Thursday, September 28, 2017

WSC 2017 Cruiser Summary

Solar car racing has never really been about practicality. Sure, teams have built and raced multi-seat solar cars in the past - Honda famously won WSC 1996 with their back-to-back two-seat Honda Dream, and 6m x 2m ISF6000 two-seat cars went head-to-head against 5m x 1.8m ISF5000 single-seat cars at ASC and WSC from 2001 to 2005. But none of those cars were designed to be practical; they were still race cars that were designed to be as light as fast as possible.

All of this changed in 2013 when WSC developed the Cruiser Class for their event. Cars in the Cruiser Class would be allowed to recharge their battery off the grid(!), and would be required to seat more than one person. In addition, for what I believe was the first time, the solar cars would be judged based on something other than which car crossed the finish line first. They would be judged on the sum total distance of people carried in the car (person-km), the amount of grid energy used, and the subjective practicality of the car.

In 2013, elapsed time was worth 56.60% of a teams final score, energy usage was worth 18.87%, person-km was worth 5.66%, and practicality was 18.87%. The formula was tweaked slightly for 2015 (elapsed time 70%, energy usage 15%, person-km 5%, practicality 15%), but the event was largely the same.

2017 Changes


There have been a lot of changes to the Cruiser class in 2017. For starters, I wouldn't call it a race anymore. Teams must arrive in Adelaide between 11:00 and 14:00 on the 6th day, and no credit will be given for arriving earlier. Without the "trying to go faster than the other teams" aspect, I'd simply call it a competition...

Scoring this year is 20% practicality and 80% "energy efficiency", where the "energy efficiency" score is calculated as person-km divided by energy usage - energy usage being the capacity of the battery pack, multiplied by the number of times a team charges off the grid + 1 (to account for the energy in the battery at the start line). While we're talking about charging off the grid, that has changed for 2017 as well - a team can charge off the grid anywhere, any number of times (in 2013, there were three locations, and in 2015 there was a single opportunity in Alice Springs). And battery size limits have been eliminated as well - a team can pick any size pack they want.

Speed and energy usage were such a big part of the old scoring formulas and person-km was such a small part; I really think that the old rules favored 2-seat cars*. 2017 is a completely different ballgame: instead of a dinky 5% of the score, person-km has a multiplicative effect on a factor that accounts for 80% of the score!

Looking at the energy efficiency score, I don't see any downsides to cramming as many people in the car as possible. Sure, a 4-seat car will have more aerodynamic drag than a comparable 2-seat car, but it won't have double the drag. It won't weigh twice as much either - many things will weigh the same regardless of how many people are in the car (the solar array, the lights, the steering wheel and pedals, etc). So by placing more people in the car, the drag-per-person and the mass-per-person both go down, which should reduce the energy-usage-per-person - exactly what the teams are being scored on. I'd also expect cars with more seats will place better on practicality, so more seats appears to be a win all around.

*Yes, Eindhoven won with 4-seat cars in both 2013 and 2015, but I'd argue this was due to the quality of the car and the team, rather than designs with more seats having an inherent advantage. In 2013, they carried more people, used the same amount of energy, and yet went FASTER than two of the three 2-seat cars that finished the course. 2015 was a similar story - the Eindhoven team was simply in a class of their own.

How many people can teams fit?


So how many people is it possible to fit in a solar car? Several things place an upper limit on the number of people that makes sense. Each person added increases a team's potential score by a smaller amount - the score increases towards a horizontal asymptote as people are added, rather than increasing linearly. But the score theoretically does keep rising as people are added... Practically speaking, however, an increasing number of people will at some point result in aerodynamic compromises to keep the car within the 5m x 2.2m planform, and at that point adding more people will rapidly stop making sense. Even earlier than that, a design will run into a weight problem - I think that's what ultimately puts an upper limit on the number of people that a team can reasonably fit.

