Smart grids can solve multiple problems now and throughout the lifetime of the energy transition. The supply from hundreds (existing + new wind, solar, etc.) and ultimately millions (rooftop solar, EV batteries) of power sources can be matched in real time with demand from all major customer devices (air conditioners, water heaters, batteries, EVs). Dynamic pricing keeps the peaks low and the troughs high. It takes the pressure off total load, allowing grids to evolve more slowly and next-gen technologies to come online when they’re ready instead of us rushing to build new capacity with old high-emission generators. That’s why in the U.S. the DoE’s Pacific Northwest National Laboratory has modelled Texas’s grid, explains Jim Conca. The results reveal that peak loads can be reduced by up to 15%, saving up to $5bn/year in Texas alone, or up to $50bn/year if deployed nationally. That’s equal to the annual output of 180 coal-fired power plants. Through the “transactive” agreement between consumers and utilities, both save money. The smart grids and AI used to price and control time-of-use are the enablers.
A recent study by DOE’s Pacific Northwest National Laboratory shows that grid operators can use something called transactive energy coordination to engage and use thousands or millions of large-area flexible distributed energy resources (DERs), such as air conditioners, water heaters, batteries, and electric vehicles, to help them operate the electric power system.
It can help keep the U.S. electric grid stable and reliable and would be a win-win for both consumers and utility operators.
The largest ever simulation of its kind, modelled on the Texas power grid that failed so spectacularly last year, concluded that consumers stand to save about 15% on their annual electric bill by partnering this way with utilities.
In this system, consumers would coordinate with their electric utility operator to dynamically control big home energy users, like heat pumps, water heaters and electric vehicle charging stations, even in their own garage.
“Transactive” agreement between consumers and utilities
This kind of flexible control over energy supply and use patterns is called “transactive” because it relies on an agreement between consumers and utilities. But a transactive energy system has never been deployed on a large scale, and there are a lot of unknowns. That’s why the Department of Energy’s Office of Electricity called upon the transactive energy experts at PNNL to study how such a system might work in practice. The final multi-volume report was released last week.
Hayden Reeve, who led the team at PNNL, said “Because Texas’s grid is quite representative of the nation’s energy system, it not only enabled the modelling and simulation of transactive concepts but provided a reliable extrapolation of the results and potential economic impacts to the broader U.S. grid and customers.”
Saving the U.S. $50bn/year
The simulation showed that if a transactive energy system were deployed on the Electric Reliability Council of Texas (ERCOT) grid, peak loads would be reduced by 9 to 15%, translating to economic benefits of up to $5 billion annually in Texas alone, or up to $50 billion annually if deployed across the entire continental United States. The savings would equal the annual output of 180 coal-fired power plants nationally, and would go a long way to avoiding blackouts.
The vulnerability of centralised power sources, and also wind & solar
By now, most people have experienced or witnessed how weather extremes and natural disasters can wreak havoc on our grid. That vulnerability is magnified by our reliance on a few centralised power sources and a grid system that sometimes struggles to match supply with demand. Further, decarbonisation of the electric grid will mean that more and more power will come from different kinds of renewable energy sources, like wind and solar. So, avoiding sudden spikes or dips—power brown or black outs—becomes paramount.
The study results indicate that a transactive energy system would reduce daily load swings by 20 to 44%. And as more electric vehicles come into use, the study, perhaps counterintuitively, showed that smart vehicle charging stations provide even larger electric peak load reductions because they offer additional flexibility in scheduled charging times and power consumption since the customer can just key in when they need the vehicle fully charged.
A smart grid is a shock absorber
“A smart grid can act as a shock absorber, balancing out mismatches between supply and demand,” Reeve said. “Through our study, we sought to understand just how valuable effective coordination of the electric grid could be to the nation, utilities, and customers. Working with commercial building owners and consumers to automatically adjust energy usage represents a practical, win-win step towards the decarbonisation of the electrical, building, and transportation sectors without compromising the comfort and safety of participating homes and businesses.”
One key component to this strategy is adoption of smart appliances and load controls. These dynamic resources can learn how to consume energy more efficiently, adjusting their use for brief periods to free up electricity for other needs.
For example, instead of charging an electric vehicle in the early evening when energy demand and price is high, transactive energy participants would rely on a smart load control to delay charging their vehicle until demand is low and electricity cheaper. This approach not only reduces stress on the existing grid infrastructure, it allows utilities more time to plan for next-generation energy storage and distribution infrastructure that is currently in development.
In a transactive energy system, the power grid, homes, commercial buildings, electric appliances and charging stations are in constant contact. Smart devices receive a forecast of energy prices at various times of day and develop a strategy to meet consumer preferences, while reducing cost and overall electricity demand.
A local retail market in turn coordinates overall demand with the larger wholesale market. All parties negotiate energy procurement and consumption levels, cost, timing and delivery, in a dynamic pricing scheme.
While this concept may seem futuristic, it is already here, being deployed in a demonstration project in the city of Spokane’s Eco-District as well as in Chinese cities like Shanghai.
Modelling 100 power sources and 60,000 customers
Using Texas’s primary power grid (ERCOT) as the basis for PNNL’s analysis, researchers created highly detailed models that represented the ERCOT power network, including more than 100 power generation sources and 40 different utilities operating on the transmission system. The analysis also included detailed representations of 60,000 homes and businesses, as well as their energy-consuming appliances.
Researchers used the models to conduct multiple simulations under various renewable energy generation scenarios. Each simulation demonstrated how the energy system would react to the addition of differing amounts of intermittent power sources, such as wind and solar. The research team also developed a detailed economic model to understand the yearly cost impacts for operators and customers.
Modelling the installation costs, too
Finally, they looked at upfront costs associated with labour and software expenses, as well as the costs for buying and installing smart devices in homes and businesses.
Overall, the PNNL research showed clear benefits of reimagining how the electric grid could accommodate a future where renewable energy is a much bigger contributor to the grid and where electricity powers more of our transportation needs.
The world’s largest digitally-connected energy storage network
But PNNL isn’t the only company working on AI-assisted grid energy operations. Stem operates the world’s largest digitally-connected energy storage network, referred to as clean energy intelligence (see below) using its Athena AI software.
Athena’s machine learning algorithms generate multiple forecasts – about weather, prices, solar generation, energy demand, and other factors – then formulate a strategy for maximising energy use and value (see video).
According to Stem’s Chief Technology Officer, Larsh Johnson, “AI software is being deployed to accommodate the addition of battery storage onto the grid as that technology becomes more mature and is commercialised. AI technology determines the most optimal and economic times to recharge the batteries and to release the energy onto the grid. AI software helps the operators forecast what individual customers’ load patterns are going to be, when they’re going to be consuming power and what the cost of power will be during different times of day.”
It’s the AI that makes a smart grid smart.
James Conca is an earth and environmental scientist and a regular contributor to Forbes magazine