State of Hydrogen - Dec. 2016
Previous blogs on hydrogen:
- MyFC JAQ and NaSi technology
- Is Toyota failing to deliver the Mirai?
- Hydrogen follow-up: Toyota's future
- Why fuel cell cars don't work (part 1)
- Why fuel cell cars don't work (part 2)
- Why fuel cell cars don't work (part 3)
- Why fuel cell cars don't work (part 4)
Nikola's hydrogen fuel cell truck
There has been quite a lot of exciting news around hydrogen, especially this week. Let's just start off with what is probably the best news for hydrogen as a vehicle fuel in many years: The Nikola One was revealed with an actual working prototype. I've been very cynical about hydrogen ever since my days with Forze, but Nikola is a good example of an actual use case for hydrogen as a fuel. The truck itself isn't particularly interesting (it's just another sleeper trailer truck). They boast all kinds of performance advantages, but in practice this won't really matter for long-haul trucking; the few times when you have to navigate steep slopes or accelerate quickly are so few and far between that trucking companies won't be willing to spend lots of extra money on it.
The business model, however, is pretty genius. The truck is only direct leased (which shifts maintenance and thus reliability issues to the manufacturer - a good thing) and it gets a million miles of free fuel. Not a million kilograms - a million miles, i.e. efficiency in real-world driving conditions immediately becomes a priority. Another big part of the business model seems to be a proprietary routing system, where trucking companies can (presumably) pay into the Nikola environment and plan out the most efficient trucking routes; taking into account driver pay, hauling and docking fees as well as proximity to fueling stations at the right time. This sounds great and exactly how logistics should work, but I don't think a proprietary system will work at all. This is already how logistics works in most of the world: lots of incompatible, crappy, unmaintained proprietary systems. Nikola doesn't really have a USP here either; Rotterdams Havenbedrijf and the Dutch flower and vegetable auctions already have these kinds of systems and have had those for decades now. It's just another standard.
Of course, then there's the question of: where do they get their hydrogen from? This goes by the Tesla business model: they're building out 364 fueling stations along major hauling hubs and routes. On the total size of the US, that means you can build out a grid of stations with roughly 400mi in between. At an 800-1200mi specified range, that's just fine for trucks. The hydrogen is sourced from large-scale solar power plants they're planning to build which hydrolyze water into hydrogen and cryogenically store it.
Let's verify Nikola's claims
First off, the financial benefit to the consumer. Their main claim is that the truck essentially pays for itself. Their lease is much more expensive than a normal truck - $5000/mo vs $2200/mo for a comparable semi. However, they say their free fuel compensates for the typical $10k/mo fuel costs of a normal truck. Is that reasonable?
Let's use some averages. On average, semi-trucks drive about 66000mi/yr at 5.8MPG. That's 11380gal of Diesel at, currently, $2.7/gal or $2560/mo in fuel. Wait a second, Nikola is claiming $10k/mo in fuel. Even if we go up to the 98th percentile in miles driven (close to 100kmi/year) and up the Diesel price to $4/gal - the highest it's been in the US in the last decade - that's still not even close to their fuel cost claim. In fact, more recent so-called class 8 semi trucks get much better mileage - between 8 and 10MPG - for long-haul jobs. The lower scores are mostly down to old vehicles and relatively short, low-miles-per-year use cases. Either way, the $10k/mo fuel claim is extremely dubious.
And, well, the same can be said for the lease payment claim. $2200/mo is at the high end of what you'd pay for a detachable sleeper semi, especially a light duty one like this. And with a lease contract, you still need insurance. Unfortunately, insurance scales with the value of the truck, and is about 3-5% of the truck value per year. The Nikola One is tentatively a $400k valued truck (which is in line with its lease price), so that's $20k/yr vs $6250/yr for a $125k truck. I'm only talking about damage premiums, the rest are not dependent on vehicle value. Tabulating this:
Given 66000mi/yr, $2.7/gal diesel, $400k Nikola One vs $125k Freightliner CA113 which does 7.0MPG, the costs per month are:
|Expense item||Nikola One||Diesel truck|
Hey, suddenly the Nikola One is a lot more expensive. But it actually gets quite worse. You see, leasing a new vehicle is rarely a good idea (and lease-to-purchase is especially shitty). This is fairly well-known within the US trucking industry, although some things have slightly changed since the linked forum post from 2005. This is why if you go searching online for leasing contracts on semi trucks, you won't find them. They're all listed with their buying price, not leasing prices, because that's a lot cheaper in the long run. And if you buy trucks new, you're buying them for the long run.
