Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods
The accurate prediction of photovoltaic (PV) power is essential for planning power systems and constructing intelligent grids. However, this has become difficult due to the intermittency and instability of PV power data. This paper introduces a deep learning framework based on 7.5 min-ahead and 15 m...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-12-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/24/9630 |
| _version_ | 1797455384448860160 |
|---|---|
| author | Xianchao Guo Yuchang Mo Ke Yan |
| author_facet | Xianchao Guo Yuchang Mo Ke Yan |
| author_sort | Xianchao Guo |
| collection | DOAJ |
| description | The accurate prediction of photovoltaic (PV) power is essential for planning power systems and constructing intelligent grids. However, this has become difficult due to the intermittency and instability of PV power data. This paper introduces a deep learning framework based on 7.5 min-ahead and 15 min-ahead approaches to predict short-term PV power. Specifically, we propose a hybrid model based on singular spectrum analysis (SSA) and bidirectional long short-term memory (BiLSTM) networks with the Bayesian optimization (BO) algorithm. To begin, the SSA decomposes the PV power series into several sub-signals. Then, the BO algorithm automatically adjusts hyperparameters for the deep neural network architecture. Following that, parallel BiLSTM networks predict the value of each component. Finally, the prediction of the sub-signals is summed to generate the final prediction results. The performance of the proposed model is investigated using two datasets collected from real-world rooftop stations in eastern China. The 7.5 min-ahead predictions generated by the proposed model can reduce up to 380.51% error, and the 15 min-ahead predictions decrease by up to 296.01% error. The experimental results demonstrate the superiority of the proposed model in comparison to other forecasting methods. |
| first_indexed | 2024-03-09T15:52:40Z |
| format | Article |
| id | doaj.art-81ef8c009a524fafacd4cfd0b26d7bed |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T15:52:40Z |
| publishDate | 2022-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-81ef8c009a524fafacd4cfd0b26d7bed2023-11-24T17:52:42ZengMDPI AGSensors1424-82202022-12-012224963010.3390/s22249630Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning MethodsXianchao Guo0Yuchang Mo1Ke Yan2Fujian Province University Key Laboratory of Computational Science, Huaqiao University, Quanzhou 362021, ChinaFujian Province University Key Laboratory of Computational Science, Huaqiao University, Quanzhou 362021, ChinaDepartment of the Built Environment, College of Design and Engineering, National University of Singapore, Singapore 117566, SingaporeThe accurate prediction of photovoltaic (PV) power is essential for planning power systems and constructing intelligent grids. However, this has become difficult due to the intermittency and instability of PV power data. This paper introduces a deep learning framework based on 7.5 min-ahead and 15 min-ahead approaches to predict short-term PV power. Specifically, we propose a hybrid model based on singular spectrum analysis (SSA) and bidirectional long short-term memory (BiLSTM) networks with the Bayesian optimization (BO) algorithm. To begin, the SSA decomposes the PV power series into several sub-signals. Then, the BO algorithm automatically adjusts hyperparameters for the deep neural network architecture. Following that, parallel BiLSTM networks predict the value of each component. Finally, the prediction of the sub-signals is summed to generate the final prediction results. The performance of the proposed model is investigated using two datasets collected from real-world rooftop stations in eastern China. The 7.5 min-ahead predictions generated by the proposed model can reduce up to 380.51% error, and the 15 min-ahead predictions decrease by up to 296.01% error. The experimental results demonstrate the superiority of the proposed model in comparison to other forecasting methods.https://www.mdpi.com/1424-8220/22/24/9630Bayesian optimization algorithmbidirectional long short-term memorydeep learningsingular spectrum analysisshort-term photovoltaic power forecasting |
| spellingShingle | Xianchao Guo Yuchang Mo Ke Yan Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods Sensors Bayesian optimization algorithm bidirectional long short-term memory deep learning singular spectrum analysis short-term photovoltaic power forecasting |
| title | Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods |
| title_full | Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods |
| title_fullStr | Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods |
| title_full_unstemmed | Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods |
| title_short | Short-Term Photovoltaic Power Forecasting Based on Historical Information and Deep Learning Methods |
| title_sort | short term photovoltaic power forecasting based on historical information and deep learning methods |
| topic | Bayesian optimization algorithm bidirectional long short-term memory deep learning singular spectrum analysis short-term photovoltaic power forecasting |
| url | https://www.mdpi.com/1424-8220/22/24/9630 |
| work_keys_str_mv | AT xianchaoguo shorttermphotovoltaicpowerforecastingbasedonhistoricalinformationanddeeplearningmethods AT yuchangmo shorttermphotovoltaicpowerforecastingbasedonhistoricalinformationanddeeplearningmethods AT keyan shorttermphotovoltaicpowerforecastingbasedonhistoricalinformationanddeeplearningmethods |