Very Short-Term Solar Power Forecasting Using Hybrid LSTM-SVM

Tuba Tanyıldızı Ağır

Makale detayları

Very Short-Term Solar Power Forecasting Using Hybrid LSTM-SVM

Yazarlar:

Tuba Tanyıldızı Ağır

Yayıncı:

Bitlis Eren University

Ana Başlıklar:

Photovoltaic Power Systems

Dergi:

Bitlis Eren Üniversitesi Fen Bilimleri Dergisi Vol.14 Issue 2, pp.677-696

Belge Tipi:

Makale

Yayın Tarihi:

6/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001351269

pISSN:

2147-3129

Anahtar kelimeler:

Özet:

 Solar energy is one of the most preferred energy sources among renewable energy sources. Very short-term power forecasting has an important role in the voltage and frequency control of solar energy. However, it provides stability to energy by correcting energy fluctuations in the energy market. In this study, long short term memory LSTM, support vector machines SVM and hybrid LSTM-SVM model were used to estimate PV power in the very short term. The inputs of the models were 60-minute pressure, humidity, temperature, cloudiness and wind speed of Şanlıurfa province in 2022.At the output of the models, the 60-minute power value of the PV panel was obtained. The performances of hybrid LSTM-SVM, LSTM and SVM were compared using mean square error MSE, root mean square error RMSE, normalized root mean square error NRMSE, mean absolute error MAE and correlation coefficient R. In the very short term, PV panel power Hybrid LSTM-SVM, SVM, and LSTM predicted 0.9649, 0.8836 and 0.7255, respectively. The proposed hybrid LSTM-SVM model outperformed the classical LSTM and SVM. The performance metrics of the hybrid LSTM-SVM model, MSE, RMSE, NRMSE, MAE and R, were 9.0098e-04, 0.0300, 0.0318, 0.011 and 0.9823, respectively. The hybrid LSTM-SVM model had high stability and accuracy in very short-term solar power forecasting. Hybrid LSTM-SVM can be used as an alternative method for very short-term solar power forecasting.

 Solar energy is one of the most preferred energy sources among renewable energy sources. Very short-term power forecasting has an important role in the voltage and frequency control of solar energy. However, it provides stability to energy by correcting energy fluctuations in the energy market. In this study, long short term memory LSTM, support vector machines SVM and hybrid LSTM-SVM model were used to estimate PV power in the very short term. The inputs of the models were 60-minute pressure, humidity, temperature, cloudiness and wind speed of Şanlıurfa province in 2022.At the output of the models, the 60-minute power value of the PV panel was obtained. The performances of hybrid LSTM-SVM, LSTM and SVM were compared using mean square error MSE, root mean square error RMSE, normalized root mean square error NRMSE, mean absolute error MAE and correlation coefficient R. In the very short term, PV panel power Hybrid LSTM-SVM, SVM, and LSTM predicted 0.9649, 0.8836 and 0.7255, respectively. The proposed hybrid LSTM-SVM model outperformed the classical LSTM and SVM. The performance metrics of the hybrid LSTM-SVM model, MSE, RMSE, NRMSE, MAE and R, were 9.0098e-04, 0.0300, 0.0318, 0.011 and 0.9823, respectively. The hybrid LSTM-SVM model had high stability and accuracy in very short-term solar power forecasting. Hybrid LSTM-SVM can be used as an alternative method for very short-term solar power forecasting.

