Predicting the right load, in addition to saving on investment costs, also allows for proper time planning for the project. Conventional methods and algorithms for load prediction can be mentioned in two methods: pso and fireball algorithm. The old method is to predict the burden of using statistical results, but in recent year new methods have been proposed based on artificial intelligence. The purpose of artificial intelligence methods is to use the computer not only as a calculator but also as trained by the load changes due to different conditions in the past and able to find suitable patterns under different conditions and based on the load. The data obtained from the Sabzevar climate are simulated by MATLAB software and the response of all simulated modes is calculated, then the simulated output is plotted to allow better observation and comparison. The results of 8 different modes are applied to two types of fuzzy inference system, 2 types of optimization algorithms and 2 different classifications. The errors are examined in each case since the purpose is to calculate and predict the load. It is important to consider the absolute mean error parameter (MAPE) as the predictive error parameter. Then by comparing the mean absolute error, we can obtain the results of the Sugeno fuzzy inference system and show the desired performance.

Keywords: