Physics-_based digital twin modeling of PV modules without historical data

The instability and unpredictability of photovoltaic (PV) power generation pose significant challenges for power producers and grid operators. An accurate and efficient digital twin model could offer a viable solution to this issue. This study presents a generalizable method for constructing digital twin models of PV modules that operates independently of historical data. The proposed model systematically incorporates influencing factors under real operating conditions and integrates a triple-submodel structure—comprising wind speed distribution, thermal calculation, and power generation. By applying principles from fluid mechanics, thermodynamics, and solar cell theory, it performs multi-physics simulations to accurately replicate the operational behavior of PV modules in a virtual space. Experimental results indicate that the model achieves an error margin within 1 % at the module level and a mean accuracy of 96.35 % at the station level. Since the method does not rely on historical data, it is particularly suitable for power prediction in scenarios where data are scarce—such as under rare extreme weather conditions or in remote underserved areas—thereby facilitating the adoption of PV systems in such challenging environments. • A generic digital twin method for PV modules, independent of historical operational data for construction. • Integrates a triple-submodel structure for multi-physics simulation using fluid and thermodynamics principles. • Achieves high accuracy with sub-1 % error at module level and 96.35 % mean accuracy at station level. • Enables reliable power prediction in data-scarce scenarios like extreme weather or remote areas.