At its control centre in Ireland, GridBeyond provides demand-response services in a unique way. It allows for specific processes or machines to be temporarily shut down to balance the grid, rather than taking a company’s entire facility offline.
In early June, Australia’s new government committed to offering low-cost loans to boost the nation’s energy transmission system and support the growth of renewables, to the tune of $20 billion (USD 13.36 billion). But transmission is nothing compared to what needs to be invested in distribution networks, to support ebbs and flows of rooftop solar, and demand from electric vehicles. Software that enhances visibility over both the big and small ends of the network is emerging as a cost-moderating saviour in its abilities to analyse, model, forecast, and balance the grid as we know it.
At the big end of town, John K Ward, research director of the energy systems research program at Australia’s Commonwealth Scientific and Industrial Research Organisation (CSIRO), is connected to many local and international energy endeavours, and serves Australia’s representative on the International Energy Agency (IEA) Smart Grid Network. A strong recent theme of problem solving in the energy transition is that datasets and digital solutions from different countries will help global grids transform more quickly – and reduce effort wasted on reinventing the wheel in multiple jurisdictions.
In Australia, Ward says demand alone, through population growth, and electrification of industry and vehicles, will double the amount of energy moving through the system by 2050. If Australia’s optimistic green energy production goals were achieved, through exports of hydrogen for example, the system would be buzzing with six to ten times the volume it now carries. “Getting that increase in throughput without a commensurate increase in investment,” he says, “comes down to how you increase the utilisation of existing infrastructure and optimise investment in new infrastructure, including how you shape usage, storage and production of energy, so that you can have the minimum required to get that amount of energy through the system.” In the first instance, this means monitoring, via judicious deployment of sensors that report energy flows and disturbances. Second, the ability to control what happens in response to that information.
What we’re talking about is a simulator of the energy system, which can forecast what level of renewable generation can be expected in the coming 10 minutes or an hour or more – and generate potential scenarios of control/response faster than real time. The greater your understanding of the system’s future state and your ability to explore future scenarios, says Ward, “the greater your confidence can be that you’re going to be able to securely operate the system in any of those scenarios”.
Today’s digital systems, though, can perhaps simulate one scenario. The decision-making tool of the near future depends on getting the right data, without delay, and building models complex enough to generate multiple scenarios. The increasing prevalence of inverter-based solar systems at both rooftop and large scale, means algorithms will work with real-time data as well as implement look-ahead capabilities to predict system security into the future. Ward says the detailed analysis required to deal with the actual control system code of inverter based resources connected to the system, as well as the myriad inputs of real-time data exponentially increases computational load. The challenge is to make the leap “to reflect what those systems are actually capable of – not to oversimplify your models.”
For the software engineer, and the future system operator, Ward says, “It’s quite a different mindset, to go from retrospectively investigating events after they’ve happened, to proactively operating a system on the basis of faster than real-time detailed assessments of different potential scenarios.” Asked how this work is progressing, Ward replies: “It’s going as fast as we can!”
Monitoring edge assets
Visibility over what’s happening at any given time in the electricity system is a prerequisite. This enables total system control, and to effectively utilise existing transmission lines and distribution networks, and to inform decisions on where investment in increased capacity is most needed.
At the distribution frontline, Future Grid co-founder and CEO, Chris Law, says his company’s digital solution addresses both visibility of the low-voltage system and its management – which he says, “has become a burning platform for most utilities”.
Transmission is not the big problem, says Law, despite the major costs inherent in beefing up the high-voltage, compared to the low-voltage network. “The infrastructure that sits in the distribution network is massive – poles every 10 meters with wires to everyone’s home. Probably now in the trillions of dollars in Australia, it’s the biggest investment the country has ever made in its infrastructure.” And it’s under significant stress due to the number of solar panels already generating renewable energy from the country’s rooftops.
Merryn York, executive general manager of system design at the Australian Energy Market Operator (AEMO), confirms that AEMO’s 2022 Integrated System Plan forecasts an increase in rooftop solar capacity “from 15 GW today in the NEM, to 37 GW by 2030, rising to 69 GW by 2050.”