 Solar energy is one of the most preferred energy sources among renewable energy sources. Very short-term power forecasting has an important role in the voltage and frequency control of solar energy. However, it provides stability to energy by correcting energy fluctuations in the energy market. In this study, long short term memory LSTM, support vector machines SVM and hybrid LSTM-SVM model were used to estimate PV power in the very short term. The inputs of the models were 60-minute pressure, humidity, temperature, cloudiness and wind speed of Şanlıurfa province in 2022.At the output of the models, the 60-minute power value of the PV panel was obtained. The performances of hybrid LSTM-SVM, LSTM and SVM were compared using mean square error MSE, root mean square error RMSE, normalized root mean square error NRMSE, mean absolute error MAE and correlation coefficient R. In the very short term, PV panel power Hybrid LSTM-SVM, SVM, and LSTM predicted 0.9649, 0.8836 and 0.7255, respectively. The proposed hybrid LSTM-SVM model outperformed the classical LSTM and SVM. The performance metrics of the hybrid LSTM-SVM model, MSE, RMSE, NRMSE, MAE and R, were 9.0098e-04, 0.0300, 0.0318, 0.011 and 0.9823, respectively. The hybrid LSTM-SVM model had high stability and accuracy in very short-term solar power forecasting. Hybrid LSTM-SVM can be used as an alternative method for very short-term solar power forecasting.

 Solar energy is one of the most preferred energy sources among renewable energy sources. Very short-term power forecasting has an important role in the voltage and frequency control of solar energy. However, it provides stability to energy by correcting energy fluctuations in the energy market. In this study, long short term memory LSTM, support vector machines SVM and hybrid LSTM-SVM model were used to estimate PV power in the very short term. The inputs of the models were 60-minute pressure, humidity, temperature, cloudiness and wind speed of Şanlıurfa province in 2022.At the output of the models, the 60-minute power value of the PV panel was obtained. The performances of hybrid LSTM-SVM, LSTM and SVM were compared using mean square error MSE, root mean square error RMSE, normalized root mean square error NRMSE, mean absolute error MAE and correlation coefficient R. In the very short term, PV panel power Hybrid LSTM-SVM, SVM, and LSTM predicted 0.9649, 0.8836 and 0.7255, respectively. The proposed hybrid LSTM-SVM model outperformed the classical LSTM and SVM. The performance metrics of the hybrid LSTM-SVM model, MSE, RMSE, NRMSE, MAE and R, were 9.0098e-04, 0.0300, 0.0318, 0.011 and 0.9823, respectively. The hybrid LSTM-SVM model had high stability and accuracy in very short-term solar power forecasting. Hybrid LSTM-SVM can be used as an alternative method for very short-term solar power forecasting.

Makale detayları

Very Short-Term Solar Power Forecasting Using Hybrid LSTM-SVM

Yazarlar:

Tuba Tanyıldızı Ağır

Yayıncı:

Bitlis Eren University

Ana Başlıklar:

Photovoltaic Power Systems

Dergi:

Bitlis Eren Üniversitesi Fen Bilimleri Dergisi Vol.14 Issue 2, pp.677-696

Belge Tipi:

Makale

Yayın Tarihi:

6/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001351269

pISSN:

2147-3129

Photovoltaic panel

Very-short-term energy forecast

LSTM

SVM

Hybrid LSTM-SVM

Sayı

Makale detayları

Very Short-Term Solar Power Forecasting Using Hybrid LSTM-SVM

Yazarlar: Tuba Tanyıldızı Ağır

Yayıncı: Bitlis Eren University

Ana Başlıklar: Photovoltaic Power Systems

Dergi: Bitlis Eren Üniversitesi Fen Bilimleri Dergisi Vol.14 Issue 2, pp.677-696

Belge Tipi: Makale

Yayın Tarihi: 6/2025

Belge Dili: İngilizce

Document ID (Org No.): 2001351269

pISSN: 2147-3129

Photovoltaic panel

Very-short-term energy forecast

LSTM

SVM

Hybrid LSTM-SVM

Sayı

Yazarlar:

Tuba Tanyıldızı Ağır

Yayıncı:

Bitlis Eren University

Ana Başlıklar:

Photovoltaic Power Systems

Dergi:

Bitlis Eren Üniversitesi Fen Bilimleri Dergisi Vol.14 Issue 2, pp.677-696

Belge Tipi:

Makale

Yayın Tarihi:

6/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001351269

pISSN:

2147-3129