Forecasting the electricity generation of photovoltaic plants






Науковий ступінь

Рівень дисертації

Шифр та назва спеціальності

Рада захисту

Установа захисту

Науковий керівник

Члени комітету



It was calculated the table of dependence of RMSE, MAE and NMAE errors on the number of input membership functions and their forms for training, test and the whole sample. The number of input membership functions varied from 2 to 30, and the type of membership functions was chosen from the set (MATLAB notation): gbellmf, gaussmf, gauss2mf, trimf, trapmf, dsigmf, psigmf, pimf. To determine the optimal number and form of membership functions, the NMAE error of the test sample was used, which was calculated to be in the range from 3.92% to 4.15%. A minimum was achieved on 5 triangular trimf membership functions. The model was tested for sensitivity to the error of the input data using the chosen optimal number and the form of membership functions. A generated random sequence was added to the input data, which added an NMAE error of 1.81% to 8.19% to the input data. The NMAE error of the initial data in the test sample varied from 4.19% to 5.78%, i.e. the model studies showed a sufficiently low level of variation in the output values relative to the error of the input data.


Ключові слова

ANFIS, RES, forecast, ANN, прогноз, кафедра електропостачання і енергоменеджменту

Бібліографічний опис

Forecasting the electricity generation of photovoltaic plants / Iu. Kuevda, S. Baliuta, P. Zinkevich, O. Stoliarov // Actual problems of renewable energy, construction and environmental engineering : VI International Scientific-Technical Conference, 24-27 November 2022, Kielce (Poland, Ukraine, Croatia, Slovakia, Lithuania). – Kielce, 2022. – Pp. 37-38