Paper at https://arxiv.org/abs/2406.09624 ("DrivAerNet++: A Large-Scale Multimodal Car Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks")
Its a few hundred gigabyte. You can only login with an institutional account. My research institute is very small, can I borrow someone's login credentials? (email in profile).
Being licensed as it is, it needs to be publicly available. This availability can be on request, but that request can be made by anyone.
The watering down of the legal definitions of Open Source and Creative Commons is a death of a thousand cuts. Every single misrepresentation and false claim needs to be called out and shut down. It's hard to stress how important these concepts are to the current state of the world. Many people say "stop being pedantic" or "the source can be read, therefore it's open source". The reason you currently have a secure server operating system that can use state of the art encryption on affordable commodity hardware is that "Open Source" has a legal definition. Please, let's not break it.
Not trying to be pedantic, but you literally asked for false claims to be called out and shut down.
There is no open source license (that I know of) which requires public availability. If there is a requirement of providing sources, it's always a requirement of providing sources with a derived work. If you are not releasing derived works to anyone you don't need to release any source to anyone either. So called permissive licenses don't even have any requirement on providing sources whatsoever.
That's because license requirements are on _other_ users (who are being licensed to use it) to follow, not the owner (who _owns_ it and doesn't need to follow the license terms the owner requires of others).
Nevertheless, not every license is open source, open source does have a definition, and it's not about whether the owner is following their own license.
Can someone show what the most aerodynamic cars in the dataset look like? Does it specify the drag coefficient? Curious to know how many of the 8000 shapes reach < 2.0, should be at most a handful.
Somewhat related: Can someone tell me why all those new electric cars are so poorly designed in terms of size and weight? Most of them weight over 2t and have a large frontal area (VW ID.4 is the prime example). I just want an economical family vehicle (lightweight, hatchback design) and there aren't that many options. Isn't that something loads of people want or am I the exotic one?
I believe the reason is that batteries to make up the 200 mile range people expect are just very heavy. I think older electric vehicles with ~70mi range would have been substantially lighter. I would wager the car you want was actually the 2017-2019 E-Golf.
Modern highway safety standards also tend to increase the hight of windows which tend to also make cars taller to not kill visibility. Look at the 2019 Mazda 3 AWD redesign, especially the hatchback. It looks squished compared to the previous years because they didn't raise the vehicle height, but all the windows start higher in the car body. That trend is true of ICE vehicles as well though.
E-golf had an updated battery in 2017, that increased the capacity by 50%. The 2014 version was sold with 200km (125mi) range, but the real range was closer to 125km (75mi). The 2017 version was sold with a 300km (185mi) range, but 190km (115mi) real range.
The range of both depend on if you drive in cold weather and if you drive on a highway. I bought and E-golf in 2017 that I drove long distance once on the highway when it was -20C (-4F). I then got around one hour of driving and one hour of charging... Do not recommend!
If you want more stats about the cars, this webpage is great:
Had it for just over a year, but so far happy with the Renault Megane e-Tech[1]. It has a kerb weight of around 1700kg, of which 400kg is the battery[2].
It's smaller than the ID.4 so fits in our small garage, great turning radius which is nice in the city, and physical buttons for the most important stuff like climate and media. It can tow up to 900kg and supports a roof rack.
Here in Norway they haven't been very aggressive with the advertisements, which I guess "helps".
Unlike the Kia Nero EV and Hyundai Kona, two similar competitors, the Megane e-Tech was designed to be 100% BEV from the start, which I think is advantageous for an EV.
I think the flat front is mostly design language to make the car seem more “masculine” or SUVish, and is even more noticeable because of the shorter hood on EVs.
The weight is just a consequence of current battery technology- no way to get that down much without hurting the range.
Look at Aptera motors for an ultralight efficient EV.
Carmakers have been tending towards SUV and crossover designs because they make more money on them. The ID.4 is a crossover, basically an SUV on a car frame rather than a truck frame. I too wish they made more compact cars, but it seems that they would rather produce what is more profitable for them rather than what customers want.
I’m not really disagreeing with your broader point, but consumer demand has been tracking toward trucks and SUVs for quite a while now. I would imagine there is a bit of a “tail wagging the dog” situation there, and it’s also probably relevant that the average age of a new car buyer has been rising as wealth distribution has skewed further upward.
But that’s great news, if there’s truly a gap in the famously difficult automotive market, maybe a scrappy new player can break through with a simple, small sedan—one that can be made cheaply, and sold affordably—and then their supply chain can be forced into a series of buy outs and hostile takeovers, creating the next generation of wealthy donors to political action groups. It’s the ciiiirrcle of liiiife etc..
I think part of the issue is that the people who want simple non-SUVs are the people who mostly buy used cars.
The majority of people buy their cars used, but the car makers make the cars they can sell to the minority who buy new cars - and they have very different tastes. New car buyers like gadgets and they like SUVs.
Crossovers in that zone are a pretty reasonable compromise of cost, handling, safety, and ergonomics. They have essentially replaced station wagons and are mostly built the same way, just a little taller.
I agree that they have replaced station wagons but as a station wagon owner they aren't really even close to interchangeable. I don't know why they replaced them.
Compare my Mazda 6 wagon to the Mazda CX-9 crossover. My wagon is almost 25% more fuel efficient (7L/100km vs 9L/100km) but also has MUCH more space. Minimum cargo space of 506L vs 230L. Maximum cargo space of 1648L vs 810L. It has a shorter wheelbase and smaller turning radius. It does have worse towing 550kg vs 750kg but I almost never see anyone towing anything.
Looking at the current Mazda lineup the replacement for the Mazda6 (29 mpg) is the CX-50 which adds AWD and gets 28 mpg or 38 mpg for the hybrid. I was suprised the Mazda6 would have so much more cargo area but in the spec sheet it lists 22 ft vs 31 for the CX-50. So to me the CX-50 looks like more capacity in about the same physical and carbon footprint, what am I missing here?
