基于数字孪生模型电流动态时间规整差异度的变压器早期故障辨识

Early fault identification for a transformer based on current dynamic time warping difference of a digital twin model

现有的早期故障辨识方法对于程度较为轻微的故障灵敏度低,且辨识方法多为离线检测。结合数字孪生技术建立了三相三柱式变压器内部无故障的孪生模型,基于动态时间规整算法提出新的故障特征量,提出早期故障保护方案。首先,利用实体设备的结构参数与实时的电气量,建立不同运行工况下的状态空间模型。然后,通过实际数据对孪生模型进行修正,使虚拟世界与物理世界实现同步,综合使用离散迭代法和Z变换法求解模型。其次,利用动态时间规整算法量化预测量与实测量的差异,以此作为故障特征量,并设计新型的故障保护方案,提高故障辨识的灵敏度。最后以三相三柱式变压器为例,通过ANSYS仿真和动模实验验证了所建模型的正确性,通过多次实验结果证实所提故障判据的有效性。

Existing early fault identification methods have low sensitivity for minor faults,and most identification methods are for offline detection.In this paper,a twin model of three-phase three-column transformer without internal fault is established by using digital twin(DT)technology.Based on a dynamic time warping(DTW)algorithm,a new fault characteristic quantity is proposed,as is an early fault protection scheme.First,a state space model in different operating conditions is established using the structural parameters of physical equipment and actual time electrical quantities,and then the twin model is modified through the actual data to synchronize the virtual with the physical world.A combination of a discrete iteration and Z-transform methods is used to analyze the model.Second,the dynamic time warping algorithm is used to quantify the difference between the predicted quantity and the actual measurement.This is used as the fault characteristic quantity,and a new fault protection scheme is designed to improve the sensitivity of fault identification.Finally,taking a three-phase three-column transformer as an example,the correctness of the model is verified by experiments based on an ANSYS simulation and dynamic model,and the effectiveness of the fault criterion proposed in this paper is verified by several test results.