Thoughts on EVs without range limits
Disclaimer: I read a lot and I poke my nose into a lot of things — but as with many other geeks, my knowledge about many areas of science runs pretty shallow. I am very aware of this and usually keep my opinions to myself. I hate when people with limited knowledge try to teach me about things I happen to know in depth — so in choosing to inflict this on others, I do so with an open mind and with readiness to be corrected if some or all of my assumptions are wrong.
We are slowly edging towards the next generation of mass-market electric vehicles with the average range somewhere around he 300–350 km mark. This is a great development, but I am worried that while this may get us over the 10–15% adoption rate in North America, we will still not see people switching from internal-combustion engine (ICE) to battery-electric vehicles (BEV) — due to both continued range anxiety and bad experience people are bound to have when attempting to use their cars on longer trips. Unfortunately, attempting to solve one will exacerbate the other.
The problem with batteries
Energy storage is a hot topic now, with a lot of exciting breakthroughs happening all the time, but solutions remain elusive for cars, where good candidates must satisfy these three main requirements:
- energy density
- cost
- number of charge/discharge cycles
Cars need small, lightweight batteries that can quickly store and release a lot of power, can be charged/discharged (cycled) 3–4 thousand times before they start losing capacity, and cost cheap enough to keep vehicles affordable. Current best of breed are Lithium-Ion batteries that give us a 60 kWh battery pack good for about 350 km of ideal weather range at a cost comparable to average market sedans ($35,000-$45,000 USD).
The range of 350 km (215 miles or so) translates into 3 hours of driving — not nearly a match for the average range of an ICE vehicle, which are good for about 900 km on a full tank of gasoline. Moreover, while with the latter you can just stop and refill in under 10 minutes, charging electric vehicles is (and is going to remain) problematic.
The problem with chargers
Home EV chargers draw about as much power as your water heater or your cooking range, and top out at about 7 kilowatt (kW), which makes it pretty easy to calculate how long it would take to recharge your vehicle. Take the total size of your battery pack, divide by how much power is coming in through the charger, and then add another hour or so (because the last 10–15% must be trickle-charged due to how Li-ion batteries work). So, a car with a 60 kWh battery pack will take 60/7+1=9.5 hours to recharge on a “level 2” EV home charger.
A regular North American outlet will only give you 1.5 kW of power, so if you drive to visit your relatives with only a wall socket available to you, recharging your Chevy Bolt will take 60/1.5=40 hours (don’t bother adding the extra hour, as 1.5 kW is pretty much the definition of “trickle charge”).
Thankfully, there is fast-charge technology that can get you going faster. Nissan Leaf-compatible ChaDeMo fast-chargers provide about 50 kW of power, and will get your 60 kWh car charged up and ready to go at 80% of range in about an hour.
Tesla superchargers push 120 kW and will soon get upgraded to 150 kW, which will get you on your way in about 20–25 minutes. New standards are under development for 200 kW fast-chargers — and that should reduce your charge times to 15 minutes or less.
These charging times are for “perfect conditions,” at the ambient temperature of about 15–18 degrees Celsius (60-65°F). Too cold, and your battery cannot recharge at full speed. Too hot, and your battery must spend extra time cooling off so it doesn’t catch on fire. If you’ve ever held your phone while it’s charging, you know how worryingly warm it can get. Now imagine the kind of thermal physics you have to worry about when delivering 150-200 kW of direct current.
However, even if we had a perfect battery and always recharged in perfect conditions, we’d still have a problem on our hands.
Holiday rush
According to the US Bureau of Transportation, every major holiday Americans drive on average about 270 miles to visit their families. If you’ve ever travelled during such busy times and had to pull off to a gas station to fill up, you probably remember it being a bit of a mad house. I bet you probably had to wait your turn to gas up, which was thankfully quick enough, as it only takes about 5 minutes to pump enough gas to get you going for another 600 miles.
Now imagine how things would be if a large portion of the fleet were to be converted to electric vehicles. Even in the best possible case, recharging a 60 kWh car at 150 kW Tesla Supercharger speeds would take 15–20 minutes or more, and drivers probably wouldn’t want to sit in their cars waiting. They’ll want to stretch, get a coffee, grab a snack, or use the facilities. That would easily bump the average time at a charge station to 30+ minutes.
We’d need three times the number of supercharger stations than there are currently gas stations if we are to handle average holiday traffic — which would not be economically viable as for the rest of the year these stations would be sitting mostly vacant due to people recharging at home. You’d be sinking a lot of up-front costs into a network of stations that would never pay for themselves and would continually cost resources to upkeep — money much better spent improving the urban grid where most of the recharge load is going to be.
