التنبؤ بالطلب على الطاقة الكهربائية باستعمال الشبكة العصبية من نوع GMDH في محطة ديزلات شمال العمارة العراق

Abstract EN

Forecasting plays an important role in the planning and operation of the power system to achieve high productivity in electric power plants.

and because of its important role in scheduling the production processes.

of power systems, and there is a need for accurate forecasting of the electrical load for the safety and reliability of the power system, and given the diversity of demand for energy.

For the power plant under consideration, which does not appear similar in the same period of each month and for the same month for each year, the fact that the plant is considered to enhance the capacity of the national grid in the following main roles (continuous, peak.

and reserve) made it more difficult to predict the demand for future loads with high accuracy, as a self organizing GMDH neural network developed to model complex.

multivariable.

nonlinear systems based on input-output relationships was used to train and test daily energy demand in North Amara Diesel Power Plant/Southern Region/Iraq.

for the period starting from 1/ 1/2019 and ends on 12/31/2021.

as 90% of the data were divided for training and 10% for testing, and the neural network was trained using the neural quadratic function.

and it was used m root mean square error (RMSE) and mean absolute error percentage (MAPE) as performance indicators for testing prediction accuracy.

the neural network model gave the mean absolute percentage error (MAPE) 0.148 and the correlation (R) value of 0.699.

while the regression analysis method gave the mean absolute percentage error (MAPE) 0.203 and the correlation (R) value of 0.395.

the obtained results show the effectiveness of the GMDH-type neural network model in predicting the regression analysis method.