Practical calculation of lost energy for large PV power plants

As PV power plants increase in size and number, there is a need to simply and accurately assess the performance of these systems for contractual purposes. For example, during grid instabilities, the power plant may be able to supply power to the grid but is unable to do so, and in many cases is still owed revenue for the undelivered energy. To accurately determine the energy a power plant delivers to the grid, a monthly multi-linear fit is proposed. This fit uses meteorological inputs of irradiance, ambient temperature and wind speed to determine the coefficients of energy generation. This technique is applied to the 21 MWac Blythe PV power plant. By calculating the multilinear fit on a monthly basis, errors due to irradiance seasonality, soiling, maintenance and known outages are minimized. The resultant modeled energy curve is clipped at both zero irradiance and at the plant capacity to accurately reflect the performance of the installation. Using this technique the mean absolute error is found to be <;2.5% of the plant capacity, ensuring an accurate estimation of energy and revenue generation.