PREDIKSI DAYA REGENERATIF PADA PENGEREMAN REGENERATIF KENDARAAN LISTRIK MENGGUNAKAN MACHINE LEARNING
Total View This Week0
Total View This Week0
Institusion
Institut Teknologi Perusahaan Listrik Negara
Author
Romy Wicaksono, Irham
Pratama Putera, Rizki
Pratama Putera, Rizki
Subject
Teknik Elektro
Datestamp
2022-11-03 04:10:41
Abstract :
This study conducts a simulation to forecast regenerative power of a regenerative braking using a model that has regenerative braking system. This study proposes analyzing forecasted results of regenerative power using training data obtained by the model and doing validation using test data. Power forecasting simulation conducted by using regression; furthermore, regressions used for power forecasting are SVR (Support Vector Regression) and linear regression. Although this study uses machine learning for power forecasting, the forecasted data have difference results compared to experimental data based on regression metrics. According to power forecasting simulation using SVR, the metrics obtained for ultracapacitor installed in parallel and single ultracapacitor, respectively, are: 1) Coefficient of determination scores 0.69 and 0.85; 2) MAE of 3.68 and 2.09; and 3) MSE of 23.16 and 11.75; In the other hand, according to power forecasting simulation using linear regression, the metrics obtained for ultracapacitor installed in parallel and single ultracapacitor, respectively, are: 1) Coefficient of determination of 0.71 and 0.91; 2) MAE of 3.49 and 1.75; and 3) MSE of 21.83 and 7.28.
This study conducts a simulation to forecast regenerative power of a regenerative braking using a model that has regenerative braking system. This study proposes analyzing forecasted results of regenerative power using training data obtained by the model and doing validation using test data. Power forecasting simulation conducted by using regression; furthermore, regressions used for power forecasting are SVR (Support Vector Regression) and linear regression. Although this study uses machine learning for power forecasting, the forecasted data have difference results compared to experimental data based on regression metrics. According to power forecasting simulation using SVR, the metrics obtained for ultracapacitor installed in parallel and single ultracapacitor, respectively, are: 1) Coefficient of determination scores 0.69 and 0.85; 2) MAE of 3.68 and 2.09; and 3) MSE of 23.16 and 11.75; In the other hand, according to power forecasting simulation using linear regression, the metrics obtained for ultracapacitor installed in parallel and single ultracapacitor, respectively, are: 1) Coefficient of determination of 0.71 and 0.91; 2) MAE of 3.49 and 1.75; and 3) MSE of 21.83 and 7.28.
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
