Scientists in Malaysia have developed a new deplete method for mapping PV fitness. By applying the new approach in the Middle East, they discovered that around 5.8% of the region has a very high suitability and 11.5% is very suitable for the development of PV energy.
A research team led by the National University of Malaysia has developed a new deplete method for PV fitness mapping.
“The novelty of our research lies in the development of a fully optimized and interpretable deep learning -pipeline, combining tabet, optuna and shap for mapping Solar PV site,” said the corresponding author Khairul Nizam Abdul Maulud PV -Magazine. “This is the first study to apply this integrated approach to the regional scale in the Middle East, use Real-World PV implementation data in addition to Environmental, Social and Governance (ESG)-on the technomic variables-in techno-economic variables.
The first step when making their new model was to identify the suitability criteria. Based on a literature research and interviews with experts, that group has decided 12 criteria. Six of them were technical conditive factors, namely solar radiation, estimated power, air temperature, cloudy days, wind speed and height. The other six were economic conditioning factors, namely slope, surfaceness, land coverage, proximity to roads, proximity to cities and proximity of schedules.
The research team obtained 612 PV sites from the Global Solar Power Tracker database, together with 612 non-PV sites. With the help of ArcGIS, each was taken into account for the 12 technical and economic criteria. The dataset was then divided into a training subset (70%), a validation sub -set (15%) and a test sub -set (15%). The model itself is built with Tabet, Optuna and Shap. The first is a deep -place model optimized for table data, the second automates the selection of the best model settings, and the latter was used to understand how the model does its predictions.
“Results demonstrated the high predictive performance of the proposed approach across both validation and testing datasets, achieving classification accuracy (ca) and area under the receiver operating characteristic curve (auc) values of 0.875 and 0.947, and 0.947, and 0.947, and 0.947, and 0.947, Respectively, thus outperforming tabpfn, ft-transformer, and seven ml Models, including RF, ”The academics stressed.
By applying their model to the entire middle -east, they discovered that about 5.8% of the region has a very high suitability and 11.5% is very suitable for the development of PV energy. The very high suitability areas are mainly concentrated in the coastal countries, Central Anatolia and parts of Saudi Aarabia and Iran, while the very suitable areas are found in particular in the center of and south of Turkey, the alluvial plain of Iraq and the Nile Basin in Egypt.
“We were particularly surprised to discover that the proximity of infrastructure such as roads, cities and electricity networks had a greater influence on the suitability of PV site than solar radiation itself,” Maulud said. “This is in contrast with conventional assumptions and is probably due to the uniform exposure to solar energy in the middle -east. Moreover, surface line turned out to be negative, an underexposed factor in earlier studies.”
Maulud added that the research team is planting various follow -up studies. “These include the inclusion of additional exclusion and environmental risk factors on the micro-scale, integrating time series of earth observation data to assess temporary changes in suitability and to expand the model for hybrid systems such as Wind-PV or Agro-PV.
The new methodology was presented in “Improving the suitability of solar PV in the middle -east using an optimized Deep Learning FrameworkPublished in the Alexandria Engineering Journal. Apart from the National University of Malaysia, scientists from the Iraq’s University of Thi-Qar and the University of Sharjah of the United Arab Emirates also participated in the study.
This content is protected by copyright and may not be reused. If you want to work with us and reuse part of our content, please contact: editors@pv-magazine.com.