In Toyota's lineup the RAV4 is basically a tall Camry, and in Subaru's lineup the Outback kind of just got a bit taller and reclassified as a crossover alternative to the Legacy. Of course these companies all have larger and smaller options, but in general it seems like the adoption of larger vehicles is not because smaller ones aren't available, it's because for various reasons customers want to buy them. In any case worrying about exactly what vehicle mix is selling to me feels like rearranging deck chairs on the Titanic.
I don't disagree, but I just prefer compact cars. If I had to buy a new vehicle and there were no compacts available, I'd go for a crossover, but I won't like it.
I prefer to drive smaller cars but with stroller, kids' car seats, etc there are some physical constraints on getting everything to fit in a reasonable way.
Gasoline has 100x more energy density per kg than Lithium batteries. As you said EVs are 3x more efficient so per kg dofference is 33x. Electic motors are much more compact and lighter than ICEs but not enough to close the 33x gap for the same weight. Thus for a similar range EVs need to be heavier.
My ICE car is 1300kg and it can drive ~1000km with a full tank of 62 liters.
> Gasoline has 100x more energy density per kg than Lithium batteries.
This is true and also misleading since much more of the energy stored in gasoline is unavailable for conversion to work relative to an electrical battery. Additionally, an electrical battery can absorb some of the vehicle's kinetic energy, which a vehicle with only an ICE cannot do.
That table says that 122kwh is only 4 gallons of gasoline. Most ICE cars that compare to an ev with a 122 kWh battery size hold 20 gallons of more.
So still 5x energy density in the ICE vehicle
Or am missing something?
Sorry for unclarity. I wasn't comparing the weight if the total car in EV or ICE scenario. It's more like 2x for a similar range I would think, because EV drivetrain is so much lighter than ICE one.
* Crash testing & crumple zones. Going up against a "half ton" MRAP in a electric velomobile is suicidal.
* Styling & consumer preferences. The SUV reigns supreme. Partly because of CAFE's perverse incentives.
* CAFE isn't directly forcing anyone to optimize EVs, and "watt hours per mile" is not widely demanded by consumers, because EV charging is ~40% the cost of gas per mile and because EVs are still exclusively a mid to upmarket thing. Everybody penny pinching is driving a 2003 Toyota with >100k miles which has appreciated in value over the past five years.
The trend to SUVs also exists in countries where SUVs aren't exempt from stuff similar to CAFE. Consumers overwhelmingly prefer SUVs and large, heavy vehicles. They're the single most popular segment of cars in essentially any market that can afford them.
I wish it weren't so, but the revealed preference of the median car buyer is to simply buy as much car by the ton as they can afford.
The American car industry largely created, popularized, and continues to heavily promote the concept of huge SUVs and impractically huge pickups with high ground clearance, high weight, tall hoods and tall beds. It's almost synonymous with our national identity, in the modern automotive era.
And that domestic market seeking overseas sales is shaped by CAFE, which in combination with manufacturing conditions has almost prohibited them from making the sort of small
pickup trucks favored by the entire world before CAFE, and which remains common in the rest of the world. Sedans, minivans, even the tradesman's current best option in light commercial vans, are barely produced domestically anymore in comparison to the Lifestyle Truck, a thing which we gradually invented sometimes between 1980 and 2010 out of enclosure of small pickups, then the Soccer Mom replacing the minivan with a more vertically inclined Isuzu, then the Army HMMWV, then the F-150 and everything else just inflating in size annually.
CAFE's min-max optimization is why every marketing executive in the sector wants you to adopt an offroading hobby. That seeps out. Cultural export.
It seems hard to differntiate between chicken and the egg here. Do consumers really prefer big vehicles outright? Or do they prefer them based on the marketing of those vehicles as 'safer' or 'capable' or whatever? We know margins are higher on larger cars, so there is an incentive even in non-CAFE areas.
Part of it feels like an arms race played out on the road as well. I finally had to buy a new[er] car after my 2001 Volvo wagon died. Modern cars are massive, heavy, and have shit visibility. They also have surprisingly similar or even smaller interior volumes compared to similar exterior sozed cars of 15+ years ago. But let me tell you, test driving even like a modern RAV4 (not a huge car by modern standards) made me feel like a prince on the road compared to my low ride wagon. Didnt feel like every other car on the road could run me over without noticing either.
Seems like cars will just keep growing til they hit the limit of roads, parking lots, and garages?
Because companies that aren’t Tesla, Lucid, or some Chinese companies jerry rig batteries into existing designs with hardpoints for an IC engine, the tall hood for radiators and pedestrian collision requirements for a big IC engine block, 20000 microcontrollers from 50 different suppliers, etc
weight is somewhat compensated by recuperation and a consequence of necessary battery weight for desired capacity / travel reach.
and while wind resistance depends on shape and "front area", front area actually is the cross section surface, the silhouette if you wish, as seen from back or front.
thus a teardrop like shape has a large flat part at the front but still it has near perfect, lowest resistance.
EVs reduce their front silhouette ("area") with a narrower cabin, low overall height,
cameras as mirrors on vans.
EV Battery packs are very heavy and dense. The ID4 has an aerodynamic Cd score of 0.28 - Mach-E has a Cd of 0.29 and Model Y has a Cd of 0.23, so its in the middle of the pack.
I think the reason electric aircraft are struggling is because of physics. Gasoline has far more gravimetric energy density than lithium ion batteries. And weight matters a lot more for flying than driving. One of the benefits of gasoline is you burn it off while flying, which makes the plane lighter, which increases the range. This is not a negligible effect - pilots are taught to take this into effect when planning flights. It's calculated by all commercial airlines.
Now on top of that add that it costs significantly more and you have what is, by at least the metric above, an objectively worse product for a higher price. Yes, there are other advantages but limited range is a much bigger deal for flying than it is for driving.
There's also safety. Planes can take off at a higher maximum weight than that at which they can safely land. In an emergency shortly after takeoff (when most emergencies occur), fuel may have to be dumped to reduce landing weight. So that's a double-whammy for battery-electric planes; they can't safely take off over their landing weight, which limits their already-short range.
I am pretty sure, that “lightweight” and “family” are not easy to combine unless you’re family of three without the need for a luggage.
