The National Energy System Operator (NESO) has published details of a new tool it will be using to improve stability studies for grid operators, using neural networks.
As more generation types come online, more advanced ways of assessing operational standards will be required. Much of the generation that comes online will be inverter-based resources (IBRs), which connect to a converter rather than a synchronous spinning turbine.
IBRs, like solar and wind generation and battery storage output, have faster dynamic behaviours than synchronous generators like gas or biomass. NESO’s current systems can handle these latter energy sources but the operator says that it is not equipped for a network dominated by IBRs.
Instead, it could use Electromagnetic Transient (EMT) simulation that can capture the faster fluctuations in power. However, that system is slow to run simulations and more complex models. So, NESO is working with software and consultancy firm Transmission Excellence, the University of Bristol and the University of Bath to develop a methodology for the use of neural networks to create a model for EMT that will be faster and offer the same accuracy as its slower parent model.
Neural networks are a type of machine learning that, theoretically, process data in a way that mimics the human brain. NESO says that by using this type of machine learning it can create accurate models to use in future stability studies. Increasing the speed at which EMT simulation happens will reduce safety margins and lead to possible cost savings, particularly across constrained system boundaries, NESO claims.
Ofgem funds LCP-led GeoGrid project
A group including LCP Delta, Northern Powergrid and E.ON have been awarded strategic innovation funding (SIF) for geothermal long-duration storage to help balance the grid.
The three companies, along with the University of Leeds, Leeds City Council and Star Refrigeration, were awarded the funding through Ofgem’s SIF scheme, which has also been awarded to transmission projects by National Grid in recent months.
This project, dubbed GeoGrid, will leverage geothermal long-duration storage to provide cross-vector balancing and advanced network management. The aim of the project is to lower network connection costs, reduce costly reinforcement and enhance grid resilience.
The trial will take place at the University of Leeds Geothermal Campus and hopes to demonstrate low-cost options for storing curtailed energy, which is projected to cost bill payers as much as £2.2 billion by 2030 if network infrastructure is not able to deal with the increased load from renewable energy targeted to come online.
Andrew Turton, head of consulting at LCP Delta, commented: “This project is of strategic importance for the UK electricity network as it addresses some of the most pressing challenges in our transition to net zero.
“By combining innovative geothermal storage with advanced modelling and systems thinking, we’re helping to lay the foundations for a more resilient, efficient, and decarbonised energy system.”