But if we don’t provide such a network of recharge stations, we’ll eventually reach critical mass where motorists will be stranded in vehicles unable to find a charge spot. The headlines will write themselves: “family with young children stranded on Christmas eve unable to recharge their electric car.” It will be a nasty black eye for EVs and will be a serious setback for many years.
My idea: Just in Time (JIT) power delivery
I think it’s a mistake to treat cars like spaceships or airplanes that must be self-powered because they are traveling through some kind of void. A vehicle journeying long-distance is going to do it on top of well-maintained roadways that already have extensive infrastructure servicing them— we just need to find a way to deliver power to the vehicle while it travels.
But the question is how to do it, right? After all, we can’t add a third rail or an overhead wire for individual cars — that would be impractical (not to mention terribly exposed to elements in a climate like that of Canada where I live). Such power needs to be delivered via a wireless mechanism that allows for a certain degree of misalignment and distance jitter between the transmitter and receiver — all while being powerful enough to beam roughly 20 kW needed for an electric vehicle to maintain highway speeds.
What we need is a powered transmitter that can travel with the car.
Magneto Dynamic Coupling (MDC)
Let me start first by mentioning what Magneto-Dynamic Coupling is. In most basic terms, if you take two permanent magnets and put them next to each-other, rotating one of them will cause the other to rotate as well because it will be affected by the changes in the magnetic field. This is exact same physics that makes the magnet on top of the table move when you move the magnet beneath, to the amusement of your audience.
Such wireless power transfer technology exists today and you can buy it on the market — produced by Elix Wireless. See their promotional video of the charger in action.
The cool thing about wireless power transfer using MDC is that it is pretty low-tech and doesn’t generate massive electromagnetic interference when transmitting power the way inductive chargers do. One of the main reasons we haven’t yet seen a wide adoption of wireless car chargers using inductive transmitters is due to powerful EM fields generated by the transmitters and the way they tend to interfere with electronics. High-power inductive charging is only safe when properly aligned with the receiver, but misalignment would fry anything electronic that happens to be in the line of the beam. Considering that cars are built to transport humans with both wearable and integrated circuitry (think someone with a pacemaker), we really don’t want to take a risk of turning on a powerful inductive charger while someone is in the vehicle.
An MDC-based transmitter, on the other hand, does not generate strong EM fields and is safe for humans even if they are standing right next to it while it is operating.
The less cool thing about MDC-based wireless chargers is that they produce a low hum of 65 dB (at 1 meter away), which is about how a vacuum cleaner would sound across a room. The transmitter operates at 115 HZ, so I would expect it to sound something like this. However, this is not something to worry about if we’re going to use them at highway speeds where the noises are already an order of magnitude higher.
Using MDC for Just-In-Time power delivery
Elix power chargers are already a real product used for wireless battery charging, but I say we use them for just-in-time power delivery to moving vehicles. They are cylindrical objects about 15 cm in diameter and would fit perfectly inside a reinforced vacuum pipe buried right under the surface of a highway.
Each transmitter shuttle would be mounted on a moving platform with power supplied by electrified rails on which the platform would be moving, the same as subway cars.
Same, viewed from the side:
The car travelling above this pipe with the moving MDC transmitter would be outfitted with two MDC receivers along the centre of the vehicle, in addition to on-board batteries capable of providing the car with 200–250 km (120–170 mi) of independent range for travelling over roads where the MDC shuttle network is not available (and for any extra power boost).
Two 10 kW receivers should provide more than sufficient power for a vehicle moving at highway speeds, as it takes on average 16–18 kW to travel at the speeds of 110–120 km/h (60–65 mph). When moving uphill or accelerating, additional power would come from the on-board battery. Similarly, when moving downhill or decelerating, the MDC receivers would be used to recharge the battery.
The receivers would be lowered from the underside of the vehicle to reduce the distance from the transmitter while the car is moving on maintained highways. Optimal distance between the transmitter and receiver is about 10–12 cm. Encountering hard debris larger than 5–7 cm in height would be extremely unlikely on busy highways.
The vacuum pipe with MDC shuttles would form a loop serving both directions of traffic.
Since many major highways are divided, the loop would serve “fast lanes” of both directions:
The loop would serve a sufficiently long stretch of straight road (say, 5–10 km), and perhaps the turns, too, if aligning with a moving vehicle proves possible there (more on that later). Each stretch would be served by an individual loop, with all of them forming a what I call “JIT Power Chain”.
The loops would be serviced by automated shuttle management stations located in the median that would provide the necessary power to the shuttles, maintain the vacuum in the pipe, and serve as independent communication centres for negotiating the power shuttle service for the cars. They would also have storage areas for unused shuttles and a simple robotic mechanism to take shuttles from storage and deploy them into the pipe network.
