When demand exceeds supply, BAD THINGS CAN HAPPEN. This was the lesson we learned at the beginning of the pandemic when the demand for disinfecting wipes, masks and ventilators exceeded the supply. Chip shortages continue today to disrupt the automobile and consumer electronics sectors. It is essential to balance the demand and supply of goods to have a normal, stable society.
This same need for balance applies to electric power grids. This was a stark reminder of when Texas suffered a devastating and fatal winter freeze in February 2021. The rising demand for electric heat was matched by the supply issues caused by low wind-power production and frozen natural-gas equipment. This imbalance led to more than 2,000,000 households being without power for days. It also caused at least 210 deaths and economic losses of up to $130 billion.
Similar mismatches between supply and demand led to massive blackouts in August 2003 in the northeastern United States, Canada, India and Venezuela.
Three reasons are why the situation is unlikely to get better soon. The first is that as more countries decarbonize, electricity demand will rise. For economic and policy reasons, traditional coal and nuclear power plants are being phased out, removing reliable sources from the grid. Third, although wind and solar-photovoltaic are excellent for the climate and the fastest-growing sources for electric generation, their variability creates new challenges in balancing the grid.
How can grid operators maintain a balance between supply and demand while shutting down old power plants and ramping up variable generation? There are several options. One option is to modernize what we did in the past. That would be to build a massive, central infrastructure. This would involve installing large amounts of energy storage such as grid-scale batteries or pumped-hydro plants to store excess renewable power and connecting that storage with high voltage transmission lines so that the grid can supply and demand. This approach is being taken by China, which is an international leader. However, it is extremely expensive and requires a lot of political will.
We believe there is a better way. Our research at the University of Vermont focuses on coordinating demand and supply in real-time rather than scaling up the power-grid infrastructure. Our technology takes two ideas that make the Internet fundamentally scalable–packetization and randomization–and uses them to create a system that can coordinate distributed energy. These two data-communication concepts enable millions of users to connect to the Internet from billions of devices without central scheduling or control. These same principles could also be applied to the electric grid. Millions of electrical devices could be used to balance electricity flow within the local grid using low-bandwidth connectivity, small controllers and simple algorithms. Here’s how.
Billions of electric loads fuel the grid. They can be divided into two main categories: residential and commercial loads. Residential loads are the more distributed of both. The United States alone has more than 120 million households. These households account for approximately 40 per cent of the country’s annual electricity consumption. Residential customers don’t often think about optimizing their electricity loads while going about their daily lives. These residential loads can be called “devices” for simplicity. They could include lights, televisions, water heaters, and air conditioners. When operating a residential electric water heater. A typical unit uses about 4.5 kilowatts to heat water. The appliance runs for about 10 hours per day and consumes 10.8 kilowatts. The homeowner pays less than $2 per day for the water heater, assuming the rate is about 15C/kWh. The utility charges electricity at a variable cost, ranging from 4C/kWh to more than $100 per kWh during peak seasons. Sometimes the cost of electricity can be negative. Grid operators pay utilities to use excess power from solar or wind plants when there is not enough.
Utility companies have offered demand-response services for years. They can turn off water heaters and air conditioners at customers’ homes on a schedule. This is especially useful during peak times when demand is high. This approach works well if we only want to reduce load during peak periods.
If our goal is to balance the grid in real-time, when renewable generation fluctuates with the sun and wind, then operating devices according to a schedule based on past behaviour will not work. We need a more flexible approach that is responsive to peak demand. It also provides additional benefits such as frequency regulation, renewable smoothing, price responsiveness and renewable smoothing.
How can grid operators coordinate many flexible, distributed kilowatt scale devices with their own needs to provide a gigawatt-scale aggregate grid resource that responds to a variable supply? We found inspiration in digital communication systems as we pondered this question.
Digital systems can be described as your voice, email, or video clip and are represented by a sequence of bits. This data is broken down into packets transmitted over a channel. Each packet is then routed independently through the network to its intended destination. After all, packets have been received; data can be reconstructed in its original form.
This is a very similar problem to ours. Every day, the Internet is used by millions of people across the globe and on billions of devices. Users have their own needs and usage patterns, which we can call demand. However, the network itself is dynamic with its bandwidth or supply. However, the Internet’s demand and supply are constantly matched without using a central scheduler. As we have seen, the supply of power is becoming increasingly variable as a result.
We recognized this similarity and developed a technology called “packetized energy management” (PEM) that coordinates the energy use of flexible devices. Coauthor Hines is an expert in power system reliability. He has studied how transmission-line problems can cause systemic blackouts and cascading outages. Frolik, a communication system specialist, worked on algorithms that dynamically coordinate wireless sensor data in a way that uses very little energy. After a chance conversation, we realized that our interests overlapped and started to work together to find solutions to the problem associated with EV charging.