Real-Time Estimation of Areal Photovoltaic Power Using Weather and Power Flow Data

To stably operate the power grid under a mass introduction of photovoltaic systems, a method to estimate the areal photovoltaic power in real time is necessary. The estimation method proposed in this paper achieves high accuracy by separately estimating the high- and low-frequency components of photovoltaic power. The high-frequency component is estimated using power flow at the distribution substation, whereas the low-frequency component is estimated using solar irradiance, ambient temperature, power flow, and if available 30-min integrated values of the electric load measured by smart meters. To evaluate the proposed method, simulations using measured data were performed. These were performed under four scenarios depending on the acceptance of a 15-min delay and the availability of smart meter data. In the best-case scenario, where 15-min delays to estimate the areal photovoltaic power is accepted and smart meter data are available, the estimation was performed with yearly root-mean-square error of approximately 2.2% of the total system capacity. Even in the worst-case scenario, where 15-min delays were unaccepted and smart meters not installed, the estimation was performed with yearly root-mean-square error of approximately 3.7% of the system capacity.