An adaptive method for the identification of the main features of photovoltaic modules

The recognition of the main features shown by a photovoltaic module through its current vs. voltage curve is of main interest in many applications, e.g. diagnosis and dynamical reconfiguration. Indeed, the automatic detection of the fact that it works in mismatched conditions, or some data directly related to its health conditions would be very useful. The curve is usually acquired by a switching converter, performing a voltage/current scanning, and it consists of a number of samples. The processing of a high number of samples requires computation resources that are than those ones available through actual embedded systems and also long data transfer sessions towards a centralized logging system. In this paper an algorithm which does not need any prior knowledge about the module, but which is able to identify the main features from the current vs. voltage curve of the module thereof is presented. The solutions presented in this paper overcome some limitations shown by an algorithm recently presented in literature. The results are validated by using simulated as well as experimental curves.