Autoregressive Integrate Mean Average as Approaches for Modeling Energy Production in the Electricity Network by Seasons of the Electric Energy Company in Lomé, Togo
- 1 Department of Electrical Engineering ((Polytechnic School of Lomé (EPL)), University of Lomé, Lomé, Togo
- 2 Department of Electrical Engineering (Engineering Sciences Research Laboratory (LARSI), University of Lomé, Lomé, Togo
- 3 Department of Electrical Engineering (Regional Excellence Center for Electricity Management (CERME)), University of Lomé, Lomé, Togo
- 4 African Institute for Mathematical Sciences, Research and Innovation Centre (AIMS RIC), Kigali, Rwanda
Abstract
The work accumulated in this article presents the results of modeling the tduction of electrical energy for the CEET network in Lomé, Togo, taking into account the seasons of the area. The objective is to evaluate the forecasts over the periods in relation to the seasons of the year. Four seasons are detected, namely: Long dry season, long rainy season, short dry season, and short rainy season. The data used comes from the operating values of the aforementioned network during the year 2021. Following a monthly characterization of the data structured in season, ARIMA is used as an algorithm to create models. The latter is subject to performance evaluation metrics such as RMSE, MAE, MSE, and R². The results of the characterization show that in the rainy season, production remains higher than consumption and we have the opposite phenomenon in the dry season. Certain contrary special cases have been observed. Regarding modeling, several results were explored; given the effect that the order of the moving average q had on the matrix, [p d q], of ARIMA. All things considered, the results of the models obtained are very interesting, due to their coefficient of determination, which is greater than 73%. This being said, we have: R² = 89.73%; MSE = 55.35; RMSE = 7.54%; MAE = 4.20, for the long rainy season and: R² = 89.10%; RMSE = 7.55%; MAE = 4.59 and MSE = 56.94, for the two cumulative rainy seasons. Results higher than that of the year which is: R² = 88.06%; MSE = 94.88; RMSE = 9.74%; MAE = 5.4. On the other hand, for the dry seasons, we find results lower than those of the year giving: MAE = 6.06; RMSE = 14.20%; MSE = 201.57; R² = 86.11% for the long dry season; MAE = 6.00; MSE = 159.87; RMSE = 12.64% and R² = 86.55% for the two dry seasons combined. Isolated cases of the models, contrary to the remarks of the characterization, are also observed for the short rainy season with the results: MAE = 6.00; MSE = 159.87; RMSE = 12.64%, and R² = 86.55%. We deduce that this comes from the fairly short period whose data is not consistent for the ARIMA algorithm, confirmed by the results of the two rainy seasons which are: MAE = 4.59; MSE = 56.94; RMSE = 7.55% and R² = 89.10%. To this end, we find the need to extend the data collection period, without forgetting the climatic anomalies which are prevalent all over the world.
DOI: https://doi.org/10.3844/ajeassp.2024.100.126
Copyright: © 2024 Apaloo Bara Komla Kpomonè, Bokovi Yao, Ghafi Kondi-Akara Victoire and Moro Ouma Cecil Naphtaly. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
- 79 Views
- 29 Downloads
- 0 Citations
Download
Keywords
- ARIMA
- Consumption
- Modeling
- Moving Average
- Production
- Season