This project develops a physics‑informed deep learning framework (PINNs) to estimate harmonic voltages and currents across transmission networks using limited power quality monitor data. By combining sparse measurements with network models, it enables real‑time harmonic “heat maps” and visibility at unmonitored nodes. The scalable software solution improves accuracy, supports optimal monitor placement, and enhances compliance and planning. It reduces reliance on costly monitoring infrastructure while improving power quality management and facilitating integration of inverter‑based and low‑carbon technologies.

Benefits

This project delivers enhanced visibility of harmonic conditions across the transmission network by extending effective monitoring coverage from ~25 substations to the full network (~350 nodes) without requiring widespread installation of PQMs.

The solution enables near real-time system-wide harmonic estimation with high accuracy (target <1% error), supporting early detection of exceedances against Grid Code and ENA G5.5 limits. This reduces compliance risks, avoids connection delays and penalties, and improves the accuracy of harmonic specifications for customers.

The quantified financial benefits are significant. As shown in the CBA summary, without innovation NGET would likely install ~200 PQMs over 10 years to achieve full coverage, resulting in:

£6.0m PQM installation cost
£4.65m engineering assessment cost
Total: ~£10.65m baseline cost

With the proposed solution:

PQM installation is reduced to ~£2.4m (<80 units)
Assessment effort is reduced to ~£2.42m
Total cost (including project and deployment): ~£5.46m

This delivers:

Benefit over 10 years: ~£4.71m
Annual benefit: £471,400
Benefit–cost ratio: 8.39
Operationally, the project reduces harmonic assessment times from days or weeks (offline studies) to near real-time analysis, enabling faster decision-making, better targeting of mitigation actions, and improved planning of PQM deployment.

Overall, the project transforms harmonic management from a reactive, model-driven process into a proactive, data-driven capability, delivering measurable cost savings, improved compliance performance, and enhanced network reliability, while supporting efficient integration of low-carbon technologies.