The Michelin Radial X and Schwalbe Energizer S both have a load rating of 150kg (and I suspect the Bridgestone Ecopia is similar), so a 4-wheeled car can weigh a maximum of 600kg. Each occupant weighs 80kg (if they're lighter, they're ballasted up to that amount - reg 3.12.8), so a 4-seat car has 320kg of occupants, and a 5-seat car has 400 kg of occupants. This does not leave a lot of weight budget left over for the car itself! Typical Challenger cars weigh 135-150kg; add in some extra weight for a larger battery, more seats, stronger chassis and suspension, etc... and you'll be over 600kg really quickly. A 6-seat car would need to weigh less than even the lightest Challenger cars to stay within the load rating of the tires, and that just ain't happening. A 6-seat car would have to move up to real car tires, which would increase weight, aerodynamic drag, and rolling resistance - and that would result in a huge downward step function in score.

Of course, this is all probably irrelevant at WSC - given the "roll cages" they still allow, do you think they'll care about something as trivial as load ratings on the tires??? Eindhoven claims Stella Vie weighs 375kg, so with 5 people inside, it weighs 775kg - 30% over the maximum allowed load on the tires (and that assumes a perfect 50/50 F/R weight bias). If WSC was being strict, they'd only allow Eindhoven to drive with two people inside a car that heavy!

But even if WSC won't prevent teams from blowing past the load ratings on the tires, I'm not sure how far I'd want to push the limits. Some of the worst crashes in WSC history have been due to tire blowouts, and with such heavier cars, the kinetic energy to deal with in a collision is much higher... Honestly, the 4/5 seat cars give me a little bit of the heebie-jeebies; we've already seen a suspension failure from UNSW and some sort of structural chassis failure from Bochum, and the event hasn't even started yet.

(EDIT 9/29: A comment below notes that the load rating on the Bridgestone Ecopia is higher than the 150kg load rating on the Michelin and Schwalbe tires. They don't say how much, but anything over 195 kg would allow Eindhoven to carry all five people without blowing the load rating.)

How small can the battery get?


So we've looked at how many people a team can practically fit. What if we move in the other direction? What if instead of adding people to maximize the person-km/energy-usage equation, we minimize the number of people and radically reduce our energy usage instead?

Given the arrival window, a team needs to have an average speed of at least 65kph. This speed would be good enough for about 9th or 10th place in the 2015 Challenger class. So, imagine if Nuon elongaged their driver canopy and stuck a "cheater" seat behind the driver's seat so it's technically a 2-seat car... how small do you think they could shrink their battery and still be able to place around 10th?

Back-of-the-envelope calculations tell me that the point a cheater-car beats a good 5-seat car is around a 2.5kWh pack. If you built a high-quality cheater-challenger car for the Cruiser class with a sub-2.5kWh pack and managed to finish the competition without recharging off the grid*, I think victory would be possible**. I heard that a few different teams strongly considered building this sort of car, but none of them ended up doing so. I think this sort of car is just too risky - because if you mess up your strategy, end up behind the minimum speed, and have to charge off the grid just once, your score gets cut in half. Which brings me to yet another way that more seats are better this year: strategy risk.

*Is 2.5kWh enough to finish the competition with? Once the pack gets small enough that you can't store all the energy from an evening and morning charge, performance of the car starts to drop off quickly, and I'm not sure if 2.5kWh is past that line...

**I don't speak Dutch, but it appears that Eindhoven thought about doing this in the past - lots of highly aero-optimized 2-seat concept sketches are shown in this video.

Cruiser Strategy


Strategy has really been turned on its head in the Cruiser class this year. In a race, it's easy to adjust your strategy. Have a little extra energy in the pack? Speed up a little bit to burn it off, and place higher. Looks like you're going to have an energy shortfall? Slow down a hair to save some energy. But in the 2017 Cruiser class, you can't adjust your speed. No extra credit is given for arriving early, so there is precisely zero advantage to driving faster than 65kph (other than to build up a buffer to account for unforeseen issues). And you can't drive slower, because otherwise you simply don't finish. The only knobs a team can turn is how many people are in the car, and how many times they charge off the grid - and both of those are BIG knobs to turn, resulting in LARGE step functions in a team's score.

