| dc.creator |
Mwifunyi, Rukia J. |
|
| dc.creator |
Kissaka, Mussa M. |
|
| dc.creator |
Mvungi, Nerey H. |
|
| dc.date |
2020-09-07T09:31:46Z |
|
| dc.date |
2020-09-07T09:31:46Z |
|
| dc.date |
2020 |
|
| dc.date.accessioned |
2022-10-20T13:45:32Z |
|
| dc.date.available |
2022-10-20T13:45:32Z |
|
| dc.identifier |
Mwifunyi, R. J., Kissaka, M. M. & Mvungi, N. H. (2020). Short-term load forecasting for improved service restoration in electrical power systems: A case of Tanzania. In 2020 International Conference on Artificial Intelligence and Signal Processing (AISP) (pp. 1-5). IEEE. |
|
| dc.identifier |
DOI: 10.1109/AISP48273.2020.9073069 |
|
| dc.identifier |
http://hdl.handle.net/20.500.12661/2474 |
|
| dc.identifier.uri |
http://hdl.handle.net/20.500.12661/2474 |
|
| dc.description |
Abstract. The full-text article is available at https://ieeexplore.ieee.org/document/9073069 |
|
| dc.description |
Reliable operation of the power system and efficient utilization of its resources requires load demand forecasting in a wide range of time leads, from minutes to several days. Underestimation of load demand forces the power system to operate in a vulnerable region to the disturbance. In the Tanzanian electrical power distribution network, peak hour load demand values are used during service restoration resulting in prolonged load shedding. This study aims at developing a short-term load forecasting model to be used during service restoration for improved service reliability. Several methods have been devised for short-term load forecasting including conventional statistical approaches and data-driven approaches. Data-driven approaches perform well in load forecasting due to its ability in learning features for the dataset with nonlinear characteristics like load demand dataset. The study has adopted an experimental design approach in developing the short-term load foresting model using six years datasets from 2014 to 2019 with twenty minutes resolution from the Tanzania power distribution network. A total of 141,749 datasets were used and three deep learning models namely Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN) and Gated Recurrent Unit (GRU) were used during the experiments. It has been observed that the LSTM outperforms the RNN and GRU with forecasting accuracy of 96.43%. The future work will be the development of a distributed algorithm for service restoration considering stochastic nature of load demand using developed forecasting load model. |
|
| dc.language |
en |
|
| dc.publisher |
IEEE |
|
| dc.subject |
Forecasting |
|
| dc.subject |
Service restoration |
|
| dc.subject |
Load forecasting |
|
| dc.subject |
Load forecasting model |
|
| dc.subject |
Power systems |
|
| dc.subject |
Recurrent neural networks |
|
| dc.subject |
Load demand forecasting |
|
| dc.subject |
Demand forecasting |
|
| dc.subject |
Power distribution network |
|
| dc.subject |
Tanzania power distribution network |
|
| dc.subject |
RNN |
|
| dc.title |
Short-term load forecasting for improved service restoration in electrical power systems: A case of Tanzania |
|
| dc.title |
In proceeding of the 2020 International conference on Artificial Intelligence and Signal Processing (AISP) |
|
| dc.type |
Conference Proceedings |
|