Those new electric cars need big batteries just to be somehow usable. Recently consumed 35 kWh for 65 miles ride on model Y with trailer in cold weather. That makes <150 miles total winter range with trailer with fully charged battery.
Most petrol family cars are less than 1500 kg in Europe. Americans are used to overweight cars and that's spreading to here too. Though it's often a problem since most city parking spots are smaller than huge cars like Tesla model Y.
Based on your experience? My parents use their Golf 7 to drive down south to Italy from Germany every other year. Two adults, two teenagers, and one small dog fit in better than into my dad's late 2000s Mercedes S class
Half of Germany drove to Italy with their VW Beetles back then. It’s a traditional thing. However there are some neighborhoods where with second child family buys their Mercedes V class or VW T6. I prefer those over the ones having Ford Fiesta as a family car.
> Can someone tell me why all those new electric cars are so poorly designed in terms of size and weight? Most of them weight over 2t and have a large frontal area (VW ID.4 is the prime example). I just want an economical family vehicle (lightweight, hatchback design) and there aren't that many options.
If you don't need anything right now, don't feel any pressure to go out and buy something new. If you do, get whatever works for you, even if that is an ICE vehicle. Don't feel bad; you're doing what is best for your family.
But know that the "future" in EVs is likely closer than you think. Look at the Canoo vehicles - I'm not saying buy one, just look at it and think about the design. Look at the VW ID.Buzz. Again, don't buy one (it's an overpriced usability mess), just look at it and think about the design. You know what is most interesting about it? The length is a foot to a foot and a half less than the existing minivans on the market, but it has more interior room. That is interesting, isn't it?
Well, the Ford CEO said this in the Q2 2024 investor call [1]:
> ...Now when people hear about affordability and they think about small and unaffordable, I'd like to address that now. We are designing a super-efficient platform, leveraging innovation across our product development, supply chain, and manufacturing teams. With no engine or drivetrain, a smaller vehicle can have a much roomier package, actually the interior package of a class above with a small silhouette. That's a big advantage for customers versus ICE. And we're focusing on very differentiated vehicles priced under $40,000 or even $30,000.
And then in the Q3 2024 investor call [2]:
> we're deep into the design and engineering of our next-generation vehicles. Boy! are we excited about these coming out in the next few years. In 40 years in the industry, I've seen a lot of game-changer products, but the mid-sized electric pickup designed by our California team has got to be one of the most exciting. It's incredible package and consumer technology for a segment we know well. It matches the cost structure of any Chinese auto manufacturer building in Mexico in the future. How do we know that? Because 60% of the [BOM] (ph) has already been quoted.
So, this pickup is the first one to come out, and that's not what you are looking for. But the EVs that come after it may be exactly what you are looking for. And of course Ford is not going to be alone. All the car companies are going to be working from this point of view. So, can you wait 4-5 years for the EV that actually works for what you need? Perhaps your next ICE car is your last ICE car?
Using a computational fluids model like OpenFOAM is overkill for reasonable estimates of aerodynamic performance and stability, provided you stick to normal-ish designs. The term you want to Google is "vortex lattice" and/or "potential flow" solvers, of which there are many open source softwares available specifically geared towards typical aircraft configurations, with simple, easy to use interfaces.
Some good ones to look at to get started with nice guis
- VSP Aero
- XFOIL (for 2D airfoil analysis only, also generates inputs for 3D VLM solvers)
- AVL
And some others if you want perhaps a little higher fidelity, and don't mind text/code interfaces:
-FreeWake
- Datcom (not technically a potential flow models, but a database)
To get the most value out of these softwares, you will need some background in aeronautics. You need to have an understanding of what a potential flow models can and cannot accurately model. It also helps to have knowledge about what forms aircraft stability/control and performance data is typically communicated in, as these softwares will use that terminology.
The 2D model alone maybe be sufficient for OPs needs, but I'd advocate for going at least a little further with some form of 3D model. The wing tip effects are significant enough to warrant some form of "next step" after the initial 2D analysis. Although simple algebraic relationships are likely good enough to satisfy this end.
However, the 3D tools above give you a little bit more, that I feel warrant their consideration by the op. These models can yield performance of wing-tail-body configurations that include wake effects. They also can give stability analysis in both longitudinal/lateral axes. The stability effects being especially important if OP wants to design a flying-wing, which are a tad more challenging to get right than just slapping a "good-enough" sized tail on a classic config plane.
For this type of problem, unless you are trying to set range or climb rate records, understanding lift and moment (and control derivatives) is more important than accurately modeling drag. Panel methods shine at this, and are both easier to use and much much faster than CFD — enough that wrapping them in optimizers is trivial. I personally prefer AVL, but there are other open source alternatives as well.
OpenFOAM I think is the go to for open source CFD, although I’ve never tried it myself. There’s also XFOIL which, since you’re talking about a flying wing, might be enough for your use case.
Even for a finicky flying wing, I think if you keep an eye on the way the pitching moment shifts with pitch, I don’t see why picking an airfoil shape for the frame and hand calcs couldn’t get you to a design that flies reasonably.
So to clarify, this does not include models of actual cars. This is a set of models generated by randomising some parameters. Basically it's a set or random car shapes based of one template.
Isn't this more efficient? Cheaper parts, everyone benefits from improvements to it.
I wouldn't mind if there were only a handful of designs around with some tweaks depending on performance/comfort.
Also, nowadays many different models share parts, like lights or rearview mirrors, showing there's a need for efficiency and scale bigger than the made-up need for fashionable cars.
It is. I would prefer the advantages of a “cheap as possible but robust” basic car design. I want a car that is air gapped and not phoning home, that is electric (or good hybrid), and gets me from A to B. I want it to be reasonably comfortable and inexpensive to replace the parts, which comes about via mass production and minimized labor costs. Unfortunately almost all of those go against what car manufacturers want to do. That’s why I drive a 2010 corolla for now and until it blows up.
Good design is emotional. It adds spark to most of our everyday sightseeing, inspires and is too often overlooked- at least in recent times. Regarding the recent trends among car design... We are in a dark age.
