DOME will examine whether measuring on-line impedance spectra of a gird can give early warning of emerging oscillations, and beyond that, whether it is possible to identify which aspects of which equipment should be re-tuned to damp those oscillations. This is a data-driven method that will not require owners/vendors of wind farms to disclose their internal control models. The analytical methods for mode identification and participation assessment have been created in previous academic research, this project will assess whether these methods are capable of practical implementation. DOME is a desktop study that will use data gathered by Transmission Owners (TOs), example systems models and small-scale laboratory testing. The project will report on the viability of practical implementation and on how field trials could be conducted.
Benefits
The implementation of online monitoring could bring several benefits, including:
- Early Warning and Reduced Risk for Possible Oscillations: By continuously monitoring the power grid in real-time, online monitoring can detect the early signs of oscillations and alert grid operators before they escalate into a more significant issue. This can reduce the risk of system instability and potential blackouts.
- Reduced Renewable Curtailment: Online monitoring can provide grid operators with more accurate information about the behaviour of renewable resources such as wind farms and solar PV.
- Increased System Security: By providing grid operators with more accurate and timely information, online monitoring can increase the overall security of the power grid.
- Contributing Toward Net Zero Target: Online monitoring can help to integrate more renewable energy into the power grid, which is critical to achieving net-zero emissions targets.
Learnings
Outcomes
1 final report accompanied by 2 self-contained supplementary technical reports were submitted to NESO.
1 technical report for WP1, and 1 joint technical report for WP2&3 were submitted to NESO.
1 pre-designed Excel spreadsheet which calculates uDEF and generates plots from the PMU data was submitted to NESO.
1 journal paper introducing the hybrid data-model driven admittance identification method delivered from this project was submitted to IEEE Transactions on Industrial Electronics. The paper is now under review with preprint version available at: Hybrid Data/Model-Driven Whole-System Admittance Identification via Single-Point Injections.
1 technical document about uDEF was published on-line as a preprint: Unified Dissipation Energy Flow. TechRxiv. August 12, 2024.
Oscillation data from simulation case studies developed in this project was provided to NESO for further studies.
Dome final report is available on ENA Smarter network portal
Lessons Learnt
Free oscillations in modern power systems, especially those with high inverter-based resource (IBR) penetration, are increasingly complex, and available oscillation tracing methods such as DEF and its variants do not guarantee accurate identification in all scenarios.
Future Direction: Further development and benchmarking of oscillation source location methods, particularly under varying system conditions, should be prioritised, including validation with real-world events and models.
For projects that require data from industry partners, such as the PMU data required in DOME project, accessing the data may be constrained by confidentiality concerns. Non-disclosure agreement may be needed, which introduces additional reviewing and administrative steps and may delay project progress.
Future Direction: A thorough risk assessment regarding data accessibility should be conducted at the project planning stage. Contingency plans should be carefully developed to address scenarios where the required data cannot be obtained in a timely manner.
The lack of ground truth significantly hinders validation of oscillation-tracing methods.
Future Direction: The development of high-fidelity small-signal models replicating observed oscillations would be valuable for method validation and algorithm improvement.
The proposed unified Dissipative Energy Flow (uDEF) method successfully generalises existing DEF methods and retains energy conservation properties, allowing source/sink identification when properly configured.
Future Direction: Further work is required to systematically study the impact of different configurations of the weight matrix in uDEF method, especially under varied network and control conditions.
Accurate time-domain identification relies on comprehensive knowledge of the event (input), system topology, and load/generator parameters. Missing information (e.g., at PMU-unmonitored IBR buses) must be approximated, which can introduce error.
Future Direction: Advanced methods for dealing with missing data, such as machine learning-based estimation, can be explored and tested in practical applications.
Frequency-domain identification would benefit from dedicated signal injection equipment for long-term monitoring.
Future Direction: The design, testing, and deployment of such equipment should be considered a priority for infrastructure readiness.
Oscillation events are growing more frequent globally, and DEF’s limitations are acknowledged by multiple system operators.
Future Direction: Cross-system operator collaboration and data sharing (where possible) should be encouraged to improve benchmarking and robustness of new methods like uDEF.
The impedance-based approach is widely considered valuable for understanding oscillation mechanisms in IBR-dominated networks.
Future Direction: Future projects should prioritise making impedance-based tools more user-friendly and actionable for real-time grid operation and control room decision support.