Research On Parameter Identification Of Vv Traction Transformer Based On Measured Data

Traction transformer,as the carrier of electric energy conversion,is one of the core devices connecting traction power supply system and power grid.Its accurate electrical parameters are the key to establish an accurate model of traction power supply system,which lays a foundation for power flow calculation and electrical transient characteristics.The traditional methods of obtaining transformer parameters include theoretical calculation and Maxwell simulation analysis,etc.However,the theoretical calculation is difficult to reflect the actual parameters of the running-online transformer,or Maxwell simulation analysis needs to build a detailed internal structure of the transformer.The above method has great limitations in practical engineering application.With the continuous development of monitoring technology,the acquisition of transformer electrical monitoring data becomes easier,which the research on parameter identification of transformer based on measured data has a good application prospect.However,the measured data generally have problems such as poor synchronization,amplitude,phase errors,which makes the traditional identification methods such as phasor measurement algorithm and ridge estimation algorithm have the problem of low accuracy or even unable to identify the parameters.Therefore,the research on identification methods with strong adaptability to amplitude error,phase error and asynchronous data has important engineering practical value for transformer parameter identification.In this paper,the equivalent circuit model of Vv traction transformer is established.Firstly,the parameter identification algorithm of traction transformer is deduced based on synchronous data.According to the actual situation,the applicability of parameter identification algorithm under different sampling data characteristics is analyzed.The factors affecting transformer parameter identification are explored,including voltage phase angle difference,reactance resistance ratio,data phase non-synchronous angle deviation,data amplitude deviation,etc.Further,the scheme of monitoring voltage phase angle difference is proposed as identifying resistance parameters.The influence mechanism of data measurement deviation on identification results is revealed,and the relationship between relative error of identification parameters and data deviation is quantified.Then,aiming at eliminating the influence of non-synchronous angle,a transformer parameter identification algorithm suitable for non-synchronous data is proposed.The algorithm uses the principle of least squares estimation and data to construct a set of overdetermined equations with high redundancy,which can effectively eliminate the influence of non-synchronous angle.Combined with the actual cases,the error of the identification results of the algorithm is analyzed.Finally,based on the measured data,the effectiveness of the algorithm proposed in this paper and the algorithm proposed in the references is tested,the outliers of the measured data are detected and eliminated by using the Grubbs-rule,and the sample determination coefficient is formulated as the criterion of data screening.Based on the data screening mechanism,the accuracy of this algorithm is further improved.The effectiveness and engineering applicability of the proposed algorithm are verified by simulation and measured data.