In the absence of speed as a variable that can be tweaked on the fly, the Cruiser teams are left with weight - in other words, how many people are in the car at any one time. They can only change this so many times - a seat must be continuously occupied for an entire leg between control stops for the person-km to be counted, and there are a limited number of control stops, as ScientificGems illustrates here. Therefore, the Cruiser cars with more seats have an advantage when it comes ot adjusting strategy over the course of the race, as they have finer control over the weight of the car. It's probably smart to design the car and strategy such that at the optimal speed, one of the seats is empty some of the time - this allows the team to increase their scoring performance if they're using less energy than expected. If a team starts off the competition planning to have all seats occupied all the time, they don't have a good way of increasing their energy efficiency score if the conditions end up allowing for it.

On the battery front, I'm not sure whether it's better to use a large battery that you plan on never charging off the grid, or a smaller battery that you intend to charge nearly every night. 

A smaller battery is lighter, and allows for more strategy modulation - skipping a charge or adding an extra charge has a smaller effect on score*. However, every time a team charges their battery off the grid, it's assumed they charged it from completely empty to completely full; there's no net-metering credit for only partially charging the battery. So every time a team charges overnight, they effectively "throw away" the "free" solar energy from that evening and morning charge, as well as whatever residual energy was left in the pack when they started charging.

A team that never charges off the grid doesn't "lose" any energy to the scoring equation, but they have to lug around a heavier battery throughout the competition. The strategy is riskier as well - if they fall behind and have to recharge off the grid a single time, their score is cut in half**.

*But still far larger an effect than adding or removing a person for a single leg of the competition, even in a 2-seat car.

**For what it's worth, I think Eindhoven is going with the "big battery, no recharges" strategy this year. They did 2015 with a 15kWh battery and a single recharge. This year they have a 12kWh battery, but competition speeds will be about 10kph lower, which means less energy expenditure over the event (and I think their aerodynamics have improved significantly as well). The array is smaller, but due to the lower speed they'll have more time under the sun, and I think they'll collect about the same amount of solar energy overall. I wouldn't be surprised if Eindhoven does the whole competition without recharging, and I would be surprised if they charge more than once.

Spectating


Due to how the competition is done this year, it will be impossible to tell how it's going just from the road position of the cars. I expect that all of the cars will be bunched very tightly together; all trying to drive no faster than necessary to finish before the deadline in Adelaide. Unless WSC publishes the battery capacity of each of the competitors prior to the event, and over the course of the event publishes when teams grid charge and how many people arrive in each car at each control stop, we will have literally no idea who is in the lead until WSC announces the winner at the awards ceremony. I implore WSC to publish this information as the competition progresses, so the public can follow along. If they don't, the Cruisers are going to drop off of everyone's radar the moment they cross the start line.

The Cars


There are 14 entries in the Cruiser Class this year:

Team
# of Seats
40: Eindhoven 5
45: Lodz 5
9: PrISUm 4
11: Bochum 4
23: Tehran 4
75: UNSW 4
14: Flinders 3
5: SunSPEC 2
30: Arrow 2
35: IVE 2
42: TAFE SA 2
49: Siam Tech 2
94: Minnesota 2
95: Apollo 2

Unlike previous years in which 2-seat cars comprised most of the field, half the field is doing more than two seats this time around. Unlike the Challenger class, none of the teams have taken the multi-junction array option.

Not too much more to say, so let's get to our top picks:

Eindhoven

Eindhoven is the undisputed champion of the Cruiser class, having won both previous events, and the car they've designed for WSC this year does not disappoint.

