M.Sc. Tezi Görüntüleme | |||||||||||||||||||||
|
|
||||||||||||||||||||
Summary: In this study, an artificial neural network (ANN) and ANFIS model was established to estimate the electricity market clearing price (MCL). Production quantities of primaryenergy sources are used as input variables of the model. The input variables are modeled by creating three categories. In the first category, the amount of production of each primaryenergy source separately, in the second category, compounds of primary energy sources are formed and their production quantities, in the third category, new partial MCLs are found bymultiplying each primary energy source of MCLs obtained by modifying each primary energy source by the ratio of total production and a new MCL value is obtained by collectingeach partially obtained MCL. When these results are examined it appears that there is a close connection between MCL and primary energy sources. With hydraulic primary energysource, the right natural gas is inversely proportional to the primary energy source. When the results of our models were compared, the most successful results were obtained withinthe third category and MAPE error rates were found as 3.74% for ANFIS and 5.80% for ANN. The fact that the results are close to reality will help to obtain more accurate resultsin the electricity market clearing price forecasting models based on primary energy sources. The accuracy of the results obtained from the model depends on the availability of primaryenergy sources. Because of the resulting supply shortage or some primary energy resource shortage, it has been shown that excessive increases in production adversely affect themodel, due to the fact that the other primary energy sources are being replaced. |