But how energy is generated, transmitted, and distributed as renewable power becomes more important means that the power industry is itself having to adapt and change.
Research conducted by Ernst & Young back in 2020 identified a series of key challenges for the industry, such as the rapid growth in renewable energy sources and the need for innovative storage technology, as well as the impact of electric vehicles (EVs) and the costs associated with moving electricity efficiently around the grid.
The rapid electrification of transportation, buildings, and industry along with other technological advancements means that the power industry is looking at a broad range of new technologies such as battery storage and artificial intelligence to not only grow capacity but to maintain current production levels and distribution. Consequently, it also needs to attract talent to deliver innovative new services and technologies as well as address new challenges such as cybersecurity concerns.
Looming over all of this is the issue of climate change and the use of renewable sources of energy and the implications that has for the wider power grid. With the rise of renewables, the demand for more predictive modelling rather than relying solely on historical data to better manage energy production and distribution is becoming an important issue.
According to Mckinsey and Company climate risk and resilience modelling are now more important than ever, providing operators with the information needed to make ‘the right calls at the right time’, whether through the better mapping of assets, creating vulnerability curves and risk heat maps, modelling detailed grid-impact scenarios and identifying and testing resilience measures.
Unpredictable weather
In the past year Europe has seen record-breaking temperatures and experienced increasingly unpredictable periods of weather, highlighting both the risks to and the vulnerabilities of the existing power grids.
Nvidia is one of a number of companies that are working to enable smarter, more resilient power grids and is involved in projects across all aspects of the industry - from renewable power generation to smart meters – that are increasingly using artificial intelligence (AI) and machine learning.
According to Marc Spieler, Nvidia’s Global Energy Director, while the industry is embracing new technology it remains very ‘conservative’.
“When you look at the current grid infrastructure it is quite old and was designed to send electricity in one direction. That is, you generate power and then send it our across the grid.
“Today’s modern grid network is distributed and has to be able to pull and put energy back into the grid when it’s required. The networks we have just weren’t designed to do this,” explained Spieler.
According to Spieler smarter grids are required that are better at leveraging and distributing energy, especially renewable energy such as solar and wind. Energy usage derived from renewables needs to become more reliable and secure.
Smart grids use sensing and measurement technologies to deliver dual communication between end-users and power companies and can automatically detect and respond to problems, addressing maintenance requirements quickly and efficiently.
“If we don’t start dealing with the problems impacting the grid, congestion will become a serious problem. By building ‘smarter’, you will create more resilience in the network and that requires greater visibility as to how electricity is being consumed. The transmission of electricity is a balancing act, so we need to be better at managing distributed power needs in terms of supply and demand,” explained Spieler, “and by adding more intelligence into the network, you will also be able to offer users more ways to participate and to actually feed surplus power into the grid.”
At the end of 2021 Nvidia announced that it was collaborating with Utilidata, a grid-edge software company, in the development of a software-defined smart grid chip powered by its AI platform. The chip is intended to be embedded in smart meters to enhance grid resiliency and to better support the integration of renewables – or distributed energy resources - such as solar, storage, and electric vehicles (EVs) into the existing power grid.
The US Department of Energy’s (DOE’s) National Renewable Energy Laboratory (NREL) plans to test the software-defined smart grid chip as it looks to scale and commercialise its Real-Time Optimal Power Flow (RT-OPF) technology, which was developed to ensure stable and efficient grid operations incorporating renewable sources of power.
“To date, the scalability and commercial potential of technologies like RT-OPF have been limited by single-use hardware solutions,” said Santosh Veda, Group Manager for Grid Automation and Controls at NREL. “By developing a smart grid chip that can be embedded in one of the most ubiquitous utility assets – the smart meter – there is the potential to enable the wider adoption of the technology and to redefine the role of edge computing. Enhanced situational awareness and visibility from this approach will greatly benefit both the end customers and the utility.”
Processing capabilities at the edge are growing massively. Nvidia is developing chips capable of conducting over 21 trillion operations per second.
“We’re designing for the future,” explained Spieler. “These devises will need to be deployed over a 10–15-year period so we need to be able to update the software over the air rather than have to replace the hardware. The future grid is a software defined one.”
