Search

Abstract Various dielectric response processes in the oil‐paper insulation are sensitively affected by the insulating states (ageing degrees and moisture contents). However, the existing research is still incomplete in revealing the microscopic mechanisms of various dielectric response processes in oil‐paper insulation with different insulating states. Given this issue, the genetic algorithm is first adopted to extract the Dissado–Hill (D–H) model parameters by simulating the frequency domain spectroscopy (FDS) of oil‐paper insulation. Then, the change laws of the extracted D–H model parameters are adopted to reveal the microscopic mechanisms of various dielectric response processes. Microscopic mechanisms of four dielectric response processes are studied, which are quasi‐dc relaxation, loss peak relaxation, optical frequency relaxation, and DC conductance. Meanwhile, due to the dielectric response processes being dominated by various polar particles (methanol, ethanol, furfural, and water molecule), the contents of various ageing by‐products dissolved in the insulating oil are measured to support the above analysis. In this respect, a dielectric theoretical reference for the FDS technique to research the insulating states of oil‐paper insulation is provided.

Abstract The ageing kinetic model is regarded as an effective tool to describe the degradation of transformer paper insulation. However, the temperature of the transformer winding is unevenly distributed, and the existing ageing kinetic models only depend on experimental data of accelerated thermal ageing at a constant temperature. Given this issue, this work proposes an improved second‐order kinetic model accounting for the non‐uniform thermal field. First, oil‐paper insulation samples with different ageing stages are prepared under the non‐uniform thermal field. Then, the sample's experimental data of moisture dispersion and distribution of degree of polymerisation are collected. On this basis, the improved second‐order kinetic model is derived to simulate insulating paper's ageing process subjected to an axial temperature gradient. Further, the relationship of the parameters in the improved second‐order kinetic model with the temperature and moisture content at different winding heights is obtained. Finally, the proposed model is validated, and the results show that the proposed model is more scientific than the traditional models to describe the degradation of paper insulation. In that respect, this work illustrates an improved second‐order kinetic model to provide a reference for the ageing assessment/life prediction of transformer paper insulation under the non‐uniform thermal field.

Abstract The polarization information contained in the frequency‐domain spectroscopy (FDS) could be utilised for characterising the ageing degree of transformer oil‐immersed insulation. However, the polarization information will be obscured by its conductance effect, it is thus hard to analyse the polarization behaviour by using the traditional FDS technique. Given this issue, an alternative approach is proposed to analyse the polarization information using FDS decoupled by logarithmic‐derivative spectroscopy (LDS). Findings proved that the polarization loss information generated by polarization behaviour can be separated from the total dispersion loss. The proposed feature parameters are later used to analyse the polarization loss information under ageing impact. Also, the polarization loss surface is constructed by using these feature parameters to characterise the ageing degrees of oil‐immersed samples. Finally, the potential possibility of using the proposed method to analyse the polarization loss and ageing degree are respectively discussed.

Abstract The charge transport and accumulation in oil–paper can cause the insulation degradation. So far, the most widely used model to simulate space charge transport and accumulation is the bipolar charge transport (BCT) model, which can well simulate the space charge dynamics. However, there are two shortcomings in the algorithms for solving the BCT model. One is that there is almost no use of vectorisation technology, which may increase the complexity of the algorithm, the other is the usage of fixed step size which might bring extra computation cost. In view of this, an adaptive time‐stepping transient upstream finite element method (FEM) is developed to solve the BCT model considering trapping/detrapping, as well as the recombination phenomenon under DC condition in this article. Then, a vectorisation technology is used in the method to optimise the algorithm. Moreover, the adaptive time‐stepping method is introduced in simulation to reduce computation time and calculation amount. Simulation results are obtained by programing and later presented, which are basically consistent with the corresponding experimental results. Therefore, the proposed method is expected to promote the optimization design of the oil–paper insulation system.

Abstract The low‐frequency polarisation information of oil‐paper insulation could be easily obscured by the conductance effect, and the contained polarisation information thus can not be readily extracted from the frequency domain spectroscopy (FDS). Given this issue, an alternative idea is reported to extract the low‐frequency polarisation information by using logarithmic‐derivative spectroscopy (LDS). The present findings proved that the parameters extracted by using the LDS can be applied for studying the low‐frequency polarisation behaviour under the moisture effect. In that respect, the novelty of this work is in the exploration of the LDS as a potential tool to extract feature parameters for analyzing the low‐frequency polarisation information of transformer oil‐paper insulation.