A state estimation method for photovoltaic power generation using independent component analysis

A large number of Photovoltaic Generations (PVs) are expected to be introduced in distribution systems (D/Ss). Therefore the state estimation is important problem for stable and reliable system operation. However, it is difficult to estimate the total power of PVs connected to a D/S because metering spots and metering items are limited in existing D/S. In this paper, we propose an estimation method for sum of PV-outputs connected to a D/S. This method enables to estimate PV-outputs by analyzing a power flow data which include PV-outputs and various distribution loads using independent component analysis (ICA). The estimation by ICA needs the same number of observation signals as estimation signals. So we propose the method which is able to estimate PV-outputs and load-changes from active and reactive power flow those are recognized as two observation signals measured at one observation point on distribution lines. And the information of insolation as a priori knowledge is added to the method improving the estimation accuracy.