Photo: Bart van Overbeeke, CC BY-NC-ND 2.0
(image source)
Photo: Bart van Overbeeke, CC BY-NC-ND 2.0
(image source)
Photo: Bart van Overbeeke, CC BY-NC-ND 2.0
(image source)

Stella Vie takes advantage of 2017's larger bounding box and smaller 5sqm array to fit an extra person in a much more curvaceous car than 2015's Stella Lux. Yes, this is a 5-seat car, with 3-across in the second row! The quality appears just as high as on Eindhoven's previous cars, and as I detailed above, I think that designing for more people was definitely the smart strategy this year.

This is the team to beat. Full stop, end of story.

Bochum

Bochum has been credited with inspiring the Cruiser class - the team fielded a side-by-side 2-seat car at WSC in 2011, before the class existed. This year, they've built their first 4-seat car, Blue.Cruiser.

Photo: Stephan Schwabe
(image source)
(image source)
(image source)

Bochum's previous two Cruiser cars were hampered by their solar arrays. Both 2013's SunCruiser and 2015's SunRiser were built with 3sqm multi-junction arrays, and they were competing against cars with up to 6sqm silicon arrays. They were cute cars (particularly SunRiser), but they simply didn't have the array performance necessary to be competitive. This year, Bochum is getting more serious about winning - they fit the full 5sqm of silicon cells allowed; Blue.Cruiser won't be at a power disadvantage. The car also appears more aggressively streamlined than in the past, and it packs 4 people into an impressively tight cockpit (There's enough room in the back to nap comfortably, however).

Bochum had some sort of incident while testing in the outback - I don't have many details, but I've heard through the grapevine that the rear array panel blew off, and perhaps some structural chassis damage occurred. The damage sounds minor enough, and the latest news from their Instagram is that they're back on the road. I expect that Bochum will do very well in the Cruiser class this year - maybe even challenging Eindhoven for the win.

PrISUm

Iowa State University has been in the solar car game for a long time - I remember seeing an "Iowa State University: Shading your array since 1989" t-shirt at an event some years back. They've attended Sunrayce and the American Solar Challenge every single time the event has been run, from the inaugural 1990 event onward.

Their Cruiser entry into WSC 2017 is the first time they have ever competed internationally; they're bringing a 4-seat car named Penumbra.

(image source)
(image source)

On the one hand, PrISUm is a quality team - they've done very well at ASC lately, finishing 2nd in 2012 and 3rd in 2014 (and 7th in 2016's "Rain Rayce", but dang that was a rough year for everyone). The car looks extremely well finished inside and out; it should place very highly on the practicality portion of the event. On the other hand, I really am not a fan of the rear end of Penumbra. I can't imagine that it will be able to come even close to matching the aerodynamic performance of Eindhoven or Bochum, and aero performance still matters a ton in the Cruiser class.

But I think the most important factor is how prepared this team appears to be: PrISUm was one of the first teams to unveiled their car, and they took the car on a 3-week driving tour of all 99 counties in Iowa way back in June. This is may be one of the best prepared cars and best trained teams in the Cruiser class this year, and that alone is probably enough to put them into the top five. I wouldn't be surprised if they managed a podium finish, especially if the next two cars I list have issues on the Stuart Highway.

(EDIT 9/29: See this comment chain for a discussion of their solar array; I hadn't realized that Iowa State fits significantly less than the allowed 5sqm of cells on the exterior of their car. I stand by my prediction that they'll likely finish in the top five, however)

UNSW

Like Bochum and Eindhoven, Sunswift has competed in both of the previous editions of the Cruiser class. They finished 3rd with their 2-seat eVe in 2013, and brought it back to finish 4th in 2015. This year, they've built a 4-seat car named Violet

(image source)

Violet is very sleek, and it looks like UNSW has really engineered it to compete for the victory under this year's Cruiser scoring formula, but the team really let the construction schedule come down to the wire. They were one of the last teams to unveil, and actually had to delay unveiling due to a suspension failure while testing the car. I have concerns about the readiness of both the car and the team, but I still think they have a better shot than most of ending up in the top five.

