AI for Energy - a Power Combo
“Then there is electricity,—the demon, the angel, the mighty physical power, the all-pervading intelligence!” exclaimed Clifford. “Is that a humbug, too? Is it a fact—or have I dreamt it—that, by means of electricity, the world of matter has become a great nerve, vibrating thousands of miles in a breathless point of time? Rather, the round globe is a vast head, a brain, instinct with intelligence! Or, shall we say, it is itself a thought, nothing but thought, and no longer the substance which we deemed it!”
These are words from Nathaniel Hawthorne’s book, first published in 1851 -? The House of the Seven Gables (link).?It’s beautiful the ways these words imagined the impact of electricity for our planet. (I believe similar sentiments can be expressed for AI today.) It was circa 1925, 50 years since the creation of the Edison Electric Illuminating Company of New York, when electricity was widely available for most residents in America. Almost a century later now, how do we fare globally? Not very well. Even today 760m people across the world (link) have zero access to electricity and ~1.5B have intermittent access. With AI, there is a path towards energy abundance and affordability. It’s very interesting, intriguing, and inspiring to be living in such times where we all can play a role in enabling that.
Here are the three distinct “energy” scenarios that we humans experience today. These point of views are all heavily influenced by my own personal experiences and my hope is we collectively converge towards the third scenario. With AI, I am optimistic.
Limited Availability and Affordability. For most of my childhood in Nepal, energy/electricity was in scarce supply - a function of limited availability and affordability. When it was available and for those who could afford, it was valued, conserved, and sometimes treated as a novelty. But mostly the importance of energy/electricity was felt through its absence. It meant less than 12 hours a day of functioning light switches, burning kerosene lamps, lots of bonfires, excitement around the first introduction of propane tanks - these are all part of my memories. This isn’t a sob story. It wasn’t just my story but it was true for everyone I knew.?
Pure Abundance. For over a decade now, I have lived in the US. From being amazed in the beginning to getting used to all that’s available here, there’s nothing but gratitude I feel to have had this privilege - the privilege of experiencing true and pure abundance. Today, I end up spending more time deciding what kind of wifi light switches I use with no concern about the availability of power. I take it for granted the 24/7 availability of it. The comfort that comes from seamless heating and cooling, the skills that are honed from using tools of all kinds - oh! the power tools… more projects than I have time for. Just the privilege to take it all for granted.
Abundant and Affordable. Our vision for our planet should be abundant AND affordable energy for everyone. According to the International Energy Agency (IEA), renewables (solar and wind, collectively) will surpass coal-based power stations in 2024 (a year earlier than previously projected). Then there is nuclear - the promise of safer fission and the possibility of fusion. Either way, the path towards global abundance is now a viable one. With renewables and micro-grids, the huge capital expenditure won’t be a bottleneck for smaller and developing economies.
However, affordability, among other things will be highly dependent on the grid’s ability to predict (and balance) energy supply and demand. The challenge - the sun doesn’t always shine and winds don't always blow. This variability makes it harder to balance supply and demand. And without accurate predictability, it’s harder to schedule power and bring down the overall cost. AI can help - neural networks can be trained based on weather patterns to make these predictions days in advance.?Google tried it for their wind farms and demonstrated a 20% improvement in the economic value attained from their ML-powered wind farms.?
While accurate supply/demand prediction remains key, enabling global affordability will require continuous cost reduction (of course, along with short-term aids and subsidies). There are over 50 specific use cases around maintenance, monitoring, optimization, and efficiency improvements within and outside the grid that can all contribute towards lowering costs. There have been a lot of sensors, smart meters, and communication networks installed in the last 20 years -? so there’s no shortage of data. For example - ?the global fleet of wind turbines is estimated to produce more than 400 billion data points per year.
领英推荐
More uses for AI across power systems are likely to increase in the future. A few that come to mind.
These are just a few examples. For every data/energy curious person today, there are plenty of opportunities to dive-in. If you are interested in this space in any way, it’s the best time to invest in AI - anywhere from streamlining data collection, their processing and storing, to analyzing these extreme data sets using better-than-ever AI toolkits available today. In effect, all of this will change energy economics for good - when the unit price goes down, it’s only logical to get more people to use it.
So, let’s go - analyze your energy usage, build an algorithm to automate your lights, open source it, share it, have others improve upon it. In effect, you, me, your energy data, my energy data, and the ever-improving and now, the uber-accessible AI will all play a role in making energy available and affordable for everyone on this planet.