Every story these days seems to connect its message or theme to artificial intelligence (AI). Some of these connections have validity, while others are a “reach” for others who want to appear on-trend and relevant. The latter are of interest primarily from a “really, is that so?” perspective, yet some legitimate ones are worrisome.
One AI-linked theme that has credibility is possible synergy — I hesitate to use that word, but sometimes it makes sense — between the management of automotive electric-vehicle (EV) charging and AI (Figure 1). For example, it is logical to use AI enhancements (whatever they may actually be) to extend battery life by smarter managing the critical cycling and minimum/maximum levels and matching them up with the driving patterns of that vehicle.
On the other hand, some commentators are claiming that AI will not only improve EV charging but will also allow it to work many wonders with no apparent downsides. Whatever you wish or dream, AI-driven charging will make it so.
For example, a recent article argued that AI-driven charging working with software-defined vehicles (SDVs) will optimize the performance of the entire EV energy path. Among other capabilities, it states that “AI-enabled smart energy solutions also manage charging schedules based on user preferences, predictive analytics, and the availability of on-site renewable energy sources such as wind or solar.”
Maybe that is so under ideal conditions, but I suspect that there are too many variables, unknowns, contradictory objectives, and “people problems” for the AI to determine an optimum solution to this complex problem in the real world. After all, one person’s view on optimization and priorities is another person’s undesirable tradeoff situation.
Further, it’s unclear if whoever or whatever is setting the priorities and chain of command for the AI-driven charging system is putting the needs of the vehicle user ahead of those of the system (what we’ll call the “general good”). Perhaps the system will decide that the needs of the grid at that time — and, by extension, other users — are more important than yours.
One of the harsh lessons of optimization studies is that there are at least two paths to macro (global) optimization. One is to optimize locally, and then those many local optimizations add up to a global solution; think of the drop-off route of the package-delivery truck. The other says that sometimes you must accept sub-optimum performance at one or more locations to trade off for global benefits. In other words, someone has to sacrifice (or be sacrificed) for the more significant large-scale benefit; you just hope it is not you.
What about V2G?
The article also lauds the potential beneficial impact of AI on vehicle-to-grid (V2G) energy management (Figure 2). V2G has always seemed to be an illusory and perhaps even dangerous dream. The concept is that everyone will connect their EVs to the grid even when not charging. Therefore, they can bolster the grid when its primary power source is reduced or down, functioning as a “free” and distributed battery energy storage system (BESS). In turn, the grid will replenish those batteries as soon as possible.
This seems to be a somewhat circular arrangement where each party in the transaction points to the party across the table and optimistically assumes, “Hey, I’ll just get the energy I need from them.” It assumes lossless back-and-forth transactions; it’s like A borrowing money from B, and then B borrowing money from A. Don’t laugh; there have been cases of corporate malfeasance where large-company subsidiaries have done this to “doctor” their balance sheets, and it doesn’t end well.
I don’t see owners of charged EVs easily relinquishing control and allowing the grid to take that hard-won energy back, even if it is for the “common good,” as the AI algorithm insists. The AI-based system may have a different perspective than yours on retaining the energy in your vehicle’s batteries.
Even if all this is worked out, there’s the presumption that little or nothing can be done wrong or go wrong with the system. Yet experience has shown us that “corner cases” happen, especially when many users and variables exist. All the rules, exceptions, and special cases soon add up and start to look like the cycles, epicycles, and more epicycles that were used to predict the motion of heavenly bodies in the geocentric Copernicus system — that is, until Galileo said, “Let’s simplify all this complexity by starting fresh, and with a heliocentric approach.”
What if you insist on charging and the system says, “Sorry, it isn’t a good time for you to charge right now, so just come back later”? Once the system becomes that smart, it may presume it knows better than you what makes sense and is a priority at any given time.
This entire situation may become similar to the dramatic ending scene from the classic 1968 movie 2001: A Space Odyssey (a movie that is now more cited than seen). The omniscient and apparently sentient HAL 9000 computer that runs the spaceship has taken over the craft and the project itself, implementing what it feels is the optimum strategy to ensure the mission’s success.
In a scene near the end, which is somewhat prescient to our AI scenario, the remaining astronaut (Dave Bowman) recognizes the danger in which HAL has placed him and would like to escape the ship — but finds he can’t, as his low-key, almost monotonic dialogue with HAL’s all-seeing and-hearing, somewhat creepy red sensor reveals, Figure 3: “Bowman: Open the pod bay doors, HAL. HAL: I’m sorry, Dave. I’m afraid I can’t do that. Bowman: What’s the problem? HAL: I think you know what the problem is just as well as I do. Dave Bowman: What are you talking about, HAL? HAL: This mission is too important for me to allow you to jeopardize it.”
I’m not saying that AI doesn’t have a legitimate or beneficial role in charging vehicles, managing the grid, or directing their interactions — just don’t oversell it to others or yourself. When dealing with critical functions such as large-scale power storage and transfer, it’s important to be realistic, somewhat humble, and very cautious.
Related EEWorld content
- Vehicle-to-grid is technically feasible, but what’s the reality?
- Are we ready for Vehicle-To-Grid (V2G) technology?
- How does vehicle-to-grid technology work to improve sustainability?
- Energy From Electric Cars Could Power Our Lives — But Only If We Improve The System
- With so many mandates, can successful designs still be achieved?
- How too many tradeoffs can kill a project
- What battery chemistries are used in grid-scale energy storage?
- How does MBSE work for EV and stationary battery energy storage systems?
External references
You may also like:
Filed Under: Charging, FAQs, Vehicle-to-Grid (V2G)