Optimized Wind and Solar Power Forecasting for EV Infrastructure Along Asian Highway 47

Amid the escalating energy crisis, rising environmental concerns, and the severe impacts of climate change, governments are prioritizing carbon emission reduction. A key approach involves utilizing green energy technologies for charging electric vehicles (EVs). While EVs are highly efficient, their ability to lower greenhouse gas emissions depends on the energy sources used for charging. These sources include Renewable Energy Sources (RES) such as solar, wind, hydropower, geothermal, biomass, and ocean energy, which are classified as low-carbon. In contrast, non-renewable energy sources, including coal, natural gas, and oil, are high-carbon, whereas nuclear power is low-carbon but remains nonrenewable. This study emphasizes RES as a sustainable alternative, focusing on renewable energy forecasting to optimize EV charging infrastructure along the Karnataka segment of Asian Highway 47 (AH-47). The research employs Long Short-Term Memory (LSTM) and a hybrid LSTM-CNN (Convolutional Neural Network) model to enhance wind and solar power forecasting at six strategically selected locations along the route. Comparative analysis using error matrices reveals that LSTM excels in wind power forecasting (MAE = 2688.11 W, RMSE = 4512.48 W, R² = 0.75), whereas the LSTM-CNN hybrid achieves higher accuracy for solar power (MAE = 74.88 W, RMSE = 122.57 W, R² = 0.88). For solar, LSTM yielded MAE = 84.34 W, RMSE = 134.52 W, and R² = 0.86, while for wind, LSTM-CNN recorded MAE = 3422.01 W, RMSE = 6149.20 W, and R² = 0.55. Integrating these models ensures reliable renewable energy utilization, facilitating optimal EV charging station placement, reducing fossil fuel dependency, enhancing grid stability, and lowering charging costs.