Prediction of Hourly Generated Electric Power Using Artificial Neural Network for Combined Cycle Power Plant

Energy is one the important subjects in the world because of its cost and achievable. In order to reduce energy costs, some kinds of plants may have founded and are managed related to demands and environmental conditions. Artificial neural network is one of the famous artificial intelligent in literature to solve nonlinear problems from medical to constructions. Artificial neural network uses nodes to weights to achieve the output. In this study, obtainable power per hour from combined gas and steam turbine power plant tries to be predicted. Data include 9685 features and 4 variables. Artificial neural network results have been evaluated with mean square error and two fold cross validation. Mean square error and two-fold cross validation are statistical evaluation methods to evaluate the results. Dataset divided 2 sections to test and train. Two datasets are trained and tested using two fold cross validation and generated R value to evaluate the fitting performance. R is famous comparing method to figure out the fitting ability. The obtained mean square error after two fold cross validation and R value are 3.176 and 0.96675, respectively. 