And there's the other issue: nobody makes $10k+ per month trucking, let alone in a sleeper with teaming (and thus higher personnel cost). The Nikola is exceptionally overpriced, and really only a vanity or specialty truck. Sure, there are high-value loads that can pay for such a truck, but it's a tiny fraction of the market. Long story short: I'm calling extreme bullshit on their claim that the Nikola One pays for itself compared to a Diesel truck.
A couple of other figures on their comparison page are also dubious, although closer to reality than the fuel cost malarky. The 500-750mi range figure is mostly moot, and almost all truck types have an optional third fueler that can extend range to 1500+ mi, at the cost of a bit of load capacity. The MPG number seems to be a long-yearly average, not a typical figure for new trucks (which they should compare against). Hell, Daimler's SuperTruck prototype achieves 12.2MPG with a full load. 9.0MPG trucks can be bought right now.
Their CPM figure is total madness as well; even with very short-haul trucking typical CPM is still sub-$0.50. Same with maintenance; long-haul trucks can easily get $0.06/mi average costs on a fleet. It seems like they're cherrypicking the worst-case scenarios here.
Nikola's financial viability
The second thing to look at is their business model proper. They're building very large PV installations near their planned 364 filling stations. How much will that cost? One way to find out is to just follow the energy flows.
The Nikola says it achieves 10-15MPGe - let's assume the worst case for now. Assuming 66000mi/yr at 10MPGe is almost exactly 10mi/kgH2 or 10mi per 33kWh. In total, the truck will require 218MWh of hydrogen. This is cryogenically liquefied hydrogen, a process that has been demonstrated at 75% efficiency, but has a theoretical limit of about 88%. Given that these stations are small-scale installations, let's assume the former. Using state-of-the-art high-pressure electrolysis, you can produce hydrogen at 75% efficiency from water (although only 66% efficiency has been achieved so far). There's a process in between (isentropic expansion) which has about 90% efficiency, as it requires active refrigeration. In all, going from electricity to hydrogen has a 0.75*0.75*0.9 ~ 50% total efficiency. That's pretty generous, but I assume they're really putting some effort into this.
So each truck requires 436MWh of electricity per year to make its hydrogen. At an average of about 1kWh/Wp/yr generation and - by the time they're building it - probably less than $1/Wp total installed cost, each single truck requires about $400 000 in solar panels.
Or, if Nikola uses a PPA (Power Purchasing Agreement) with large solar operators, it'll currently pay $0.04/kWh and in 2020 it'll probably pay between $0.028-$0.036/kWh for solar power. That's about $185 000 to $235 000 in total electricity cost over the 1 million mile guarantee.
Harder to quantify are the other costs. Producing hydrogen and dispensing it is also not cheap. Current hydrogen filling stations cost about $3M per station, with two nozzles each. This is a low estimate for Nikola, as they're storing hydrogen cryogenically instead of at high pressure, which adds significant expense. Not to mention the hydrogen and thermal leakage, which is quite significant. Of course, if 100 000 trucks use just those 400 stations, the per-truck cost is still really low. Nikola is planning to serve other hydrogen vehicles as well, so that adds a bit of cost sharing.
Let's just split the difference and say that Nikola has to spend about $300k over the lifetime of the truck on fuel, which we put at 15 years for now. In that period, assuming an above-average leasing percentage of 85% (at any time, about 17% of the US trucking fleet is inactive) they will make $765k in leasing income. The truck itself is valued at about $400k. Hang on, let's analyze that for a bit.