Growing complexity
The complexity of the power grid is expected to grow, especially as more distributed energy resources and intermittent renewables come online.
“Bringing AI to grid edge operations will increase resiliency and reduce energy consumption and costs to consumers through new capabilities in meters,” said Spieler and will mean that utilities and consumers will no longer be locked into proprietary systems that have tended to hamper innovation by limiting third-party applications.
Great consumer participation in energy markets and incentivising commercial users to help in load balancing are seen as being vital when it comes to building more resilience into the grid and the development of microgrids and customers taking part in ‘dynamic load support’ by turning off high load equipment during peak periods will ensure better stability, according to Spieler.
“Smart grids need to be managed in real-time and that encourages the more efficient management of generation, transmission and distribution,” suggested Spieler. “With renewables you need to monitor the reliability of the network especially as you are using more ‘intermittent’ forms of energy. In terms of transmission, you need to be able to manage load capacity as you move excess capacity around the grid.”
More intelligence, therefore, is the bedrock on which the smarter grid will be built and will enable consumers not only take from the grid but sell back to it.
“With the development of microgrids users will have the opportunity to sell excess capacity back into the grid, so long as they have invested in solar of battery wall technology. It will give them the capability to buy or sell to the grid at a time of their choosing.”
This, however, raises the issue of sufficient storage.
“We could see individual houses, towns, or regions acting like local utilities when it comes to the transmission or electricity but because storage remains an issue, all of this will have to be handled in real-time, again highlighting the importance of adding intelligence and edge processing.”
According to Spieler, the economics of power generation are changing.
“Renewables are becoming more affordable but if we want to use them effectively, we need to be able to better predict weather conditions.
“Nvidia has been heavily involved in developing digital twins – the company’s Earth 2 project is a supercomputer developed to predict climate change – that can predict conditions whether that’s cloud cover or wind strength to create a more accurate picture of our weather and what that means for renewable sources of energy generation.
“If we can better predict weather conditions, so we can better handle issues around storage and whether we need to rely on more traditional forms of power,” according to Spieler.
Smart grids are not only seen as empowering consumers by providing them with more control over their use of energy but will make sure that energy can be properly leveraged, stocked and distributed no matter how far a city may be from wind and solar farms, while at the same time reducing waste.
While the future grid will be software defined, hardware innovation still plays an important part when it comes to developing these new networks. Companies have developed solid-state circuit breakers that can modulate flow rather than just switching it on or off while bi-directional AC/DC conversion inverters have been developed that can enable cheaper and much smaller solar panels which will help in the decentralisation of power production, distribution and storage.
The trend towards a more decentralised grid, however, does bring with it additional challenges in particular the threat from cyber attacks due to the increased attack surface that will be available to ‘bad actors’
“Whatever the location the grid is defined as critical infrastructure and with more technology being deployed, whether solar inverters or electric vehicles, so there will be a need to communicate with the grid. “Consequently, cybersecurity at every level will be required as will be the ability to update the software stack,” explained Spieler.
There has certainly been a lot of research into whether smart meters, sensors and advanced communication tools being deployed could result in greater vulnerability to cyberattacks.
“Smart meters retain an immense amount of data pertaining to consumer information, utility consumption and a single breach could easily compromise the entire system,” said Spieler.
“In fact, US cybersecurity experts are now urging utilities to focus less on costs and rather on better risk management. If someone attacks a transmission line or a sub-station the consequences could be massive.”
Smart grids are also expensive to install and deploy and the financial resources required are immense.
Many argue that while it might be ideal for smart grids to be implemented everywhere and rely solely on renewable energy, fossil-fuel produced electricity will remain a key component for many businesses going forward.
As Spieler has explained AI will be at the heart of smart grid solutions, but it will also have an increasingly important role in helping consumers better understand and manage their energy consumption.
Smart grids will produce a vast amount of data, which can then be exploited and turned into valuable insights covering the weather, how demand is fluctuating and the location or generation assets and will help operators to make better decisions.
Today’s power grids are undergoing radical change and the current and ongoing energy crisis is only going to accelerate those changes not only because they are necessary as renewables play a bigger part, but as a matter of urgency considering the changing geo-political considerations that we are all now having to live with.
