Loading…

Current transformer saturation compensation based on a partial nonlinear model

► A partial nonlinear regression model is developed and used to estimate the phasor of fault current. ► It is more accurate and faster than the existing methods that only use unsaturated sections to compensate the saturation. ► A separable nonlinear least squares method dramatically reduces the comp...

Full description

Saved in:
Bibliographic Details
Published in:Electric power systems research 2013-04, Vol.97, p.34-40
Main Authors: Shi, D.Y., Buse, J., Wu, Q.H., Guo, C.X.
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:► A partial nonlinear regression model is developed and used to estimate the phasor of fault current. ► It is more accurate and faster than the existing methods that only use unsaturated sections to compensate the saturation. ► A separable nonlinear least squares method dramatically reduces the computational load of the regression calculation, for real-time implementation. ► The method can be used not only for protection relaying but also many other relevant applications. ► It has been implemented in an FPGA based embedded system as a function module. Using the module, real-time compensation can be achieved. This paper proposes a partial nonlinear model to accurately represent the nonlinear saturation characteristic of a current transformer (CT). Based on the model, the saturated section of the secondary current as well as the unsaturated section can be used in a regression process to estimate model parameters. The saturated section normally lies near the inception of a fault, therefore accurate parameters can be obtained faster compared with the methods using only unsaturated sections. The pre-fault remanent flux and DC-offset, which could significantly influence CT saturation, are both considered in the model, thus they do not affect the accuracy of the parameter estimation. The computational load of the regression calculation is significantly reduced by using separable nonlinear least squares (SNLLS) method. This provides the feasibility to implement the method for real-time protective relaying. The performance of the method has been evaluated on the data obtained from both PSCAD/EMTDC simulation and live recording with a test CT. The method has also been implemented in a Field Programmable Gate Array (FPGA).
ISSN:0378-7796
1873-2046
DOI:10.1016/j.epsr.2012.11.019