Correlation Analysis of Urban Rail Transit Electricity Consumption Based on Copula Function

Accurately quantifying the dynamic correlation between subway traction electricity consumption and its influencing factors is crucial for energy saving. Most existing studies use static or isolated analyses, failing to capture its time-varying complex characteristics. This study proposes a time-varying Copula-GARCH framework: first, the GARCH(3,3)-t model fits the edge distribution of passenger flow, traveling kilometers, maximum temperature, and traction consumption difference sequences to describe their thick-tail, peak, and wave-aggregation features; second, a time-varying Copula constructs a multivariate joint distribution for dynamic correlation analysis. Empirical research on 2024 Xiamen Metro data shows: (1)Significant positive time-varying correlation exists between traction consumption and these factors; (2) The time-varying Clayton Copula fits best, with upper-tail correlation (0.751) higher than lower-tail (0.686), meaning factor increases drive consumption more than decreases inhibit it; (3) Dynamic correlation rises sharply in summer, peaking at 0.99 (showing strong temperature impact). This model effectively captures nonlinear, asymmetric dynamic correlations, supporting refined energy management strategies.