A Data-Driven Sensitivity Calculation Method for Measurement Error Resistance under Highly Random Charging Load

As the strong randomness of loads such as electric vehicles are integrated into the distribution grid, it causes rapid and large power fluctuations. However, the measurement devices in the medium voltage distribution networks have limited accuracy and asynchronous measurements, resulting in large measurement errors. To address the issue, we proposed a datadriven sensitivity calculation method with anti-measurement error characteristics. The topology information of the power grid is generally obtainable in practice. Then the empirical statistics are used to extract analytical sensitivity matrix constraints, which are then involved in the linear sensitivity parameter regression process to reduce the impact of large measurement errors and improve accuracy. The proposed method is validated through IEEE cases and demonstrates significant accuracy advantages over unconstrained parameter regression methods.