Transmitter and receiver alignment
There will be RFID transmitters embedded either directly into the top of the pipe, or into a special non-ferrous strip covering the road surface right above the pipe. The RFIDs do not have to be sequential, they just have to be unique.
The vehicle will use these RFIDs to find the exact centre position above the pipe, for example by using two RFID receivers on both sides of the vehicle and ensuring that signal coming from the RFID chips is the exact same strength at both receivers. If signal strength on the left side of the vehicle is higher than on the right sight, that means the car needs to move to the right until signal from RFIDs is equally strong. The RFID chips would need to be placed close enough together that would allow for continuous course correction in order to achieve the best alignment. There can additionally be an optical mechanism for such alignment, such as a colourful strip on top of the highway, but it can only be used as a secondary mechanism, as it is likely to wear out, be covered by dust, dirt, or ice.
Such auto-alignment would not require a fully autonomous car, but the vehicle would need to at least operate in a “smart cruise control” mode that is able to adjust both the speed and the bearing in order to achieve optimal alignment over the JIT power track.
Once the vehicle establishes the cruise control necessary for MDC operation, it would communicate with the JIT system and request MDC shuttle service (using mobile wireless such as 4G/LTE, etc). Once a session is negotiated, the vehicle would send its speed and the latest RFID reading back to the system — perhaps with the latest GPS data as well. This would allow the shuttle management service to know the vehicle’s exact speed and position on the highway in order to prepare the shuttles necessary for the next step.
At the beginning of the next JIT loop, the vehicle would be issued its own two MDC transmitter shuttles. After approaching the vehicle, the shuttles would enter into direct near-field communication with each MDC receiver on the vehicle, using a standard protocol such as NFC or Bluetooth (whichever proves more dependable — I’m going to guess NFC). Once NFC communication is established, each vehicle-mounted receiver would continuously send the ID of the latest unique RFID encountered to each of the shuttles, forming a continuous stream of “seen RFIDs.” The MDC shuttle would read the same RFIDs as it passes them and use this data to exactly match the speed of the vehicle for best transmitter alignment.
In other words, if IDs received via the NFC link arrive at the same time as IDs read by the RFID reader on the shuttle itself (with allowance for NFC latency) the positions of each transmitter/receiver pair are exactly matched. The vehicle can additionally send speed and accelerometer data across the NFC link as extra controls to ensure the best MDC alignment.
Once the length of the loop is completed, a new set of shuttles would be issued at the beginning of the next loop and can use the precise information from the previous session to minimize the time necessary to reestablish proper alignment with receivers.
Bypassing energy reconversion
An extra possible upside of MDC power transfer mechanism is that it is not necessary to convert the rotational momentum of the receiver into electricity, if we are going to convert it right back into rotational momentum a moment later. It should be possible to connect each MDC receiver directly to the drivetrain on each axle in order to minimize energy loss. If more torque is needed than can be provided by the MDC receivers, the vehicle would switch to battery power and the MDC can be used to recharge the battery instead.
Upfront cost of deployment
Initial roll-out would require a comparatively small amount of investment:
- There would be software development costs for shuttle management stations and for communication protocols across WAN links (for establishing a session) and across NFC links (for direct vehicle communication with MDC transmitter shuttles).
- There will be some R&D costs, but they would be pretty low, as this would use readily available technologies (MDC, LTE, NFC/Bluetooth, RFID).
- Shuttle management stations should be pretty simple installations that would only require a power source (regular grid connection to begin with), a vacuum pump, and a simple robot to take shuttles from storage and deploy them to the pipe (and vice versa).
- Shuttles themselves are very simple devices that only require an MDC transmitter and a microcontroller for communicating with the vehicle and the shuttle management station. For the initial roll-out, each station needs to have just a handful of such shuttles to satisfy demand.
- Loop track should be very cheap as it just needs a strong enough pipe to withstand an occasional weight of a vehicle’s wheel (it will be buried in the ground, but close enough to the surface that it needs to allow for occasional pressure stress from a heavy vehicle). With the pipe only needing to be 15–18 cm in diameter, this should not be difficult to achieve, especially considering natural structural properties of round pipes.
- Embedded RFIDs are very cheap and since they don’t need to follow any particular sequence, any damaged section of the loop can be replaced without having to reprogram transmitters.
- The actual process of laying the pipe would be the most expensive task, but as it would only require a shallow excavation in the road, it would require minimal interruption, workforce, or machinery. It can probably be done with a robot capable of cutting through asphalt, excavating a 20 cm trench, laying (or printing) sections of the pipe track, applying fine gravel, then re-applying asphalt (or using non-ferrous covers).
