Abstract：Aiming at the problems of significant nonlinear metering error and low overshoot response rate

faced by electric vehicle charging equipment in complex environments，an out-of-tolerance metering dynamic

modeling method integrating the neural ordinary differential equations（ODE）and vehicle-pile-cloud synergy data

was proposed. Based on the vehicle-end，pile-end and cloud synergistic data，the charging environment topology

was constructed，the energy relationship between stations was analyzed，the charging equipment metering error was predicted，the voltage error matrix and metering uncertainty were established，which were inputted into the neural ODE model，and the loss function was designed for dynamic error correction. The experimental results show that the proposed model for charging equipment metering error prediction relative error fluctuation is controlled within ±1%，the out-of-tolerance metering error response rate are more than 95%，and the average absolute error maximum value is only 0.001 4 in low，normal and high temperature environments，which is significantly better than the comparison method. This study effectively improves the metering accuracy and reliability of charging equipment，and provides new ideas for high-precision metering and intelligent operation and maintenance.

        Key words：out-of-tolerance metering；vehicle-pile-cloud synergy data；dynamic correction；nonlinear dynamic properties

     Format Citation:陈稳，李鹏程，张华,等. 车-桩-云协同下的充电设备超差计量方法[J].电气传动,2026,56(07):91-96. CHEN Wen，LI Pengcheng，ZHANG Hua,et al. Out-of-tolerance metering method for charging equipment under vehicle-pile-cloud synergy [J].Electric Drive, 2026,56(07):91-96