I agree that many parts can be interchangeable but modularity can be done in a way that doesn't result in every car looking like the blobs we see about today. I wouldn't mind if there were only a handful of designs around with slight variations in trim and specs -so long as I like those designs of course! I know that is subjective so lets keep things interesting and focus on making things that get the job done but also in varying styles. Both can be achieved.
The need for fashion is made up, yes- but so is everything else. There's a balance between overly individualistic cars and the most cost efficient, reliable and uninspiring 4 door. Lately we have been leaning too far into the latter.
Aaaaand, that's why they aren't modular. Take a look at something with obvious common parts like a side mirror. They are all intentionally designed adding special shapes to make them as unique as possible to the car. Never a simple square or rectangle.
I don't buy 8GB VRAM GPUs of different form factors, but GPUs as powerful/efficient/cheap as I can get with a standardised form factor.
Business don't want standards because they care about selling non-practical bits of the car for a price, like how good it can make you feel because you think it's cool, even if it's dumb (dummy exhausts, fake carbon, LED logos and fashion oriented design)
Just give me something that looks like a Buick Grand National or a '92 Camaro..... I'll pay the extra money for gas. Why can't they be creative and reduce the efficiency just a tad.
Pointing out a material misrepresentation is hardly a nitpick. I'm quite surprised to see an MIT affiliated publication making such an error.
That said since it isn't source code I think the end result is probably a bit different. Unless I misunderstand something you could still use the data to further a commercial venture. You just can't directly commercialize the dataset itself. Conceptually similar to using a GPL'd software tool as an end user while performing work on a proprietary project.
I don't know how you can use this for a comercial task. For example CC explains you can't use a CC-NC song in a comercial video. I expect that if you use the data to design a car and sell it the authors will have a similar interpretation.
Wouldn't that be because you directly commercialize the video and the song is part of the video? Therefore you are directly commercializing the song.
It doesn't seem like there would be any legal barrier to feeding the dataset into an algorithm and making use of the result (depending on the implementation details of the algorithm I suppose).
While I don't always love 'denominalization' (if I that's the right word), the verb 'speed' is a common one, e.g., 'speed up'.
Every word we use was created at some point, often in the same way. There is nothing magical about the people and the words they created in 1824 that makes them better than us and words we create in 2024, except perhaps that the 1824 words we still use have stood the test of time.
Also, as far as I can tell, the verb form is as old as the noun form.
I read moby dick for the first time last summer. That guy verbs nouns, adjects verbs and nouns adjects like he's building language from first principles. It takes a bit getting used to, but eventually it both speeds up and broadens language. Why use different roots for each category? And with each root connoting slightly different, why not use that and pick the exact root that means best?
Datafiles available at: https://dataverse.harvard.edu/dataverse/DrivAerNet
Issue tracking at: https://github.com/Mohamedelrefaie/DrivAerNet
Paper at https://arxiv.org/abs/2406.09624 ("DrivAerNet++: A Large-Scale Multimodal Car Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks")
Its a few hundred gigabyte. You can only login with an institutional account. My research institute is very small, can I borrow someone's login credentials? (email in profile).
It is not meant for non credentialed individuals?
Being licensed as it is, it needs to be publicly available. This availability can be on request, but that request can be made by anyone.
The watering down of the legal definitions of Open Source and Creative Commons is a death of a thousand cuts. Every single misrepresentation and false claim needs to be called out and shut down. It's hard to stress how important these concepts are to the current state of the world. Many people say "stop being pedantic" or "the source can be read, therefore it's open source". The reason you currently have a secure server operating system that can use state of the art encryption on affordable commodity hardware is that "Open Source" has a legal definition. Please, let's not break it.
Not trying to be pedantic, but you literally asked for false claims to be called out and shut down.
There is no open source license (that I know of) which requires public availability. If there is a requirement of providing sources, it's always a requirement of providing sources with a derived work. If you are not releasing derived works to anyone you don't need to release any source to anyone either. So called permissive licenses don't even have any requirement on providing sources whatsoever.
That's because license requirements are on _other_ users (who are being licensed to use it) to follow, not the owner (who _owns_ it and doesn't need to follow the license terms the owner requires of others).
Nevertheless, not every license is open source, open source does have a definition, and it's not about whether the owner is following their own license.
From the repo
> Please note the (CC BY-NC 4.0) license terms, as outlined in the License section.
In spite of the title and the article says "open souce" many times, it's free as in "source abailable".
[dead]
Can someone show what the most aerodynamic cars in the dataset look like? Does it specify the drag coefficient? Curious to know how many of the 8000 shapes reach < 2.0, should be at most a handful.
> reach < 2.0
I meant < 0.20 of course, such as the 0.19 achieved by the ill-fated https://en.wikipedia.org/wiki/General_Motors_EV1
This is also available at https://caemldatasets.org/ with no access restrictions.
Somewhat related: Can someone tell me why all those new electric cars are so poorly designed in terms of size and weight? Most of them weight over 2t and have a large frontal area (VW ID.4 is the prime example). I just want an economical family vehicle (lightweight, hatchback design) and there aren't that many options. Isn't that something loads of people want or am I the exotic one?
I believe the reason is that batteries to make up the 200 mile range people expect are just very heavy. I think older electric vehicles with ~70mi range would have been substantially lighter. I would wager the car you want was actually the 2017-2019 E-Golf.
Modern highway safety standards also tend to increase the hight of windows which tend to also make cars taller to not kill visibility. Look at the 2019 Mazda 3 AWD redesign, especially the hatchback. It looks squished compared to the previous years because they didn't raise the vehicle height, but all the windows start higher in the car body. That trend is true of ICE vehicles as well though.
E-golf had an updated battery in 2017, that increased the capacity by 50%. The 2014 version was sold with 200km (125mi) range, but the real range was closer to 125km (75mi). The 2017 version was sold with a 300km (185mi) range, but 190km (115mi) real range.
