Solar power simulations for a renewable integration study in New Mexico using sparse input data

Since solar PV power generation is growing rapidly, it is important to accurately model solar power production in renewable generation integration studies which look at the impact of variable renewable generation on electric grid operations. However, solar irradiance or power measurements are often sparse both spatially and temporally, making it difficult to simulate PV power output. Here, we describe the technique used to simulate generation from up to 40 utility-scale PV plants and 9 areas of distributed PV in the state of New Mexico given only five hourly irradiance measurements plus a sixth irradiance measurement at 1-minute resolution that was used as a lookup library. Spatial smoothing based on the plant size was applied, then this area-average irradiance was converted to PV power output using irradiance to power models. In this way, PV power output profiles for each location, and for the aggregate of all locations, were produced and supplied to the integration study. Also for use in the study, day-ahead solar power output forecasts were created by adding errors representative of the current state of solar forecasting to the actual power output values.