Issue 54: Swarms and Our Electric Future

Hope the summer is kicking off for you on a swell (but not sweltering) note. 

This issue of PM is all about the Future of Electricity. It is also the focus of podcast Episode 4. You can subscribe here, listen on    

Spotify, Apple, Amazon, or watch on YouTube. And we are now on every single podcast platform out there! Do like, subscribe, and (because you’ll be helping us) share with at least one friend!

Enjoy this issue of PM.

Ps: The text below is an update to my 2017 article.

————————————

Swarms and Virtual Power Plants

Every Tesla currently driving thousands of miles is gaining knowledge of how to navigate our roads. With every additional mile driven by each car, more collective knowledge is gained and dumped in a central database where it is analyzed in real time using machine learning. We hear a lot about how cars will be self-driving in a few years. This is also the case for every solar panel and home battery currently connected to the grid.

Less discussed is the knowledge the collective system is gaining in how to manage the energy stored in the hundreds of thousands of Teslas. In his paper ‘Engineering Antifragile Systems: A Change In Design Philosophy’, Kennie H Jones suggests

A swarm is a collection of autonomous vehicles operating for a common purpose that are not centrally controlled. Thus, they must self-organize and corporate to complete a mission. Another characteristic is that a swarm is composed of large numbers of relatively inexpensive (in the grand scheme) units that are expendable: 80% failure of individuals may still result in 100% mission success. 

This, I believe, is the central reason why we should lean in on the idea of Virtual Power Plants. The distributed resources that make up the ‘plant’ can act like a swarm. While the swarm is useful as a tool for managing the car as a unit of autonomous vehicles, it is also quite useful for managing the distributed energy resource which is the battery in every single Tesla. The swarm gathers data about each individual battery and some of this knowledge is in the form of

  • the number of optimal cycles before a battery needs to be recharged

  • cost of the energy being used and

  • power being discharged from the battery and how it affects the battery life.

This data, and much more, become vasts amount of knowledge through big data analysis (and even machine learning). This knowledge becomes a real-time insight into how to manage a virtual power utility. Add Powerwalls and SolarCity panels and the swarm becomes a utility managing each Tesla Model 3/Solarcity panel/Powerwall as a distributed energy resource (DER) node on the grid. Add Chevy Bolts and BMW i3’s and this becomes less about one company’s disruption and more about the upending of an old outdated business model. And boy does our grid/utility model need some updating (Gretchen Bakke’s ‘The Grid’ lays it out). The joke that goes around in electricity circles is that Insull - early 19th-century pioneer of the centralized utility model - would still recognize the electric grid that we have today. (Book: Empires of Light by Jill Jonnes) As I mentioned in my article about the true value of Tesla, this is the disruption that is being unleashed with the rollout of distributed energy resources. And this disruption is being borrowed straight out of Nature’s playbook.

Click image to watch on Youtube.

Biomimicry

My first immersion in all things Biomimicry came at SXSW Eco 2015. I listened to a few sessions and talks including one by biologist, author and innovation consultant, Janine Benyus. Biomimicry is simply using designs inspired or copied from nature to solve human problems. According to Janine Benyus in her book ‘Biomimicry’, most biological systems have a communication system that enables the system to hold in check any waste and reward efficient behavior. She suggests that the most mature and efficient systems are not ones that broadcast the necessary system control messages from the top, but ones that, from the grassroots, disperse messages throughout the system to manage this structure of rewarding efficiency, preventing and punishing waste. In layman’s terms most organic systems are not centrally managed but still function efficiently. And the smartest amongst us have always borrowed from nature to build technology; While he was a student at Harvard in 1951, Marvin Minsky, who ended up at MIT and is one of the founding fathers of AI, used a simple form of reinforcement learning and mimicked a rat navigating a maze.

This mechanism of information dissemination, which is like a rich feedback system, enables the system to adapt quickly when changes happen. This mechanism also helps the system maintain its integrity despite any attacks. It’s what we see in how leaves manage to capture, utilize and resource the rest of a tree to which they are attached without becoming burdensome to the whole tree. This is why leaves shed from trees when they no longer contribute to the integrity of the tree. The shed leaf decomposes and becomes manure and nourishes the other leaves/tree. This system adaptation looks a lot like antifragility (Book by Nassim Nicolas Taleb); the system gets stronger from any shock to a part of the system. It’s what a swarm enables.

By mimicking nature, a swarm of electric vehicles, solar panels, batteries, and energy efficiency devices in homes can be made into a virtual power plant made up of many nodes. If a node is attacked or compromised, say a cyberattack on a home, the virtual power plant does not get compromised and just cuts off that node. But the virtual power plant learns from that node. The virtual power plant gets stronger from the real time insight it gained from the compromised, but now disabled, node. The swarm/virtual power plant benefits from antifragility (and of course, resilience).


Antifragility or ‘death by a million cuts’

When it comes to innovation and business model changes, utilities are selling a message to their shareholders that this transition to an antifragile grid will play out without too much disruption. The utility industry is failing to inform its shareholders that the future Antifragile business model owes no debts to the structures that exist today. 

Don’t mistake this view for naivete about the complexity of the widespread system that is our current grid and how difficult it will be to change it. Because Tesla is selling expensive cars to a small subset of the market does not equate to disruption. The company is probably over leveraged at this point but it speaks to a possible and disruptive future. As I discuss in my book, The Antifragile Grid, while there is still a lot of work to be done to provide the infrastructure for electric vehicle mass adoption, the infrastructure is already being laid in the smart home/a node on this grid. Considering the state of our infrastructure, generally and specifically to the utility industry,  and the trends that Reza and I discuss on the podcast I think we are a lot closer to the swarm reality than we realize. Refashioning Hemingway’s quote from ‘The Sun Also Rises’ goes “How did the old utilities die?” Two ways. Gradually, then suddenly.”.

But, and this is quite important, even as we plow our way ever forward towards a more resilient and green future in the west, we must not forget the necessary sustainability of the communities we are now relying on to (literally) power this future. Because, even if just a few countries remain fragile, we all (across the Earth, remain fragile.

Till next time

Seyi

Previous
Previous

PM 55: The 20 Laws of Cities.

Next
Next

Issue 53: Future of Infrastructure