Lodz

This team from Poland was a rookie at WSC in 2015, and was unable to complete the entire route. However, they completed the most person-km of the Cruisers that didn't finish, and the construction of the car appeared to be high quality.

This year, Lodz is the only team other than Eindhoven to construct a 5-seat solar car.

(image source)

Eagle Two looks like a well constructed car, and I have suspicion that Lodz may do much better in their second attempt at WSC (similar to WSU's improvement from 2013 to 2015).

Other Cars

I waffled back and forth on this for a while when I was writing this blog, but I ended up only picking cars with more seats for my top five predictions. However, it's worth mentioning a few of the 2-seat cars:

Team Arrow did quite well in the Challenger class in 2013 and 2015, and has shifted to the Cruiser class this year. Their car Arrow STF looks hot; it probably would have done extremely well in the previous Cruiser class, but given how the regs changed for this year... I also have the same concerns that I have with UNSW: The car was unveiled extremely late, and I'm dubious of the amount of testing that has been done on both the car and team.

Minnesota is the fourth team that has competed in both previous editions of the Cruiser class, and they've struggled to finish the event both times. Their car Eos II is visually a massive step up in quality from Eos in 2015, but it's still a 2-seat car fielded under regulations that seem to heavily favor more seats. Also, while writing up this post I realized that despite being unveiled in early July, I couldn't find a single mention of test driving the car until their recent trek north from Melbourne to Darwin... so I have some doubts about how prepared the team and car are. They'll probably finish in the top half of the field, but given their competition this year, I won't be surprised if they end up bumped out of the top five.

Finally, I'm not quite sure what to expect from IVE. They struggled badly in 2015; only completing 896 person-km. The car this year appears to be a much better constructed car, but I don't know how much they've really improved. It's also only a 2-seat car...

21 comments:

  1. Just a point of order and to make things even more complicated - charging once instead of no times will not automatically halve your efficiency score.

    It will halve your energy efficiency value but because that value is then compared to the most efficient car and multiplied by 80 it can mean many things.

    If you are still the most efficient team it would have no effect on your score at all. Any other scenario is almost impossible to foretell because it is based on what your competition does.

    Nigel

    ReplyDelete
    Replies
    1. Correct, but discussing that in detail would have made this post even longer, and I'd already spent far too long revising and re-revising this...

      Delete
    2. I can sympathise with that!

      Your work is much appreciated.

      Nigel

      Delete
  2. The Bridgestone Ecopia with Ologic has a stated load rating much higher than 150kg in the data sheet. However, based on its construction and performance I don't think it is really much more robust than the Michelin or Schwalbe. I think Bridgestone may have just erased the factor of safety from their technical specifications. If any teams are running Schwalbe tires they would have needed to purchase them as early as last year due to the end of production. Good for them if they got them. Although I would question their use (or the Michelin) on any Cruiser car due to the obvious overloading conditions. Realistically, the Cruiser cars really need to use better tires than the solar car specific options available. However that will probably triple the rolling resistance. That is why it is difficult to build a highly efficient and practical car.

    ReplyDelete
    Replies
    1. Presumably the figures given by the manufacturers for rolling resistance of solar car tyres would be based upon a loading lower than the maximum.

      If the narrow tyres are overloaded as we suspect then the actual rolling resistance will be higher due to increased deflection.

      Using a wider tyre with a load much lower than it would normally carry would mean that there would be less deflection than expected and therefore perhaps a lower R/R than predicted.

      So maybe not actually a tripling of R/R in reality?

      Nigel

      Delete
    2. If you have a current datasheet that you can you can send me, could you forward it on to me at mostdece.solar@gmail.com? I'd be very interested to see it. I was under the impression that pretty much any and all details about the "Ologic" Ecopias were covered tightly by NDAs.

      I recall in 2015, the Challenger teams were lamenting how fragile the Ecopia tires were and how often they flatted. I'm assuming the 2017 tires have been improved, but still - I'm more than a little sketched out at the loads the 4/5 seat cars are going to be putting on those tires.

