Performance assessment without pyranometers: Predicting energy output based on historical correlation

Estimating the energy that should be generated by a PV system under the prevailing conditions of irradiance and temperature is very important for system or fleet investors, energy customers, and operators. Traditionally this has been achieved by measuring irradiance and temperature in the proximity of the PV array in order to calculate the expected energy output either by using an appropriate model or by making assumptions about the system's Performance Ratio. However, this method requires the accurate installation, maintenance, and continuous monitoring of sensors thereby increasing the system's capital and maintenance costs. In this work we present an alternative methodology which can calculate the expected output of one or more systems in a regional fleet based on the measured power output from a subset of the total fleet. This can be achieved thanks to high accuracy energy measurements and the ability to correlate historical performance records.