A Machine-Learning-based Prediction of Solar Power Generation for a Novel Net Zero Energy Building

Solar Power (SP) forecasting is essential to enhance the utilization of green energy on a large scale. Nevertheless, the irradiation time series exhibits non-linearity, fluctuation, and intermittence, hence posing challenges to accurate prediction. The precise estimation of SP generation is essential for the utilities to adjust their generation scheduling, optimize the use of other power sources, and prevent grid overloading and under-utilization. Here, the prediction model uses a historical dataset and thereafter, forecasts the yearly solar irradiance for the Kolkata region in India. Subsequently, the load profile is also predicted for finding the expected energy yield of a novel Net Zero Energy Building (NZEB). Moreover, the SP generation is assessed by employing a hybrid SP array configuration scheme. Here, the NZEB infrastructure is strategically planned, with consideration given to the electrical demands of the facility’s load profile.