      An aside, the tire situation in solar car racing really sucks right now. All of the old 2.25" wide tires are out of production now, as is the Schwalbe 95/80R16. The only tires left are the Michelin Radial X (super expensive, restricted supply) and the Bridgestone "Ologic" Ecopia (Ultra-restricted supply, only available for WSC). I've heard a lot of American teams may be forced to use high rolling resistance scooter tires at ASC in 2018, simply because the special solar car tires are impossible to acquire.

      Delete
    3. Re the Tyre situation.

      I have seen a post, which I can't find now, from someone trying to get a bulk order of Schwalbe tyres together.

      Apparently they will make a batch of 50, which doesn't seem a huge order really.

      Nigel

      Delete
    4. Schwalbe had meetings this month to discuss any continuing production of tires. They decided to not produce any. There will be no bulk orders.

      In reply to Mostdece, I do not have a data sheet. I just know that one of the cruisers competing this year would have overloaded the 150kg tires, but came in under the load limit for the Bridgestone. I can't provide any more detail than that.

      In response to Nigels R/R question. I think most widely available scooter tires have a pressure rating of 30-40psi. The solar car specific tires are 70-100psi. That is a significant difference. I have also seen results from coast-down tests comparing the two and the car was able to coast about 3 times farther on solar car tires than scooter tires.

      I am interested to see what all of the teams at WSC will be using, for sure.

      Delete
    5. Hmmmm. Well, five of the seven cars that seat more than two people are on the Bridgestone tires sponsor list - Tehran and Flinders are the two exceptions. We'll have to see if WSC limits the number of people they can carry based on the load rating of their non-Bridgestone tires. Tehran's car in particular looks extremely heavy...

      Delete
    6. I would be pretty surprised if the 'B'WSC imposed that limit.

      Delete
  3. Also note that PrISUm only has 250/325 solar cells.

    ReplyDelete
    Replies
    1. Crap, really? I went and did a solar cell count on Eindhoven and Bochum to verify, but I assumed a car as boxy as Penumbra would obviously fit the full array. Is it significantly shorter than the maximum length?

      Just a lot of weirdo design decisions on that car...

      Delete
    2. Hmmm, I count 269, but yes - barely more than 4sqm of cells on the exterior of the car.

      (13W*9L panel on the hood, minus one cell on each front corner: 115 total on the hood. 12W*14L panel on the roof, minus 3 on each rear corner and 4 on each front corner: 154 total on the roof).

      Delete
    3. I counted 94 on the hood.

      Delete
    4. https://twitter.com/IowaStateUNews/status/908385894469656576

      13 wide x 9 long, yes? = 117. Minus the two missing on the corners, 115.

      Delete
    5. Ok now I'm confused, try counting this pic:
      https://scontent-ort2-2.xx.fbcdn.net/v/t1.0-9/19275213_1783551471659975_8301097377917923862_n.jpg?oh=12c90661831b2c5f6ce1f45fdf90e699&oe=5A4BE628

      Delete
    6. I think they made a new front array sometime after their big tour of Iowa

      Delete
    7. First array looks like a wet encapsulated top-contact array, those cells are often larger than the Sunpower cells:

      https://www.facebook.com/PrISUm.Solar.Car/photos/pcb.1767169229964866/1767168886631567/?type=3&theater

      https://www.facebook.com/PrISUm.Solar.Car/photos/pcb.1772433432771779/1772432569438532/?type=3&theater

      They made a whole new array, hood and roof. Note that the front corners of the old roof array are only missing 3 cells each, the new roof array cuts 4 cells on each front corner of the roof.

      Delete
    8. Ah, the mystery is solved. Now we can just speculate as to why they have the curved trunk instead of a few more rows of cells in back.

      Delete
  4. As an internet spectator of this event, this information is very helpful. Thank you.

    ReplyDelete
  5. This comment has been removed by the author.

    ReplyDelete