The range of both depend on if you drive in cold weather and if you drive on a highway. I bought and E-golf in 2017 that I drove long distance once on the highway when it was -20C (-4F). I then got around one hour of driving and one hour of charging... Do not recommend!
If you want more stats about the cars, this webpage is great:
2014 version: https://ev-database.org/car/1040/Volkswagen-e-Golf
2017 version: https://ev-database.org/car/1087/Volkswagen-e-Golf
Had it for just over a year, but so far happy with the Renault Megane e-Tech[1]. It has a kerb weight of around 1700kg, of which 400kg is the battery[2].
It's smaller than the ID.4 so fits in our small garage, great turning radius which is nice in the city, and physical buttons for the most important stuff like climate and media. It can tow up to 900kg and supports a roof rack.
YMMV and such.
[1]: https://ev-database.org/car/1521/Renault-Megane-E-Tech-EV60-...
[2]: https://cdn.group.renault.com/ren/gb/transversal-assets/broc...
Honestly sounds like it's near the sweet spot. Seems to be available in Australia as well. I'm not sure why I haven't seen any on the road!
Here in Norway they haven't been very aggressive with the advertisements, which I guess "helps".
Unlike the Kia Nero EV and Hyundai Kona, two similar competitors, the Megane e-Tech was designed to be 100% BEV from the start, which I think is advantageous for an EV.
I'd say worth a look.
I think the flat front is mostly design language to make the car seem more “masculine” or SUVish, and is even more noticeable because of the shorter hood on EVs.
The weight is just a consequence of current battery technology- no way to get that down much without hurting the range.
Look at Aptera motors for an ultralight efficient EV.
Carmakers have been tending towards SUV and crossover designs because they make more money on them. The ID.4 is a crossover, basically an SUV on a car frame rather than a truck frame. I too wish they made more compact cars, but it seems that they would rather produce what is more profitable for them rather than what customers want.
I’m not really disagreeing with your broader point, but consumer demand has been tracking toward trucks and SUVs for quite a while now. I would imagine there is a bit of a “tail wagging the dog” situation there, and it’s also probably relevant that the average age of a new car buyer has been rising as wealth distribution has skewed further upward.
But that’s great news, if there’s truly a gap in the famously difficult automotive market, maybe a scrappy new player can break through with a simple, small sedan—one that can be made cheaply, and sold affordably—and then their supply chain can be forced into a series of buy outs and hostile takeovers, creating the next generation of wealthy donors to political action groups. It’s the ciiiirrcle of liiiife etc..
I think part of the issue is that the people who want simple non-SUVs are the people who mostly buy used cars.
The majority of people buy their cars used, but the car makers make the cars they can sell to the minority who buy new cars - and they have very different tastes. New car buyers like gadgets and they like SUVs.
Crossovers in that zone are a pretty reasonable compromise of cost, handling, safety, and ergonomics. They have essentially replaced station wagons and are mostly built the same way, just a little taller.
I agree that they have replaced station wagons but as a station wagon owner they aren't really even close to interchangeable. I don't know why they replaced them.
Compare my Mazda 6 wagon to the Mazda CX-9 crossover. My wagon is almost 25% more fuel efficient (7L/100km vs 9L/100km) but also has MUCH more space. Minimum cargo space of 506L vs 230L. Maximum cargo space of 1648L vs 810L. It has a shorter wheelbase and smaller turning radius. It does have worse towing 550kg vs 750kg but I almost never see anyone towing anything.
Looking at the current Mazda lineup the replacement for the Mazda6 (29 mpg) is the CX-50 which adds AWD and gets 28 mpg or 38 mpg for the hybrid. I was suprised the Mazda6 would have so much more cargo area but in the spec sheet it lists 22 ft vs 31 for the CX-50. So to me the CX-50 looks like more capacity in about the same physical and carbon footprint, what am I missing here?
In Toyota's lineup the RAV4 is basically a tall Camry, and in Subaru's lineup the Outback kind of just got a bit taller and reclassified as a crossover alternative to the Legacy. Of course these companies all have larger and smaller options, but in general it seems like the adoption of larger vehicles is not because smaller ones aren't available, it's because for various reasons customers want to buy them. In any case worrying about exactly what vehicle mix is selling to me feels like rearranging deck chairs on the Titanic.
I don't disagree, but I just prefer compact cars. If I had to buy a new vehicle and there were no compacts available, I'd go for a crossover, but I won't like it.
I prefer to drive smaller cars but with stroller, kids' car seats, etc there are some physical constraints on getting everything to fit in a reasonable way.
> I just want an economical family vehicle
Then buy an ICE vehicle. It's a more mature, cost-effective, and power/energy-dense technology
Pretty sure the Model Y is more cost effective than any ICEV now. There are still valid reasons to prefer an ICE but it’s more expensive to do so.
But the efficiency factor of an EV is 60% as opposed to 20% for an ICE, which means you can get a similar range with much lower energy density.
Gasoline has 100x more energy density per kg than Lithium batteries. As you said EVs are 3x more efficient so per kg dofference is 33x. Electic motors are much more compact and lighter than ICEs but not enough to close the 33x gap for the same weight. Thus for a similar range EVs need to be heavier.
My ICE car is 1300kg and it can drive ~1000km with a full tank of 62 liters.
> Gasoline has 100x more energy density per kg than Lithium batteries.
This is true and also misleading since much more of the energy stored in gasoline is unavailable for conversion to work relative to an electrical battery. Additionally, an electrical battery can absorb some of the vehicle's kinetic energy, which a vehicle with only an ICE cannot do.
https://en.wikipedia.org/wiki/Gasoline_gallon_equivalent
That table says that 122kwh is only 4 gallons of gasoline. Most ICE cars that compare to an ev with a 122 kWh battery size hold 20 gallons of more. So still 5x energy density in the ICE vehicle Or am missing something?
Yes, 5x is more favorable than 33x as quoted by the GP.
Sorry for unclarity. I wasn't comparing the weight if the total car in EV or ICE scenario. It's more like 2x for a similar range I would think, because EV drivetrain is so much lighter than ICE one.
That’s not “unclarity.” That’s “moving the goalposts.”
