Structural optimization and design of 110 kV longitudinal Optical Voltage Transducer

Objective With the continuous improvement of power system transmission capacity and voltage level, previous voltage transformers have exposed some fundamental shortcomings, such as poor insulation performance, large volume, electromagnetic resonance, easy magnetic saturation, etc., which are difficult to meet the needs of smart grid development. Optical Voltage Transformer (OVT) adopts optical sensing technology, which can overcome the above defects and better meet the development needs of China's smart grid, with good development prospects. However, there are still many unresolved issues with OVT. In practical applications, small positional deviations between optical devices may occur due to vibration, unstable connections, and thermal expansion and contraction, resulting in uneven distribution of internal electric fields and affecting measurement results. However, current research methods have only reduced the unevenness and coefficient of the internal electric field, and have not improved the accuracy of measurement results when small crystal and optical path offsets occur in OVT crystals. Therefore, this article conducts electric field simulation on the longitudinal modulation OVT structure of 110 kV, analyzes the non-uniformity and coefficient inside the electric field, and proposes a new dielectric layering and wrapping method when offset occurs.

Methods The sensing unit of OVT is simulated and analyzed using ANSYS Maxwell. Firstly, the distance between electrodes is determined (Fig.2), and then a layered structure is selected (Fig.3). Next, the packaging material and structure are selected (Tab.5), and finally the thickness of the packaging is determined (Fig.6). By using calcite with relatively high dielectric constant to divide the voltage of BGO and wrapping it with Al_N, the adverse effect of electric field non-uniformity after contact between BGO with higher dielectric constant and SF₆ gas with lower dielectric constant is greatly eliminated. And an experimental setup is built to verify its simulation results (Fig.9).

Results and Discussions The non-uniformity of the electric field inside the OVT sensing unit is mainly caused by direct contact between SF₆ (1.002) with a relatively small dielectric constant and BGO (16) with a relatively large dielectric constant. Direct contact between the two is avoided by using calcite layering with a relative dielectric constant of 8.3 and Al_N wrapping with 8.8. And a set of two ends Φ10 mm×75 mm calcite layers, with the middle being Φ10 mm×10 mm BGO and an OVT sensing unit with a thickness of 0.5 mm wrapped around it. Through system simulation, it can be found that the unevenness of the field integration voltage and the coefficient are significantly reduced (Fig.6), and the errors caused by crystal shift or optical path shift due to external factors are also significantly reduced. Finally, an experimental setup was constructed and signals were collected through CMOS. The results are shown (Fig.10), which presents the standard deviation calculated before and after the improvement, as well as under different voltages. The optimized effect can be clearly seen.

Conclusions This article conducts simulation analysis on longitudinal modulation OVT with a voltage level of 110 kV. Detailed analysis is conducted on electrode spacing, additional medium, and wrapped medium, and the most suitable structure is found to effectively reduce the unevenness of the internal electric field of the crystal. In addition, compared to other studies, this article greatly reduces the errors generated in the case of crystal or optical path offset. By improving the internal structure of the OVT system, the uneven electric field along the crystal axis has been reduced from 1.157 3 to 1.008 8 (Fig.6), and the maximum electric field integration error caused by optical path offset has been reduced from 0.090% to below 0.008%; The maximum electric field integration error caused by crystal displacement has been reduced from 0.075% to about 0.006% (Fig.8), and the improvement effect is significant. And through experiments, it can also be demonstrated that the model can effectively improve the distribution of electric field inside the crystal and reduce the integration voltage error (Fig.10), which further confirms the effectiveness and reliability of our proposed method.