Monday, November 27, 2023

WSC 2023 Cruiser Wrapup

Another week, another blogpost. This time: Cruisers

Like we'd speculated in the Day 4 post, because everyone failed to finish the leg to Coober Pedy, only the first two days of the event (the Darwin to Tennant Creek stage) "counted" in the team's scores. Some teams just kept on truckin' anyway - Minnesota, Solaride, Ascend, and Apollo drove most or all of the distance after Coober Pedy anyway even though nothing mattered anymore (good for them), while teams like Sunswift, Flinders, and Sun Shuttle packed the car in the trailer all the way to Port Augusta before driving just the last leg.

Ultimately, Sunswift came out on top. With an efficiency score of 109.44, a practicality score of 83.3%, and no time penalty, their final score was 91.17.

Their nearest competitor was Minnesota with an efficiency score of 51.98 and a practicality score of 72.6%, good for a final score of 37.74 - except with an 0.59 penalty multiplier for being 52 minutes late at TC, bringing their score down to 22.38.

In third was Solaride with an efficiency score of 39.44, practicality score of 71.8%, un-penalized score of 28.32, but an 0.52 penalty multiplier due to being 65 minutes late at TC for a final score of 14.73.

I've been waffling about how much of a rant I should post here: how many feathers do I want to ruffle and how much I should censor myself, but I've just gotta get it out... IMHO the Cruiser Class has a ton of problems. Warning: Some saltiness ahead.

How the Rules Define The Race, or: The Failure Machine

Background: Currently, the rules define the score as "Efficiency * Practicality * Penalty", where "Efficiency" = Person-KM/External_Energy, "Practicality" is a value between 100% and 0% that is subjectively judged at the finish line, and "Penalty" = 0.99^Minutes_Late. Each stage (Darwin to Tennant Creek, Tennant Creek to Coober Pedy, and Cooper Pedy to Adelaide) has a target finish time and teams start accruing penalty minutes for missing those target times.

I sort of understand how WSC got to where they did with the scoring equation. The 2013/2015 sum-of-factors was kind of a mess: it was very unclear how a team should balance the various goals (minimizing energy usage, maximizing person-km, and going fast) to score as well as possible on the on-road portion - even though all of those boil down to "build a more efficient car". Locking at least one of those numbers would allow for a clearer on-road performance metric. The Cruiser teams have repeatedly expressed that they want to have freedom with the grid energy charging and number of occupants, so WSC drew a line in the sand with speed and gave the teams a fixed target time. But my opinion is that the fixed target speed/time turns the event into a machine that encourages teams to fail.

The Challenger class goal is to go as fast as possible, which places them as far away from failing to finish the event as they can get. If a team has unexpected issues, they fall back and don't finish as well as they'd like, but they have been incentivized for the entire race to build as much margin as possible ahead of the failure line. Meanwhile, the Cruiser class goal is to minimize energy usage within the target time, and that means driving as slowly as allowed. And there's not very much margin between the target time and the failure time: three hours at Tennant Creek, two and a half hours at Adelaide, and only half an hour of margin at Coober Pedy - after the longest and most challenging leg of the race for the Cruisers. That's not a lot of wiggle room between "ideal" strategy and failing out of the event!

If it's a rough year for weather (like this year, with smoke and headwinds in the middle), maybe fewer Challengers finish but at least we still have a full podium (and then some). But with the Cruisers, all the teams were planning to toe the failure line, and instead trip and fall over it together.

"But they should just have planned to drive faster in the first place to build margin for themselves!" OK but A) Going faster uses more energy. For bigger, heavier, less aerodynamic cars like the Cruisers, a lot more energy. That means more battery, which is more weight, which needs a stronger chassis to carry the loads, which is even more weight, and now the car needs heavier suspension to support all of that, and needs more battery to power the heavier car, and oops now you're in a weight spiral. A team that designs with energy margin for a year with bad weather is going to score poorly on efficiency on a year with good weather because they're going to expend more energy lugging all that extra weight around ("But they should just design a more efficient car instead of piling in more batteries!" We'll... get to that). And B), even if the teams build appreciable time margin into their strategy... if it looks like the efficiency scores are going to be close, the teams are going to want to pay off that margin and slow down to reduce unnecessary energy expenditure in order to improve their scores. Two teams can end up playing chicken with one another until they both run off a cliff. Even if a game of chicken doesn't happen, teams that build schedule margin early are still going to be tempted to slow down and start reducing their margin in order to increase their efficiency scores as they near the end of the stage. And if they break down at the wrong time... Tokai had to spend two hours on the side of the road with a suspension issue this year, and they still finished well. If they'd have been a Cruiser team? That's it, they'd have been out.

Let's look at how fast the Cruisers are expected to drive: in 2013 and 2015, teams could at least finish the Cruiser class at a ~60kph pace, and the target speed in 2017 was ~65kph. But in 2019 and 2023, the target speed on the TC to CP leg is 75kph and the fail-out-of-the-event speed on that leg is 72kph - that's the overall pace that Top Dutch set this year in 6th place in the Challenger class, so that's a pretty high bar to set for the cruiser teams! And in fact, if we look at the Challenger teams interval times on the TC to CP leg, that's where all the Challengers drove the slowest this year due to the headwinds and smoky skies. If we'd held the Challengers to the same schedule as the Cruisers, we'd have had 4 finishers this year instead of 12: only Innoptus, Twente, Brunel, and Michigan would have completed the second stage. Sonnenwagen's interval time is 37 minutes longer than the Cruiser cutoff time, Top Dutch's is an hour and seven minutes too long, Tokai would have shown up two hours and thirteen minutes past the cutoff time (due to their aforementioned suspension issues making them park on the side of the road for 2+ hours), and every other team behind was far too slow. That's rough!

