Tertiary windings of super grid transformers in transmission networks provide a cost-effective and efficient connection solution for BESS (Battery Energy Storage Systems). Numerous connections to BESS and small-scale renewable generation projects are planned. This could result in flows of active and reactive power through the transformers that were not anticipated or planned for in the original system design. This project will develop appropriate modelling methodologies for assessing the dynamic thermal performance and voltage interactions of grid transformers with active usage of tertiary connections. A methodology for combined assessment of system and thermal impacts of grid transformer tertiary connections will be completed to understand the voltage control interactions as well as the dynamic thermal performance of the transformer and risks to asset health.
Benefits
The project aims to develop appropriate modelling and methodology for assessing the dynamic thermal performance and voltage interactions of grid transformers with active usage of tertiary connections, which will help to prevent an increased risk of Super grid transformers failure and reducing the cost of the resulting loss of tertiary connections. The resulting benefits is estimated to be around £4.69m by 2034/2035 in NPV (Net Present Value) if the project is successful. There could also be potential savings that are derived from the avoidance of wider network failures induced by possible voltage volatility in a weak grid, and potential savings that could be made by transformer lifetime extension.
Learnings
Outcomes
1. Appropriate dynamic models of voltage control systems for transformers with active tertiary connection.
The project has demonstrated that an RMS phasor domain model (implemented in DIgSILENT PowerFactory) is appropriate and necessary for studying OLTC interactions with tertiary connected equipment. RMS simulation captures tap changer delays, deadbands and circulating current dynamics to a level of detail that quasi dynamic (sequential load flow) methods cannot represent. The modular model architecture and the four typical substation configurations capturing common layouts for two parallel connected SGTs provide a robust platform for comparative studies and for extending to site specific analyses.
2. Appropriate dynamic thermal model with good representation of three windings, dual-temperature cooling modes and tap changers.
A dynamic transformer thermal model was developed based on the concept of IEC thermal diagram but with significant expansion to accommodate three-winding auto-transformer design, temperature dependant winding resistance calculation, tap-changer position effect and dual-temperature cooling modes. The model outputs not only top oil temperature and hotspot temperature but also cooling mode status and loss of lifetime. The model was implemented in MATLAB software. The methodology to derive the necessary thermal parameters from factory temperature-rise tests (also called heat-run tests) was illustrated. This facilitates the future use of the model for transformers with different thermal designs.
3. In-depth understanding of the impact of active tertiary connections on the voltage control interaction and thermal performance.
4. Comprehensive analyses on voltage and thermal performances of transformers with active tertiary connections under various loading scenarios.
5. Impacts of tertiary connection with BESSs:
Across all investigated scenarios, BESSs generally have shown a minimal effect on transmission system voltages. Charging during low load periods and discharging during high load periods tends to reduce the number of OLTC operations by smoothing the transformer loading profile and reducing voltage variations. However, the effect depends strongly on circulating current thresholds and whether tertiary asymmetries (e.g., reactors) are present. Solar PV reduces transformer loading and can slightly reduce tap operations, though the effect is small. At higher penetrations, PV can increase voltage and marginally increase tap activity, but the system remains well regulated. Reverse power flow does not create problematic OLTC behaviour in the studied configurations. Pulsating load profiles of the BESS operation caused corresponding fluctuations in hotspot temperature, but the overall effect on tertiary winding loss of life was negligible across all loading scenarios investigated.
6. Impacts of tertiary connection with data centres:
Data centre loading has limited voltage impact when a constant loading profile is assumed. The constant loading results in some reactive circulating current, though this is typically less than when a reactor or capacitor is connected as the demand is real power. The inclusion of a data centre load did not result in additional tap operations. In contrast, a tertiary‑connected data centre imposed continuous near‑rated loading, resulting in persistently high tertiary hotspot temperatures. In these cases, the tertiary winding may dominate the cooling control.
7. Recommendations for dynamic voltage control and thermal modelling of transformer with active tertiary connections.
The number of OLTC operations is highly sensitive to the choice of circulating current deadbands. Tight deadbands on the circulating current control markedly increase tap activity. A revised OLTC algorithm that reduced tap operations in simulation was proposed, which requires further investigations to be implemented and generalised.
The CDF thermal simulations showed that the main winding loading affects the thermal profile and key dynamic thermal model parameters of the tertiary winding. Such an interaction was not considered in conventional IEC thermal model based dynamic thermal modelling. Suggestions for model improvements were recommended, which requires further investigations to be implemented and generalised.