Sure, but this is pretty pedantic in practice - 5x is still a huge number. There is no practical difference between 5x and 30x and 1000x.
Except for an order of magnitude, sure. But you're not getting 5x range in ICE cars.
5x is not 100x. Or am I missing something?
I literally was asking a question, no need to be an ass.
Do your own math, no need to be lazy.
It's not the math dude it's ensuring I'm understanding what the table is trying to say.
Maybe stop being an ass and try to be useful?
No, I am not your helper, especially when you call me names.
That sounds like a VW Passat B5...
…at the expense of door handles and good looking tires.
People are not buying as many compact cars.
Electric cars need large, massive batteries to have a useful range.
It's as simple as those two things.
* Crash testing & crumple zones. Going up against a "half ton" MRAP in a electric velomobile is suicidal.
* Styling & consumer preferences. The SUV reigns supreme. Partly because of CAFE's perverse incentives.
* CAFE isn't directly forcing anyone to optimize EVs, and "watt hours per mile" is not widely demanded by consumers, because EV charging is ~40% the cost of gas per mile and because EVs are still exclusively a mid to upmarket thing. Everybody penny pinching is driving a 2003 Toyota with >100k miles which has appreciated in value over the past five years.
The trend to SUVs also exists in countries where SUVs aren't exempt from stuff similar to CAFE. Consumers overwhelmingly prefer SUVs and large, heavy vehicles. They're the single most popular segment of cars in essentially any market that can afford them.
I wish it weren't so, but the revealed preference of the median car buyer is to simply buy as much car by the ton as they can afford.
The American car industry largely created, popularized, and continues to heavily promote the concept of huge SUVs and impractically huge pickups with high ground clearance, high weight, tall hoods and tall beds. It's almost synonymous with our national identity, in the modern automotive era.
And that domestic market seeking overseas sales is shaped by CAFE, which in combination with manufacturing conditions has almost prohibited them from making the sort of small pickup trucks favored by the entire world before CAFE, and which remains common in the rest of the world. Sedans, minivans, even the tradesman's current best option in light commercial vans, are barely produced domestically anymore in comparison to the Lifestyle Truck, a thing which we gradually invented sometimes between 1980 and 2010 out of enclosure of small pickups, then the Soccer Mom replacing the minivan with a more vertically inclined Isuzu, then the Army HMMWV, then the F-150 and everything else just inflating in size annually.
CAFE's min-max optimization is why every marketing executive in the sector wants you to adopt an offroading hobby. That seeps out. Cultural export.
It seems hard to differntiate between chicken and the egg here. Do consumers really prefer big vehicles outright? Or do they prefer them based on the marketing of those vehicles as 'safer' or 'capable' or whatever? We know margins are higher on larger cars, so there is an incentive even in non-CAFE areas.
Part of it feels like an arms race played out on the road as well. I finally had to buy a new[er] car after my 2001 Volvo wagon died. Modern cars are massive, heavy, and have shit visibility. They also have surprisingly similar or even smaller interior volumes compared to similar exterior sozed cars of 15+ years ago. But let me tell you, test driving even like a modern RAV4 (not a huge car by modern standards) made me feel like a prince on the road compared to my low ride wagon. Didnt feel like every other car on the road could run me over without noticing either.
Seems like cars will just keep growing til they hit the limit of roads, parking lots, and garages?
Because companies that aren’t Tesla, Lucid, or some Chinese companies jerry rig batteries into existing designs with hardpoints for an IC engine, the tall hood for radiators and pedestrian collision requirements for a big IC engine block, 20000 microcontrollers from 50 different suppliers, etc
in a nutshell...
weight is somewhat compensated by recuperation and a consequence of necessary battery weight for desired capacity / travel reach.
and while wind resistance depends on shape and "front area", front area actually is the cross section surface, the silhouette if you wish, as seen from back or front.
thus a teardrop like shape has a large flat part at the front but still it has near perfect, lowest resistance.
EVs reduce their front silhouette ("area") with a narrower cabin, low overall height, cameras as mirrors on vans.
thats how you reduce "front area" silhouette.
the shape of these is aerodynamic AF
https://web.archive.org/web/20240000000000*/http://www.yanfa...
Nissan Leaf?
So are 'most' petrol cars, people want space for a family plus fair amount of luggage and end up at Model 3 size minimum to cover that spec.
If you want a small EV what about;
Dacia Spring Electric
Leapmotor T03
Citroen e-C3
Vauxhall Frontera
Vauxhall Corsa
The list goes on. Renault 5 E-Tech
Honda-e
VW id.2 all
The latter is not on the market yet, but planned to be in 2025. Looking forward to that materializing.
The former is also discontinued.
EV Battery packs are very heavy and dense. The ID4 has an aerodynamic Cd score of 0.28 - Mach-E has a Cd of 0.29 and Model Y has a Cd of 0.23, so its in the middle of the pack.
Now you're seeing why electric aircraft are struggling.
I think the reason electric aircraft are struggling is because of physics. Gasoline has far more gravimetric energy density than lithium ion batteries. And weight matters a lot more for flying than driving. One of the benefits of gasoline is you burn it off while flying, which makes the plane lighter, which increases the range. This is not a negligible effect - pilots are taught to take this into effect when planning flights. It's calculated by all commercial airlines.
Now on top of that add that it costs significantly more and you have what is, by at least the metric above, an objectively worse product for a higher price. Yes, there are other advantages but limited range is a much bigger deal for flying than it is for driving.
There's also safety. Planes can take off at a higher maximum weight than that at which they can safely land. In an emergency shortly after takeoff (when most emergencies occur), fuel may have to be dumped to reduce landing weight. So that's a double-whammy for battery-electric planes; they can't safely take off over their landing weight, which limits their already-short range.
I am pretty sure, that “lightweight” and “family” are not easy to combine unless you’re family of three without the need for a luggage.
Those new electric cars need big batteries just to be somehow usable. Recently consumed 35 kWh for 65 miles ride on model Y with trailer in cold weather. That makes <150 miles total winter range with trailer with fully charged battery.
