Hybrid photovoltaic-thermal solar collector modelling with parameter identification using operation data

Hybrid photovoltaic-thermal (PVT) solar collectors combine photovoltaic and thermal technologies to produce both electricity and thermal energy. To integrate efficiently PVT solar collectors, an optimal control approach is crucial. Accurate information about their thermal and electrical yields under real operating conditions is necessary for applying such an approach. Obtaining this information requires detailed testing and modelling of the PVT solar collectors. However, this is time consuming and does not take into account specific local characteristics. Consequently, a data-driven approach based on grey box modelling, automated parameter identification and numerical optimization becomes important. This approach can be particularly valuable in situations where traditional modelling methods do not fully capture the actual operating and installation conditions. Therefore, this paper presents several PVT grey-box model structures and analyses their accuracy and reliability using own experimental data. It also analyses the parameter identification performance and identifies the most appropriate model variants for optimal control application within a data-driven framework.