They've got a 320kWh battery. At the current $170-200/kWh for lithium ion, that's about $60k for the battery. On top of that, there is a fuel cell of unspecified capacity, but given that it burns through 3.3kWh/mi on average at 55MPH, it has to be in the order of at least 200kW. They're almost certainly using a PEM fuel cell (because of weight issues with a SOFC), meaning the absolute cost bottom hovers around $100/kW, although true free-market costs at the moment are about an order of magnitude higher. At least for the first few thousand units, those fuel cells will cost at least $50-100k. Of course, these costs can go down significantly over the course of the next 5-10 years, but it's safe to assume that the primary powertrain is going to cost well in excess of $100k. Adding on a couple tens of thousands of dollars for the rest of the drivetrain, chassis and cabin, it seems very viable to build a few thousand of these trucks for $400k a piece and make a bit of profit on them as well.
I'm leaving out a lot of other costs - probably a significant fraction of a billion dollars in R&D and testing, infrastructure to source and build the trucks, interest, maintenance costs over the years, etc. However, to a first-order approximation, the $5000/mo business model seems to (barely) cover their costs. I still think it's way too much to ask for any kind of sleeper semi, but they're financially viable.
Other hydrogen news
Filling station update
First and foremost, we need some follow-up on filling stations. Back in October 2015 I reported on a whopping total of 12 hydrogen fuel stations being operational in the entire US, omst of them in Los Angeles. 20 were supposed to be built in 2015 (none ever did), and another 20 were supposed to open this year. Did they?
We can now look at the data! The NREL provides this data free of charge, and subsequently I'm offering a JSON-formatted list of hydrogen station information download on my website. I also have a filtering tool on my site. It's very rudimentary right now as I just made it for this blog. The awesome thing of the NREL data is that it provides literally all information - down to numbers you can call to make an appointment for fueling - available about filling stations. So we can filter out stations that don't really count.
Alright, the total number of stations in the database is 88. However, that includes a lot of planned or unavailable stations, so let's filter those out. This reduces the list to 56. Then there's a lot of private stations, mostly on industrial estates and government only filling points. Filter those out and we're left with 31. Arguably, this is the headline number. We went from 12 to 31 public filling stations in 2016. That's awesome! Just one more and we actually hit the planned 20 new stations in 2016!
Not so fast. There are a lot of iffy stations still left. For instance, two stations require you to get a proprietary PIN (these stations will only fill up certain vehicle models). One station is only available for emergency fueling (?!). A bunch of stations have soft openings, which in practice seems to mean that they're only actually dispensing once in a while and you have to call ahead to make sure. Talking about those - let's filter out all stations that require you to call ahead for a fueling appointment. Lastly, maybe a bit controversially, I'm filtering out all the proprietary card key stations, as they require a separate subscription.
What are we now left with? A grand total of 19 actual public filling stations where you can just go and get hydrogen without jumping through hoops. Like they promised the hydrogen future would be - the same as any gas station.
From this data, the most interesting conclusion seems to be that although FirstElement has opened a whole slew of new filling stations this year (13 out of 19 total!), almost all other stations have been closing down over the course of the last 2 years. Three Air Products stations remain, two state-owned stations still operate, UC-I operates a station and Propel has one. That's it.
Toyota hadn't been selling a lot of Mirai vehicles throughout 2015 and 2016 - a total of 350 up to June 2016 to be precise. With a sales goal of 1000 in 2016 and 2000 in 2017, they had to do something. So as of August 2016, the lease price was dropped from $499/mo to $350/mo, an incredibly large price drop, about the same as the 2017 Acura which is about 2/3rds of the price of a Mirai. Clearly, Toyota wants to kickstart the FCEV sales and is willing to make a loss on the early models.
This has had a tremendous effect. In August alone, 350 Mirais were leased, as many in a single month as the total sales and leases up to that point. The rest of the year has seen stronger sales as well, with already 1000 Mirai vehicles on the road before the end of the year. In other words: the sales target has been met, and the total number of fuel cell vehicles in the world has roughly doubled in 2016. That is much better than I'd expected, although it is obviously mostly because of the great lease deal. By the way - if you live in California and are in the market for a luxury sedan: given the free fuel and incredibly low lease price, this is a steal!