Most petrol family cars are less than 1500 kg in Europe. Americans are used to overweight cars and that's spreading to here too. Though it's often a problem since most city parking spots are smaller than huge cars like Tesla model Y.
People don’t use a Golf VII as a family car, they use SUVs. And most of those are closer to 2 tons than 1.5.
Based on your experience? My parents use their Golf 7 to drive down south to Italy from Germany every other year. Two adults, two teenagers, and one small dog fit in better than into my dad's late 2000s Mercedes S class
Half of Germany drove to Italy with their VW Beetles back then. It’s a traditional thing. However there are some neighborhoods where with second child family buys their Mercedes V class or VW T6. I prefer those over the ones having Ford Fiesta as a family car.
Plenty use golf sized cars as family cars. Fairly normal in the EU and UK
My family (of 4) car is a Kia Rio :). We live in an EU city and do 300km trips every other weekend. Frequently it is a tight fit, but one can manage.
Golf is THE family car. SUVs are not as common and often driven by boomers who don't live with their children anymore.
> Can someone tell me why all those new electric cars are so poorly designed in terms of size and weight? Most of them weight over 2t and have a large frontal area (VW ID.4 is the prime example). I just want an economical family vehicle (lightweight, hatchback design) and there aren't that many options.
If you don't need anything right now, don't feel any pressure to go out and buy something new. If you do, get whatever works for you, even if that is an ICE vehicle. Don't feel bad; you're doing what is best for your family.
But know that the "future" in EVs is likely closer than you think. Look at the Canoo vehicles - I'm not saying buy one, just look at it and think about the design. Look at the VW ID.Buzz. Again, don't buy one (it's an overpriced usability mess), just look at it and think about the design. You know what is most interesting about it? The length is a foot to a foot and a half less than the existing minivans on the market, but it has more interior room. That is interesting, isn't it?
Well, the Ford CEO said this in the Q2 2024 investor call [1]:
> ...Now when people hear about affordability and they think about small and unaffordable, I'd like to address that now. We are designing a super-efficient platform, leveraging innovation across our product development, supply chain, and manufacturing teams. With no engine or drivetrain, a smaller vehicle can have a much roomier package, actually the interior package of a class above with a small silhouette. That's a big advantage for customers versus ICE. And we're focusing on very differentiated vehicles priced under $40,000 or even $30,000.
And then in the Q3 2024 investor call [2]:
> we're deep into the design and engineering of our next-generation vehicles. Boy! are we excited about these coming out in the next few years. In 40 years in the industry, I've seen a lot of game-changer products, but the mid-sized electric pickup designed by our California team has got to be one of the most exciting. It's incredible package and consumer technology for a segment we know well. It matches the cost structure of any Chinese auto manufacturer building in Mexico in the future. How do we know that? Because 60% of the [BOM] (ph) has already been quoted.
So, this pickup is the first one to come out, and that's not what you are looking for. But the EVs that come after it may be exactly what you are looking for. And of course Ford is not going to be alone. All the car companies are going to be working from this point of view. So, can you wait 4-5 years for the EV that actually works for what you need? Perhaps your next ICE car is your last ICE car?
[1] https://www.fool.com/earnings/call-transcripts/2024/07/24/fo...
[2] https://seekingalpha.com/article/4730242-ford-motor-company-...
I recently wanted to design a flying wing/rc plane. Is there any FOSS to get reasonable results for aerodynamics?
Using a computational fluids model like OpenFOAM is overkill for reasonable estimates of aerodynamic performance and stability, provided you stick to normal-ish designs. The term you want to Google is "vortex lattice" and/or "potential flow" solvers, of which there are many open source softwares available specifically geared towards typical aircraft configurations, with simple, easy to use interfaces.
Some good ones to look at to get started with nice guis
- VSP Aero
- XFOIL (for 2D airfoil analysis only, also generates inputs for 3D VLM solvers)
- AVL
And some others if you want perhaps a little higher fidelity, and don't mind text/code interfaces:
-FreeWake
- Datcom (not technically a potential flow models, but a database)
To get the most value out of these softwares, you will need some background in aeronautics. You need to have an understanding of what a potential flow models can and cannot accurately model. It also helps to have knowledge about what forms aircraft stability/control and performance data is typically communicated in, as these softwares will use that terminology.
I used xfoil in my aerodynamics masters degree, it absolutely good enough for hobby applications
The 2D model alone maybe be sufficient for OPs needs, but I'd advocate for going at least a little further with some form of 3D model. The wing tip effects are significant enough to warrant some form of "next step" after the initial 2D analysis. Although simple algebraic relationships are likely good enough to satisfy this end.
However, the 3D tools above give you a little bit more, that I feel warrant their consideration by the op. These models can yield performance of wing-tail-body configurations that include wake effects. They also can give stability analysis in both longitudinal/lateral axes. The stability effects being especially important if OP wants to design a flying-wing, which are a tad more challenging to get right than just slapping a "good-enough" sized tail on a classic config plane.
For this type of problem, unless you are trying to set range or climb rate records, understanding lift and moment (and control derivatives) is more important than accurately modeling drag. Panel methods shine at this, and are both easier to use and much much faster than CFD — enough that wrapping them in optimizers is trivial. I personally prefer AVL, but there are other open source alternatives as well.
OpenFOAM I think is the go to for open source CFD, although I’ve never tried it myself. There’s also XFOIL which, since you’re talking about a flying wing, might be enough for your use case.
Even for a finicky flying wing, I think if you keep an eye on the way the pitching moment shifts with pitch, I don’t see why picking an airfoil shape for the frame and hand calcs couldn’t get you to a design that flies reasonably.
The usual way is to use a wing with known aerodynamics. Choose one on airfoiltools.com
This project might be exactly what you're looking for:
https://github.com/peterdsharpe/AeroSandbox
This guy seems to have a good handle on diy aero. Both air and water.
https://youtube.com/@rctestflight
Check out: https://ecalc.ch/wingdesigner.htm
Xflr5? OpenFOAM?
I've used neither but I hear others in the gliding and model airplane scene talk about those.
