Factoring behind-the-meter solar into load forecasting: Case studies under extreme weather

Distributed energy resources (DERs), especially distributed photovoltaics (PV), have been rising dramatically over the past years. However, behind-the-meter (BTM) PV devices are not monitored, and thus are invisible to utilities and system operators. In addition, electricity demand is likely to increase as a result of extreme hot/cold weather conditions, which stretches the grid to its limits, and thus triggers high electricity price. High electricity prices and extreme temperatures also stimulate the adoption of solar panels, which in turn add difficulties to load forecasting. This paper proposes a data-driven feeder-level load forecasting method by taking account of BTM PV under extreme weather conditions. The BTM PV penetration is first estimated, and in this study the PV penetration is defined as the ratio of total BTM PV capacity to peak load of the feeder. A machine learning model is adopted to quantify the relationship between measured PV power generation and corresponding solar irradiance. The BTM PV generation within the entire feeder can be estimated through the PV penetration and forecasted PV irradiance, which is then integrated in load forecasting. Numerical results of case studies at three distribution feeders show that the performance of load forecasting under extreme weather conditions is significantly enhanced by considering the contribution of BTM PV.