Other smaller news itemsAmong the smaller news stories in hydrogen, I'd consider Honda's decision to build a BEV on the previously hydrogen-only Clarity platform the most interesting. Almost no news has come out about this car so far, but it seems like they want to amortize R&D costs of the hydrogen car over both the FCEV and BEV version, the latter of which will undoubtedly sell much better. By the way, the Honda Clarity is supposed to launch before the end of this year for $369/mo - clearly trying to compete with the Mirai. Expect that any day now. The Clarity has an edge over the Mirai - it can supposedly also power your house from the fuel cell.
Some more, kind of unclear news has come out about more h2 filling stations as well. H2ME is planning to build out 20 more hydrogen filling stations in Europe, as well as rolling out about 1200 new FCEVs. Unfortunately, there is no good information on the progress of this project. Same goes for HyFIVE - a smaller initiative. In Germany, a Coradia iLint hydrogen-powered train has been launched as a pilot project for non-electrified railway stretches which are usually operated with Diesel locomotives. Various other news items about new hydrogen fueling stations have come and gone, but so far none have materialized or gone past press release stages. One article about almost 70 new fueling stations in the US has gotten so little follow-up coverage that I can't seem to even find it back after more than an hour of searching. This is fairly normal - if you would have believed 2014's coverage of hydrogen fueling, you'd think there would have been hundreds of stations already.
In non-news, the MyFC JAQ that I discussed before isn't out yet. It was announced in January, by the way. I would have hoped to have one already, just to take it apart.
ConclusionI think 2016 can be viewed as a comparative triumph for hydrogen. This year, so far, has been the only one in history with more than 1000 concurrent hydrogen vehicles on the road even though we came close in the early 2000s. The Nikola announcement is a distant glimmer of hope for hydrogen vehicles in general. Unfortunately, that's pretty much where all hope ends.
Fueling infrastructure build-out has been exceedingly slow - slower than I even expected. Prices aren't improving (just more loss is being incurred) and new models aren't coming out. In fact, most car manufacturers have finally pulled the plug on their fuel cell projects during 2015 and 2016.
This is a graph with reasonably well-researched predictions from 2010. Guess where we are now...
As I've said in all previous blogs: hydrogen tech and infra needs to accelerate like mad to ever become competitive with pure electrics, let alone beat them on a level playing field. 2017 will be a bitch, with multiple long-range BEVs at affordable prices (i.e. half of a Mirai or Clarity) getting launched. The Renault Zoe and Nissan Leaf 40kWh, Chevrolet Bolt and possibly even 48 or 60kWh upgrades near the end of the year. There are already hundreds of (DC) fast charging stations all over the US and Europe, with rollout accelerating quite substantially after the Paris accord. Impressed by the h2mobile effort to build 400 hydrogen fueling stations before 2023? Try close to 5000 fast chargers all over Europe in 2020.
The Nikola might be able to stem the decline and let hydrogen go in the direction it should have been going: big applications, big vehicles. Places where batteries aren't going to be competitive or even possible. But it's just one new company with their proprietary, hyper-expensive, untested system. Logistics doesn't like to gamble with those odds. I'm pessimistic as to their chances on the free market.
I've felt for some time (based on similar calculations) that hydrogen is an unlikely contestant for consumer vehicle applications. It's such an uphill battle to compete with the already-in-place electricity grid. Even though the current grid will prove woefully insufficient when EV takeup increases, at that point so many EVs are already being produced that hydrogen-based vehicles won't be able to keep up with their falling TCO even in the most rose-colored scenarios. Plus, it will probably be more economical to address such grid inadequacies with pockets of decentralized energ storage.
In the industry, however, there seem to be some great opportunities to use fuel cells to capitalize on hydrogen 'waste' streams (e.g., left-over hydrogen from industrial-scale chemical synthesis).