Can't wait to see this in GTA6 mod.
So to clarify, this does not include models of actual cars. This is a set of models generated by randomising some parameters. Basically it's a set or random car shapes based of one template.
For what use is this?
Very nice.
O great. More cars that look identical.
Isn't this more efficient? Cheaper parts, everyone benefits from improvements to it.
I wouldn't mind if there were only a handful of designs around with some tweaks depending on performance/comfort. Also, nowadays many different models share parts, like lights or rearview mirrors, showing there's a need for efficiency and scale bigger than the made-up need for fashionable cars.
It is. I would prefer the advantages of a “cheap as possible but robust” basic car design. I want a car that is air gapped and not phoning home, that is electric (or good hybrid), and gets me from A to B. I want it to be reasonably comfortable and inexpensive to replace the parts, which comes about via mass production and minimized labor costs. Unfortunately almost all of those go against what car manufacturers want to do. That’s why I drive a 2010 corolla for now and until it blows up.
Good design is emotional. It adds spark to most of our everyday sightseeing, inspires and is too often overlooked- at least in recent times. Regarding the recent trends among car design... We are in a dark age.
I agree that many parts can be interchangeable but modularity can be done in a way that doesn't result in every car looking like the blobs we see about today. I wouldn't mind if there were only a handful of designs around with slight variations in trim and specs -so long as I like those designs of course! I know that is subjective so lets keep things interesting and focus on making things that get the job done but also in varying styles. Both can be achieved.
The need for fashion is made up, yes- but so is everything else. There's a balance between overly individualistic cars and the most cost efficient, reliable and uninspiring 4 door. Lately we have been leaning too far into the latter.
I’ve always been shocked there are not way more common parts between vehicles and brands.
Alternators, starters, ac compressors and so much more could just be small, medium, large, extra large.
The price would drop dramatically.
> The price would drop dramatically.
Aaaaand, that's why they aren't modular. Take a look at something with obvious common parts like a side mirror. They are all intentionally designed adding special shapes to make them as unique as possible to the car. Never a simple square or rectangle.
No one wants that. Neither consumer nor business.
Think about every GPU having 8GB VRAM. Nobody wants that.
I don't buy 8GB VRAM GPUs of different form factors, but GPUs as powerful/efficient/cheap as I can get with a standardised form factor.
Business don't want standards because they care about selling non-practical bits of the car for a price, like how good it can make you feel because you think it's cool, even if it's dumb (dummy exhausts, fake carbon, LED logos and fashion oriented design)
Just give me something that looks like a Buick Grand National or a '92 Camaro..... I'll pay the extra money for gas. Why can't they be creative and reduce the efficiency just a tad.
That was actually the design goal of Saturn - Ferrari style with Chevrolet price tag.
A few cars have got it right after Saturn. I would say the Hyundai Veloster and Tesla Model 3 top that list.
The Saturn Sky and Pontiac Solstice were top notch design concepts.
Aerodynamics is aerodynamics, but we should embrace wild paint jobs.
in shift to electric we need to hit reset on design and just build the best car not one that is an iteration of design from our older gas cars
I believe that l the Tesla Models S and 3, and the Hyundai Ionic 5 and 6, are the only vehicles designed from the ground up as electric.
It's funny, this is the second comment that I mention Tesla and Hyundai together. Two very different companies - entertaining to find this overlap.
> Creative Commons Attribution-NonCommercial
Non-commercial licenses are not open source.
The idea is you look at the source and if you want to make money from it you negotiate a license that allows you to do that.
I wonder if you might say something more substantial? This seems like a low effort nitpick as written.
Pointing out a material misrepresentation is hardly a nitpick. I'm quite surprised to see an MIT affiliated publication making such an error.
That said since it isn't source code I think the end result is probably a bit different. Unless I misunderstand something you could still use the data to further a commercial venture. You just can't directly commercialize the dataset itself. Conceptually similar to using a GPL'd software tool as an end user while performing work on a proprietary project.
I don't know how you can use this for a comercial task. For example CC explains you can't use a CC-NC song in a comercial video. I expect that if you use the data to design a car and sell it the authors will have a similar interpretation.
Wouldn't that be because you directly commercialize the video and the song is part of the video? Therefore you are directly commercializing the song.
It doesn't seem like there would be any legal barrier to feeding the dataset into an algorithm and making use of the result (depending on the implementation details of the algorithm I suppose).
Fails at criteria 6.
[flagged]
"Please don't complain about tangential annoyances [...] They're too common to be interesting."
https://news.ycombinator.com/newsguidelines.html
Sorry dang, sorry HN.
[dead]
While I don't always love 'denominalization' (if I that's the right word), the verb 'speed' is a common one, e.g., 'speed up'.
Every word we use was created at some point, often in the same way. There is nothing magical about the people and the words they created in 1824 that makes them better than us and words we create in 2024, except perhaps that the 1824 words we still use have stood the test of time.
Also, as far as I can tell, the verb form is as old as the noun form.
(CW: pedantry)
Well actually, “speed up” is a phrasal verb that is semantically distinct from “speed”.
https://en.m.wikipedia.org/wiki/English_phrasal_verbs
But that all’s irrelevant because this article uses verb definition 5 from Wiktionary for “speed”:
> (transitive) To increase the rate at which something occurs.
https://en.m.wikipedia.org/wiki/English_phrasal_verbs
“He was speeding.” Love English!
I read moby dick for the first time last summer. That guy verbs nouns, adjects verbs and nouns adjects like he's building language from first principles. It takes a bit getting used to, but eventually it both speeds up and broadens language. Why use different roots for each category? And with each root connoting slightly different, why not use that and pick the exact root that means best?
Yes. They had verbed it by the 12th century, according to: https://www.merriam-webster.com/dictionary/speed#word-histor...
Only in the archaic sense 3a, "to prosper in an undertaking", which is surely not what's meant in the TFA.
Have you known anyone who received a ticket (citation) for "speeding" while driving?
Yes, that's another different sense. In fact it's an intransitive verb that wouldn't fit here grammatically.
I sped read the title so didn't notice