Xiang Zhang

The breathing system of an oil‐immersed transformer is of vital importance for maintaining equilibrium between internal and external pressures, with any malfunction potentially disrupting the transformer's stable operation. This study develops a thermal–fluid–structure interaction (TFSI) model to simulate the breathing phenomenon in a 110 kV oil‐immersed transformer, addressing the limitations of current manual inspection methods. This model analyses key breathing parameters such as breathing volume and maximum breathing rate, along with their dynamic characteristics under various load factors and ambient temperatures, revealing nonlinear and linear relationships, respectively. To validate the model, the breathing parameters of a 110 kV oil‐immersed transformer were monitored over a 7‐day period using a breather equipped with a breathing monitoring functionality. Comparison between simulation results and monitoring data for the period from 7 p.m. March 7 to 7 p.m. March 8 revealed discrepancies of 6.93% for the maximum daily breathing rate and 6.73% for maximum daily breathing volume. This study contributes to the field by providing a more accurate and efficient method for analysing transformer breathing dynamics, potentially enhancing early detection of breathing system malfunctions, optimising breather maintenance schedules and improving diagnosis of pressure‐related faults in oil‐immersed transformers.

The liquid temperatures and velocities are of great importance for estimating the hot-spot temperature (HST) within the transformer windings. For liquid natural cooled power transformers (ON/KN), the liquid temperatures and velocities can only be obtained by modelling the complete-cooling-loop (CCL), which refers to the insulating liquid circulation between the windings and the radiators. In this paper, a CCL based computational fluid dynamics (CFD) model was developed for determining the liquid temperatures and velocities in the natural cooling mode. The validities of the CCL CFD simulations were verified by conducting experiments under different loading conditions, at different thermal heads and of different insulating liquids. The experimentally verified simulation results showed that the top liquid temperature increases exponentially against the power loss, whereas the bottom liquid temperature increases linearly against the power loss. The liquid velocity is in an approximately linear relationship against the square root of the product of the power loss and the thermal head. Moreover, the thermal performances of different insulating liquids were investigated. The dominating material property for the liquid thermal performance of an ON/KN transformer is the dynamic viscosity. For different liquids, the closer the dynamic viscosity, the better matching of the liquid temperatures and velocities.

Real-time thermal capacity estimation is not only part of prognostics and health management (PHM) but also determines the dynamic loading capability of the traction transformer. The hot-spot temperature is the vital parameter for real-time thermal capacity estimation. In high-speed railway, the unbalanced loads of the two phases of the traction transformer cause unbalanced winding losses and therefore affect the top-oil temperature and further influence the hot-spot temperature in each phase. In order to explore the dynamic thermal performance of the transformer under unbalanced load (current), a thermal model for calculating the hot-spot temperature was proposed incorporating thermal interaction between two power supply phases. This model treats the winding loss of each phase independently, considering the temperature dependence of winding ohmic losses and eddy current losses. A thermal circuit topology was presented with independent phase windings sharing the same top-oil temperature, where an improved heat transfer correlation was used to account for the nonlinear thermal convection. The proposed model was verified by the factory temperature rise tests and field operation data for the same batch of transformers.

Due to the influence of weather and various human factors, the transmission tower will be deformed, which will affect the operation of the transmission tower. In order to monitor the deformation of transmission tower, a deformation monitoring system is designed in this paper. The system consists of 15 wireless acceleration sensors, a wind speed and direction sensor, a monitoring terminal and a monitoring center. Acceleration sensor and monitoring terminal transmit data through wireless transmission; The output analog signal of wind speed and direction sensor is directly transmitted to the monitoring terminal; The monitoring terminal realizes data transmission through 4G network and monitoring center, which can monitor the deformation of transmission tower in real time.