Anyway, thanks again for this informative write-up!
There are plenty of great opportunities for hydrogen. There are differing views on the matter - a lot of them fueled by what I'd call misconceptions about the alternatives. Hydrogen being the ultimate chemical energy density contestant would make for awesome batteries, so a lot of scholars and industry professionals alike are of the opinion that this inevitably leads to the fact that hydrogen will be the 'end point' of mobile energy storage. Therefore, it makes a lot of sense to front-load this development.
Unfortunately, this flies in the face of current tech reality. Sure, batteries have low energy density, but we don't need ultimate density. We need enough and most importantly: at a low price. Batteries are already enough and battery prices are dropping so fast that the affordability issue will quickly be addressed, insofar as it hasn't been already. With pretty much all feasibility and scaling issues already being tackled, there is only explosive growth to happen to battery electric vehicles.
Hydrogen is still grappling with the most basic of issues (e.g. producing it in quantity without causing massive GHG emissions) and has thermodynamics working against it in a big way. Even though the HHV energy density may be 2-3x as much as you can achieve with a theoretical lithium-air battery, you lose all of that advantage in producing, storing and consuming the fuel due to pretty much unbreakable thermodynamic principles. That makes hydrogen no better end point for vehicles than batteries, regardless of time frame. Additional current technological challenges give it an inherent disadvantage which can never be fixed as long as both technologies progress at comparable rates.
But all of this only pertains to mobile applications. Static or large-scale applications have the tables turned. For applications that require relatively low power density but lots of energy storage, hydrogen is great. Storing more hydrogen just means adding more cheap tanks. If you have room, they don't even need to be super high pressure. Likewise, Nikola shows something that just can't be addressed with batteries anyway, but could benefit massively from electrification. That's the direction industry should be going in.
Low-speed diesel engines can run at a pretty high efficiency, let's say 40%. The most efficient natural gas fired power plants can reach 60%, but of course it's impossible to convert that power into hydrogen and back to power without spending more fossil fuels than by just burning diesel. Also, the price of non-fossil fuels will probably not drop under that of fossil fuels because the fossil fuel price will drop when that is about the happen, just because demand & supply, I think. So, will long-distance rural trains run on diesel forever? Or is there still a chance for hydrogen?
Weight is not really an issue for heavy trains, for example, Swedish electric iron ore locomotives need 60 tonne ballast weights to increase their rail-to-wheel friction level to haul similar trains as North-American ones.
[Reactie gewijzigd op zaterdag 10 december 2016 16:04]
Here's a decent overview of electrification and diesel costs on non-even terrain: http://reasonrail.blogspo...analysis-of-railroad.html.
The exact numbers aren't super important (they are not far off for mainland Europe - BE and NL cost about €200k/km/yr, but see much higher utilization), what's important to know is the ballpark figures: electrification costs a couple million per mile to build but after that it's only about $10k/yr/mi to maintain and feed electricity. Diesel costs almost nothing to build, but is an order of magnitude more expensive in fuel, maintenance and externalities.
Hydrogen is much like Diesel, only fuel costs are even higher (about a factor of 5 right now, if you make it from really cheap solar energy, will go down to 2-3x more in the future as tech improves, not likely to get any better). But for a single train service, maybe once an hour during the day, it's still cheaper than electrification. Infrastructure costs are practically nothing compared to electrification. It's really all about the fuel (=solar panels + hydrolyzer).
The biggie, though, is that if you include a fair price of CO2 (e.g. $200/tCO2), Diesel suddenly becomes twice as expensive, i.e. pretty much as expensive as hydrogen would be now from renewable sources, given a stable and mature market. This is competitive, maybe even advantageous because of the true electric nature of the beast (which helps a lot with acceleration and inclines - something diesel-electrics still have trouble with). Not to mention lower weight and more internal space.
So yes, I see a pretty good case for running trains on hydrogen. Much like big trucks and other large applications: hydrogen makes sense when weight and/or flexibility don't really need to exist.