Aiming at digital twin applications

and focusing on the temperature rise characteristics and online simulation requirements of 110 kV three-phase common-box GIS double-fracture disconnect switch (DDS)

this paper proposes a hybrid modeling method based on electromagnetic-fluid-thermal multi-physics coupling and proper orthogonal decomposition (POD) reduced-order modeling. Firstly

a detailed physical model including dynamic contacts

static contacts

and the SF6 gas chamber was constructed using finite volume method. The steady-state temperature-rise distribution analysis was performed under various loads (0.3 to 1.1 times rated current) and contact pressures (30 to 150 N). The results indicate that the average temperature rise of phase B conductors is about 2 K higher than that of phase A and C due to the heat accumulation effect at the middle position. When the contact pressure decreases from 150 N to 30 N

the local temperature rise increases by 55.1%. Subsequently

123 sets of loss data for key heat-generating components were generated using Latin hypercube sampling. The main modes of the temperature field were extracted via the POD method

and an 8th-order reduced-order model was constructed. The root mean square error (RMSE) is below 0.01%

with a maximum error below 0.131 K. Compared with the full-order model

the solution time is reduced from 23 500 s to 0.8 s

improving calculation efficiency by approximately 2.9×104 times

which fully satisfies the real-time requirements of digital twins. Finally

the reduced-order model was integrated into an IoT-driven digital twin platform

achieving 3D dynamic visualization and real-time temperature contours display for gas insulated switchgear (GIS). The results provide efficient and accurate technical supports for fault early-warning and online monitoring in intelligent operation and maintenance. The method provides a theoretical and practical foundation for the design optimization and digital operation and maintenance of high-voltage switchgear.