More-electric-aircrafts (MEAs) are heading towards larger power, higher voltage, and more optimum architectures, it is extremely important to understand partial discharge (PD) behaviors under the challenging aeronautical power systems. PD under the pressure of 1–101 kPa and frequency of 50–1000 Hz is studied to emulate the practical situation with the consideration of insulation defects from electric machine windings. The high frequency current transformer (HFCT) is specifically employed to pick up partial discharge signals. Experimental results indicate that PD inception voltage (PDIV) decreases almost linearly with air pressure. As to the discharge amplitude, it increases first and then decreases as air pressure dropping. The highest discharge repetition rate and the number of discharges appear at 30 kPa. High frequency doubles the number of discharges without affecting PD amplitude. Low air pressure causes the phase shift and phase width of the phase-resolved partial discharge (PRPD) to increase, and the trend also is accelerated by high frequency. This experimental approach is beneficial to provide prerequisite knowledge in characterizing PD and assessing its potential adverse impact on the aeronautical power system and consequently contribute to optimizing the insulation design for MEA.

Predicting liquid flow distribution in the winding is crucial for transformer thermal design and hydraulic network models are usually used to guarantee reasonable liquid flow distribution. This paper is a continuation of the research effort in using computational fluid dynamics (CFD) to calibrate and improve the network modelling approach. The possibility is assessed of hydraulic network models in predicting the occurrence of reverse flows in pump-driven oil directed (OD) cooling modes. CFD modelling shows that the occurrence of reverse flows requires both high Reynolds number at the winding pass inlet and the “real” pass inlet velocity profile. To reflect the effect of pass inlet velocity profile on flow distribution, a winding pass is added beneath dividing and merging junctions in deriving minor pressure loss correlations. In addition, hybrid minor loss correlations are used to incorporate the effect of pass inlet velocity profile and to mitigate the inaccuracy of minor pressure loss segregation introduced by the inclusion of a winding pass beneath the junctions. Even with the above mentioned improvements, a conventional network model with one node for each junction still cannot predict reverse flows because of its architecture limit, hence a 3-node junction network model is then proposed. It is demonstrated that the newly developed 3-node junction hydraulic network model can predict the occurrence of reverse flows in OD cooled disc-type transformer windings.

Retrofilling transformers with a new liquid has been considered as an option for life extension and/or uprating purposes. The determination of the new total liquid flow rate is of vital importance in assessing the effect of retrofilling on flow and temperature distributions in the winding, especially for ON transformers. This paper presents the determination of flow rate ratios for four liquids—a mineral oil, a GTL oil, a synthetic ester and a natural ester—in ON transformer retrofilling scenarios. Flow rate ratios among different liquids are provided for a simplified condition of linear radiator oil temperature variation. For a more realistic exponential radiator oil temperature variation, the detailed transformer geometric information is needed to estimate the total flow rate. In addition, a hydraulic winding network model is established to prove that “minor pressure losses” due to change of flow directions is indeed minor for ON transformer liquid flow conditions, justifying neglecting the minor losses in the deduction of the total liquid flow rate.

Transformer loss of life is an important factor for the asset owner / operator to consider as it can be used in asset replacement planning and can help reduce the likelihood of unanticipated failures. Loss of life for a transformer is correlated strongly to the hottest spot temperature, which is a function of the loading on the transformer and ambient temperature.Accurate measurement of transformer load and the ambient temperature is not usually available to owners / operators of distribution transformers (due to cost). Instead, assumed profiles are utilised, and load profile simplifications such as the two-step calculation provided in IEEE Standard C57.91 are employed. This work looks to establish the applicability of such load profile simplification methods (`root-mean-squared' and `two-step' methods are considered), as well as the viability of averaging load and ambient temperatures into thirty-minute profiles from original minute-by-minute profiles. The analysis presented is conducted on load profile data of 81 installed distribution transformers in the United Kingdom for a period of 137 days.It is shown that, as a general rule, the simplification of the load profile to a `root-mean-squared' or a `two-step' profile does not appropriately match the calculation on a minute-by-minute basis. It is noted that only 11.9 % of the root-mean-squared and 12.2 % of the two-step model results fall within ±10 % of the original calculated value. Averaging of loads and ambient temperatures to thirty-minute profiles saw a larger portion of results matching the original, with 30.2 % within ±10 %.