Modelling and Optimization of Four-Segment Shielding Coils of Current Transformers

Yucheng Gao, Wei Zhao, Qing Wang, Kaifeng Qu, He Li, Haiming Shao, Songling Huang2017 · DOI

TL;DR

This paper introduces a novel analytical model for predicting the magnetic interference in current transformers (CTs) caused by adjacent bus-bars. By developing accurate models for both stray flux and coil leakage flux, the authors enable efficient optimization of four-segment shielding coils, leading to improved CT performance and reliability.

What Problem Does This Paper Address?

Heavy current transformers (CTs) used in large-capacity generators are susceptible to magnetic interference from nearby high-current bus-bars. This interference can cause partial magnetic saturation in the CT's core, leading to measurement errors and even permanent damage. While shielding coils are a common solution, their design and optimization have historically relied heavily on time-consuming physical experiments or complex numerical simulations. Existing analytical models are either outdated, lack accuracy for modern CTs, or are incomplete, particularly regarding the leakage flux generated by the shielding coils themselves. This paper aims to bridge this gap by providing a practical and efficient analytical framework for modeling and optimizing these shielding coils.

Methods

The core of this paper lies in the development of new analytical models and the refinement of existing ones.

  • Stray Flux Model Improvement: The authors start with an existing analytical model for stray flux [1] but recognize its limitations for modern, large-capacity CTs. They address this by using Finite Element Method (FEM) simulations to derive a new empirical equation for the "core-pickup factor" ($k_s$), making the stray flux model applicable to current mainstream CTs.
  • Coil Leakage Flux Model: A novel analytical model for the leakage flux produced by the shielding coils themselves is introduced. This model considers the coil as a combination of conductors and simplifies it into an electrical circuit model.
  • Four-Segment Shielding Coil Model: The developed flux models are integrated to create an electrical circuit model for four-segment shielding coils. This model allows for the calculation of coil currents and in-core remnant flux.
  • Optimization Framework: The analytical models are then used to perform a parameter sweep to optimize the shielding coil configuration. Performance metrics like maximum remnant flux ($|\Phi_r|_{\text{max}}$) and coil heating ($I_b^2\text{-max}$) are analyzed to determine optimal coil placement ($\beta$) and connection topology ("all parallel" vs. "separated loop").
  • FEM and Experimental Verification: The proposed analytical models are rigorously validated through FEM simulations and physical experiments on a 30 kA and a 5 kA CT sample.

Datasets

  • FEM Simulation Data: The paper utilizes data generated from FEM-based simulations (using ANSYS Workbench 15.0) to derive empirical equations for the core-pickup factors ($k_s$ and $k_{cs}$). These simulations were performed for various CT and bus-bar configurations as described in Tables 2 and 4.
  • Experimental Data: Physical experiments were conducted on a 30 kA CT sample and a 5 kA CT sample. The parameters for these samples are detailed in Tables 5 and 7, respectively. The experimental results were used to verify the accuracy of the developed analytical models.

Key Insights and Ideas

  • Bridging the Analytical-Numerical Gap: The paper successfully bridges the gap between purely analytical and purely numerical methods by using FEM simulations to derive accurate empirical correction factors for analytical models. This significantly reduces the computational burden for optimization.
  • Trade-off in Optimization: A crucial insight is the inherent trade-off between minimizing remnant flux (to prevent saturation) and minimizing coil heating. The "all parallel" topology with a 45° starting position is optimal for reducing remnant flux, while the "separated loop" topology with a 0° starting position is best for minimizing heat. This highlights the need for careful design decisions based on specific application requirements.
  • Generalizability of the Approach: The authors emphasize that the modeling and optimization procedure presented for four-segment coils can be generally applied to other shielding coil configurations, suggesting a broad applicability of their framework.
  • Importance of Coil Leakage Flux: The paper highlights that accurately modeling the leakage flux generated by the shielding coils themselves is critical for precise analysis, a factor often overlooked or simplified in previous work.

Key Results

  • Improved Stray Flux Model: The new empirical equation for the core-pickup factor ($k_s$) achieved a mean squared error of 0.091, demonstrating a significant improvement in accuracy for mainstream CTs.
  • Validated Coil Leakage Flux Model: The analytical model for coil leakage flux, along with its correction factor ($k_{cs}$), was shown to closely match experimental and FEM simulation results.
  • Optimal Shielding Coil Configurations:
    • For minimizing remnant flux ($|\Phi_r|_{\text{max}}$), the "all parallel" topology with a 45° starting position was found to be optimal, reducing the maximum remnant flux density to approximately 60% of the maximum possible value.
  • For minimizing coil heating ($I_b^2\text{-max}$), the "separated loop" topology with a 0° starting position was identified as superior, resulting in approximately 50% of the maximum heating value.
  • Experimental Verification: The analytical models for coil leakage flux, stray flux, and the overall shielding coil model showed good agreement with experimental data, validating their practical utility. For instance, the average coil currents from the analytical model were very close to experimental results (e.g., 2.554 A analytical vs. 2.458 A experimental for the 30 kA CT).

Limitations and Future Work

  • Dimensional Range: The empirical equations proposed in this paper are specifically applicable to mainstream CTs within the dimensional range defined in Table 1 and for coils with rectangular cross-sections.
  • Coil Shape: The models are based on rectangular cross-section coils. For coils with rounded corners, the core-pickup factor might differ slightly.
  • Future Work: The authors suggest future work could focus on the physical essence of the core-pickup factor to broaden the application scope of the analytical models. Investigating coils with rounded corners would also be a valuable extension.

Summary updated

BibTeX
@misc{gao2017modelling,
  author = {Yucheng Gao and Wei Zhao and Qing Wang and Kaifeng Qu and He Li and Haiming Shao and Songling Huang},
  title = {{Modelling and Optimization of Four-Segment Shielding Coils of Current Transformers}},
  year = {2017},
  doi = {10.3390/s17061218},
}

Loading related content...

Figure