基于时域回归模型的CVT传递特性获取方法

Method for Acquiring CVT Transfer Characteristics Based on Time Domain Regression Model

扫频法获取电容式电压互感器(capacitor voltage transformer, CVT)的传递特性时,存在电源功率大、频带受限的难题,因此提出了一种基于时域回归模型的CVT传递特性获取方法,并进行了传递特性曲线的可信度评估。当CVT工作在线性状态时,其传递函数可表示为s域多项式构成的分式形式,进而可建立其时域回归模型。基于单次冲击试验获取CVT一次、二次侧数据,考虑了时域回归模型在实际建模中的多波峰以及鲁棒性问题,提出了基于调幅矩阵和鲁棒矩阵的参数辨识方法。对比扫频法测量所得的传递特性曲线和时域回归模型的传递特性曲线并进行可靠性评估,结果显示误差曲线的均值处于零假设的95%置信区间以内,所提出的传递特性获取方法可为现场快速获取CVT的传递特性提供理论指导。

Aiming at the problem of large power supply and limited frequency band when acquiring the Capacitor Voltage Transformer(CVT) transfer characteristics by the frequency sweep method, a method for obtaining the CVT transfer characteristics based on the time domain regression model is introduced. When the CVT works in a linear state, its transfer function can be expressed as a fractional form composed of polynomials in the s-domain, and its time domain regression model can be established. Based on the obtained CVT primary and secondary data by the single-shot tests, a parameter identification method based on the amplitude modulation matrix and robust matrix is proposed after taking into account the multi-peaks and robustness of the time domain regression model in actual modeling. The credibility evaluation is performed after comparing the transmission characteristics curve obtained by the established time domain regression mode with that measured by the frequency sweep method. The results show that the mean value of the error curve is within the 95% confidence interval of the null hypothesis. Therefore, the proposed method for obtaining transmission characteristics can provide a reference for quickly obtaining CVT transmission characteristics in the field.