基于输入阻抗低频模型的混合型MMC-HVDC线路故障性质辨识判据OA

With the development of large-scale wind and solar power bases in deserts and Gobi regions,the demand for centralized renewable energy integration and long-distance transmission continues to grow.The full-half bridge hybrid modular multilevel converter-based high-voltage direct current(MMC-HVDC)system,which eliminates commutation failure while offering DC fault ride-through capability and cost-effectiveness,presents an effective solution for large-scale renewable energy transmission.For long-distance and large-capacity power transmission,overhead lines are commonly adopted as the transmission medium.Overhead lines are long and the working environment is complex,resulting in high fault probability.An effective fault restart method is essential to determine whether the system power can be restored.Traditional auto-restart strategy struggles to distinguish fault properties under high-impedance conditions.Therefore,this paper proposes a highly sensitive and reliable fault identification criterion to improve the system restart success rate and prevent unnecessary system shutdowns. First,by analyzing the control response of the converter after line protection activation,the mathematical models of the converter station and DC line are derived,yielding the equivalent circuit of the system.Combining the definition of input impedance,frequency-domain characteristics of input impedance under different fault types and fault properties are investigated.The analysis reveals that,regardless of the fault type,the low-frequency band characteristics of the input impedance consistently contains information that effectively reflects whether the fault has disappeared.Second,by appropriately designing a low-pass filter and selecting the transfer function order,the low-frequency input impedance model is constructed offline using complex curve fitting method based on the frequency response data of the input impedance under transient fault.Subsequently,this model is discretized via the bilinear transform to enable computation with measured sampling data.Specifically,the measured current is processed through the low-pass filter and fed into the discretized input impedance model,generating the predicted low-frequency voltage,which is compared with the filtered measured voltage to establish the fault identification criterion.Additionally,the threshold setting principle and identification procedure are also provided.Finally,simulation results verify the correctness of the input impedance frequency-domain analysis and demonstrate the effectiveness of the proposed criterion.Furthermore,comparative analysis shows that the proposed criterion,utilizing only single-end measurements,significantly improves the sensitivity,reliability,and adaptability to different fault types in fault property identification. The conclusions can be drawn as follows:(1)During the restart stage,the faulty-pole hybrid MMC station operating in fault ride-through mode can be equivalently represented as an RLC series branch,while in signal injection mode it functions as a controllable voltage source with a series RL branch.The non-faulty-pole hybrid MMC station can be modeled as a constant 1p.u.voltage source with a series RL branch.(2)Frequency-domain analysis of input impedance demonstrates distinct differences in low-frequency characteristics between temporary and permanent faults.A transfer function representing these low-frequency features is fitted using complex curve fitting method and discretized via the bilinear transform,enabling computation with sampled current data to support the fault identification criterion implementation.(3)The proposed criterion is theoretically sound,demonstrates strong noise immunity,and remains effective for fault resistance up to 1000 Ω.Requiring only single-ended data,it outperforms existing methods in sensitivity and reliability while being simpler to implement.