- Vehicles would not be using this system for free, but pay per kW consumed — which should still work out cheaper than gasoline in the long run.
- Political agreements with highway management authorities and car manufacturers is a whole another topic that I am going to leave entirely out of this write-up.
Major upsides
- The power is delivered continuously, negating massive spikes caused by fast-charge stations requiring megawatts of power at peak travel, each.
- Energy can be provided by renewables local to each JIT power loop. On divided highways, solar panels are perfect candidates for providing the majority of power as people tend to drive most during daylight hours. At night-time, when traffic is light, power can be delivered from the rest of the grid, including from wind turbines or from grid storage batteries.
- This system is entirely isolated from weather elements. There should be very little maintenance required to the tracks, unless they are directly damaged. This system is also entirely snowplow-friendly and can operate uninterrupted during major snowstorms.
- Parking lots can be outfitted with a disconnected loop, such that electric cars can be recharged without having to park in special “electric cars only” parking spots. A handful of shuttles can recharge every car parked overnight by travelling to each one in sequence.
- This would be especially handy for municipalities that want to provide a charging service for EVs owned by tenants living in apartments — just run the shuttle pipe underneath the curb parking. This would eliminate the need for urban EV owners to equip their own charging stations.
Scaling up
Once initial deployment is done and cars start using JIT power loops, adding capacity would be as simple as adding more shuttles to each management station and ensuring that there is enough electrical power supplied to each loop to satisfy demand — by adding more solar panels and grid storage batteries at each management station (stationary storage is a lot cheaper due to not needing to be light and dense enough to fit in a moving vehicle).
What ifs
What if there is a lot of traffic moving one way but not the other? It costs almost nothing to move shuttles in a vacuum pipe, so an “empty trip” along the other side of the loop would be all it takes.
What if traffic speed on one side of the highway is lower than on the other side? Cars in a traffic jam do not need 20 kW of continuous power and can use onboard batteries until the jam is cleared. There can additionally be a pipe directly connecting the two ends of the loop, running along the highway median, for dealing with such congestion — effectively it would become two independent loops with a shared pipe in the middle.
What if a section of the chain is inoperable due to malfunction? This is why having 30–50 kW onboard power packs remains important for use in the city, along secondary highways not offering JIT power, for any additional boost when 20 kW is insufficient, and for dealing with any unexpected malfunctions in the JIT system.
Why nots
Why not use plug-in hybrids for long-distance travel? Plugin Hybrid-Electric Vehicles are a lot more complex that pure EV-drivetrain cars. As such, they are inherently less reliable and require a lot more maintenance. Plus, we’re still burning fossil fuels and wasting great amounts of money refining them.
Why not use redux flow batteries for fast recharge? Redux flow batteries should theoretically allow “recharging” just as fast as with gas, but they still require an extensive network of refuelling stations and they can’t be recharged at home. To use redux flow batteries effectively, cars would probably need a hybrid power source, one from a battery rechargeable at home (e.g. Li-Ion), and one from the redux-flow battery for longer trips. This would make the vehicle too heavy and too complex, in my opinion.
Why not use hydrogen fuel cells you can refuel? Because every way you look at it, using hydrogen gas is a terrible idea, plus all of the above.
Why not battery swapping? It sounds great in theory, but batteries are likely to remain some of the most expensive components in your car. Each battery swap would effectively involve exchanging tens of thousands of dollars worth of your stuff for someone else’s stuff and I’m afraid the incentive for cheating or otherwise abusing this system would be too great.
Last words
Like I said, I’m not a mechanical engineer, highway architect, a car nut, or a chemistry junkie. I’m just an IT geek that wants a better world for his kids, and I see electric cars as one of the ways of getting us all there. I think the system I describe above is sane, workable, efficient, doesn’t cost incredible amounts of money to deploy, and doesn’t rely on unobtainium or other technologies that haven’t yet been invented.
I’m a bit conflicted about this only in the sense that this doesn’t get us away from our dependence on cars and doesn’t solve the problem of traffic jams. If something like JIT power loops can be implemented, it needs to happen hand in hand with further development into urban public transit that can get workers to their offices without the worst parts of the daily commute.
That said, some of my most fond memories as a parent is loading up the whole family into the car and heading out on an adventure. There’s something to be said about being able to load all your strollers, all your picnic blankets, all your toys and bikes and skis into a vehicle you own and trust — something that can’t quite be matched by buses, planes, trains and shared self-driving autonomous vehicles. I don’t believe the era of personal vehicles is over in North America, and I hope we find a way to keep it going without the need for internal combustion engines.