Total NIA Expenditure on project
Total Spend: £604,000
External Spend: £500,823
Internal/Indirect Spend: £103,000
£14,000 extra costs have been booked to this project due to an increase in the internal time booking.
Lessons Learnt
1. Voltage Assessments of Transformers with Active Tertiary Connections
RMS modelling is essential for capturing OLTC behaviour. The project demonstrates that RMS dynamic simulation is required to represent OLTC deadbands, time delays, and circulating current control. Quasi dynamic approaches systematically misrepresent tap counts and cannot capture circulating current interactions or tap stagger effects.
BESS generally has a modest and often beneficial impact on voltage. Across all configurations, BESS operation influences 132 kV voltage sufficiently to trigger tap changes, but the overall impact is modest. When BESS charge/discharge cycles align with system loading, they smooth transformer loading and can reduce tap operations.
Solar PV reduces transformer loading and can slightly reduce tap operations, though the effect is small. At higher penetrations, PV can increase voltage and marginally increase tap activity, but the system remains well regulated. Reverse power flow does not create problematic OLTC behaviour in the studied configurations. Data centre loading has limited voltage impact. Tertiary connected data centres introduce steady reactive loading that produces modest circulating current and a small tap stagger, but do not materially increase tap operations or degrade voltage performance. Their impact is smaller than that of BESS or PV variability.
Circulating current control is the most sensitive and operationally critical mechanism. Tertiary asymmetry (e.g., reactors, capacitors, uneven BESS operation) creates persistent reactive circulating current, often driving maximum tap stagger between parallel SGTs. Deadband width is a dominant driver of tap activity. Narrow deadbands (e.g., ±30 A) can increase tap operations by an order of magnitude. When tertiary asymmetry is removed, BESS variability becomes the dominant source of circulating current driven tap changes.
Substation topology strongly shapes OLTC interactions. Two transformer sites are more prone to tap stagger and circulating current sensitivity, while three transformer sites distribute circulating current more evenly and exhibit more stable OLTC behaviour.
2. Thermal Assessments of Transformers with Active Tertiary Connections
An enhanced dynamic thermal model was developed extending the IEC thermal diagram concept. The model accounts for three‑winding configurations, temperature‑dependent winding resistance, tap‑changer position, and dual cooling modes.
The effect of tap‑changer operation on transformer thermal modelling was assessed. While tap position influences winding resistance and losses, its impact on hotspot temperatures was minimal for the investigated cases, largely due to low loading levels and moderate tap variations.
For BESS operation, pulsating load profiles caused corresponding fluctuations in hotspot temperature, but the overall effect on tertiary winding loss of life was negligible across all loading scenarios investigated. In contrast, a tertiary‑connected data centre imposed continuous near‑rated loading, resulting in persistently high tertiary hotspot temperatures. In these cases, the tertiary winding may dominate cooling control, with the most severe thermal conditions occurring during single‑SGT operation combined with high solar injection on the secondary side.
To address unbalanced loading and mixed cooling modes, an innovative 2D–3D coupled CFD model was developed. This approach enabled detailed thermal analysis of all three windings while maintaining manageable computational effort. CFD results showed that variations in tertiary loading had limited impact on main winding temperatures, whereas changes in main winding loading significantly affected tertiary winding thermal behaviour.
CFD simulations also demonstrated that thermal interactions between windings influence key parameters in the dynamic thermal model. Higher main winding loading reduced the effective tertiary winding temperature gradient, leading conventional IEC‑based models to overestimate tertiary hotspot temperatures for the investigated transformer design. Recommendation to incorporate winding thermal interactions was therefore proposed, highlighting the need to extend existing thermal modelling method for transformers with actively loaded tertiary windings.
Dissemination
2024/2025
- Presentation at the CIGRE UK SC A2-D1 Annual Laision meeting on ”Development of Thermal Models for Grid Transformers with Tertiary Windings Connected to BESS”. The University of Manchester, 28th November 2024.
- Presentation at the CIGRE UK SC-A2 & NGN Workshop on “Optimal Control Strategies for Parallel Operation of Supergrid Transformers”. The University of Manchester, 7th March 2025.
2025/2026
- Poster Presentation at Energy Innovation Summit on “ Voltage Interaction and Thermal Dynamics of Tertiary Connection”, Scottish Event Campus, Glasgow, 5-6th November 2025.
- A dissemination workshop is held on 2nd June 2026 to summarise the project learnings.