Typically, utilities’ response to lack of capacity is to build more, but the cost of building extra capacity throughout the capillary-like distribution networks would be astronomical, and of course it would be borne by the customers. Says Law, “This is what utilities around the world are talking about, although most will never say it publicly.”
Capacity for handling export of solar from rooftops, demand from EVs, and the voltage drops that occur when everyone is self-consuming home-made energy rather than drawing from the grid, varies across the Australian system. Future Grid puts incoming data from the many mini generators and consumers in any given network into the context of grid operations, to show what’s happening at a local level. “The goal is to give utilities visibility over all these changes, so they can start to figure out how to minimise their investment costs,” says Law.
“If they can see where certain assets in the grid are overloaded, or could be better optimised by shifting load to different times of day,” Law explains, “they can choose to target different strategies to minimise the impact.” For example, when people start smart charging their cars, Future Grid can analyse which parts of the grid will cope in the medium term, and have no need for immediate action, and which areas are already under pressure and when, and how different solutions (such as incentivising people to charge at different times) could mitigate that pressure.
Where upgrading is the only viable solution, says Law, his team aims to “make it very targeted and purposeful, so we minimise the broader impact on everybody else”.
Data processing and analytics are at the core of Future Grid’s work, but now that it has data and related outcomes from customers in Australia, Thailand and the US, for example, it is applying machine learning to derive further insights that will help everyone using its technology. “I don’t think anyone’s got more smart meter data from such a diverse set of networks in the world than we have,” says Law. The result, he says, is that new utility customers “get a system that’s kind of pre-trained, that already knows what to look for, rather than having to start from scratch”.
This year, another digital company with a similar name, different premise, but complementary renewables-boosting effect, launched in Australia with data and experience gleaned from other markets. GridBeyond started 15 years ago in Ireland, and now operates in many markets, essentially monetising demand response for major manufacturers and consumers of energy in the grid by providing grid-balancing services in a unique way.
Globally and in Australia, the company has numerous competitors, but its point of difference is that it operates at an asset level: a facility doesn’t need to suspend its whole operation to offer services to the grid. Instead, data-driven options help choose processes or machinery assets that have some latency to be responsive to grid needs, while other aspects of production continue.
Data amassed in various markets over time has provided GridBeyond with an “understanding of processes in cement making, steel production, glass manufacture, refrigeration facilities, pulp and paper production, and metal refining,” says Mark Davis, managing director of GridBeyond in the UK, Ireland, and Australia. In such industries it has already identified which assets can be ramped up and ramped down, and which respond well to being switched off and on again.
Australia doesn’t yet have a capacity market for energy – which would pay participants to be able to offer a certain amount of energy back to the grid as needed. The Electricity Security Board has started the process towards a capacity mechanism, and expects to have draft legislation to government ministers by February 2023.
In the meantime, Davis says, large energy consumers can monetise their flexibility services to the grid via the Frequency Control Ancillary Services (FCAS) market: “If you have one megawatt of available flexible load, you can put that in the FCAS marketplace through GridBeyond. If that asset can be off for up to six seconds, you earn a reasonable return from the Australian grid system,” because the operator doesn’t have to pay, say, a gas generator to fill the shortfall in supply. If you manage participation at an asset level, according to asset characteristics, some assets may be able to be offline for five minutes without causing disruption, and “you can stack these revenues together to make a profit,” says Davis. He claims, “We have clients in Europe and North America that are making millions of dollars they didn’t know they could sweat from their assets and their process.” In the UK, its most established market, GridBeyond controls some 700 MW of load, across around 500 sites.
Franco Santucci, executive director at Sentient Impact Group investors, has focused on the energy sector in various roles at EY, Deloitte and others, for more than 30 years. He says, digital management of energy infrastructure “is an area of significant underinvestment” and that “more needs to be done to model the energy system and different scenarios”.
On the digital adoption side, Law says most utilities are at base camp in an Everest analogy to how they could be using software solutions. It’s a slow process, says Law, “but, one of the fundamental things I can do is get them moving in the right direction. Then as problems occur, at least they have a foundation they can build on.”
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: email@example.com.