Location
Online
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2022 12:00 AM
End Date
7-1-2022 12:00 AM
Description
This paper addresses the slow-onset crisis of global warming caused by CO2 emissions. Although electrical load is a major influence in a country’s growth and development, it is also one of largest sources of greenhouse gases (GHG), CO2 in particular. Therefore, switching to cleaner energy sources is a clear objective and forecasting electricity load and its environmental cost is a necessary task for electrical energy planning and management. This paper addresses short-term load forecasting of renewable energy (RE) production in the region of Adrar in Algeria with Adrar’s photovoltaic (PV) farm and Kabertene’s wind farm. The forecast is compared to the overall load demand, and the reduced amount of CO2 resulting from using renewable energy instead of fossil fuels is calculated. The forecasting models are Long short-term memory (LSTM) neural networks, which were trained and validated using real data provided by the national state-owned company SONALGAZ. The results show good performance for the forecasting models with PV and wind models achieving a Mean-absolute-error (MAE) of 0.024 and 0.1 respectively, and that RE can help reduce CO2 emissions by up to 25% per hour.
Modeling renewable energy production and CO2 emissions in the region of Adrar in Algeria using LSTM neural networks
Online
This paper addresses the slow-onset crisis of global warming caused by CO2 emissions. Although electrical load is a major influence in a country’s growth and development, it is also one of largest sources of greenhouse gases (GHG), CO2 in particular. Therefore, switching to cleaner energy sources is a clear objective and forecasting electricity load and its environmental cost is a necessary task for electrical energy planning and management. This paper addresses short-term load forecasting of renewable energy (RE) production in the region of Adrar in Algeria with Adrar’s photovoltaic (PV) farm and Kabertene’s wind farm. The forecast is compared to the overall load demand, and the reduced amount of CO2 resulting from using renewable energy instead of fossil fuels is calculated. The forecasting models are Long short-term memory (LSTM) neural networks, which were trained and validated using real data provided by the national state-owned company SONALGAZ. The results show good performance for the forecasting models with PV and wind models achieving a Mean-absolute-error (MAE) of 0.024 and 0.1 respectively, and that RE can help reduce CO2 emissions by up to 25% per hour.
https://aisel.aisnet.org/hicss-55/dg/disaster_resilience/10