This exposes the difficulty of running a staged event simultaneously with an un-staged event, while holding the staged cars to the same overall time limit as the un-staged cars. The Challengers have flexibility to allow their speed to wax and wane over the entire length of the race; it doesn't matter as much if the entire field has a slow day or two working through a weather system. But the Cruiser class is a series of 2-day sprints instead of a 6-day marathon, and failing a single one of those sprints causes a team to fail out of the event. Even if the Cruiser class became an elapsed time event while sticking to the current staged format, it would still be a harder event to complete than the Challenger class. The Cruisers would never be given the opportunity to build much margin past the failure line: they park for the day at Tennant Creek and Coober Pedy no matter how early they arrive, and any margin they've built up over the stage goes away.

Watching the Event

The target time is no fun as a spectator. In the ideal scenario, all the Cruiser teams are at essentially the same position on the road, and there's no way to tell from road position how they're doing on the things they're actually scored on. The only thing you can tell during the race day from road position is a binary "are they failing out of the event or not". Spectating the road position of the Challengers is fun and exciting, we're cheering as our favorite teams charge ahead and groaning as they fall back, and eagerly watching tight contests. But when I look at the road position of the Cruisers, there's nothng to cheer about; I'm just peeking out from between my fingers with dread.

It's also impossible to write about or explain to people how the teams are doing over the course of the event - even to technically-minded solar car alums! Absent road position mattering, we only really get a snapshot of how the teams are doing every two days, and it's "low resolution" summary information. For instance, as far as I can tell there's no publicly-accessible record of how many people the teams carried over each control stop interval, the results page is only set up to show us the total energy used rather than the amount charged at each stage stop, etc. The scoring is even difficult to explain in a useful way after the fact - witness ScientificGem's variety of different attempts over the years.

The Teams and the Cars

As a cranky old solar car designer with a competitive streak, when I survey the field of competitors, it feels like the teams are more confused by the rules than anything else. The cars don't seem to be designed to actually do well under the rules of the competition, and that's intensely frustrating.

I would like to start by pointing out: this year the Cruiser cars finished exactly in order of "efficiency". We could eliminate practicality, and even the minutes-late penalty time, and the 6 cars that completed the first stage are ranked in exactly the same order. And this isn't a fluke: The three cars that finished the event in 2019 ended up ranked in order of efficiency, and ditto for the three cars that finished in 2017. As long as we've had the efficiency metric and the practicality metric, the spread in efficiency scores has always been large enough that the spread in practicality scores was not large enough to alter the placement of the cars! If you want some numbers: the spread in efficiency this year between the 6 cars that made it to Tennant Creek was 6.28x. The efficiency spread between the top two cars alone was 2.11x! That's about as big as the total spread in practicality scores has ever gotten: 1.48x in 2023, 1.74x in 2019, 2.38x in 2017, 2.36x in 2015, and 1.76x in 2013.

It sure seems to me there is a ton more room right now for the teams in the Cruiser class to compete on efficiency rather than practicality! Yeah practicality being a score multiplier that can technically be zero in 2019 and 2023 is big and scary, but efficiency is a multiplier as well in the scoring equation! In practice, practicality varies between about 0.5 and 0.9 rather than between 0.0 and 1.0, but the efficiency score sure as shit can and does go to zero if your inefficient car fails to finish the event! So why does every single cruiser car nowadays look like it was designed with the practicality portion of the score as the primary concern and efficiency as a secondary concern? Why are they all trying to copy "normal car" styling cues to the detriment of efficiency?

There is something about how the scoring is set up that seems to be obliterating "solar car thought" from the teams that are entered in the Cruiser class. Post-event this year, some WSC officials brought up that they really aren't stoked by cars with 40, 50, and even 60kWh batteries in the Cruiser class nowadays (at this point it's basically an EV race with fig leafs symbolic solar cells on top of the cars), and proposed bumping the Cruiser array size up from 5m² to maybe as high as 8m². The response from some team members was "no, we have enough trouble fitting 5m² on the car as is". What??? That ain't solar car thinking! They should be absolutely stoked at the idea of getting more solar generation on the car, getting heavy batteries out of the car, and getting into a virtuous weight cycle in their designs. Teams had to fit 6m² of array on 4.5x1.8m Cruiser in 2013, they should be able to figure out how to fit 8m² on the current 5x2.2m cars. Another comment was "it's just too much to expect us to be able to go 1200km from Tennant Creek to Coober Pedy without recharging, we need to add back the third grid charge location in Alice Springs". Give me a #$%@ing break, the Challenger cars do Darwin to Adelaide without charging off the grid! They do it with comparatively eensy-weensy batteries, too. And y'all know it, you've seen it! You were at the same event! You're sitting there with them at the finish line and saying shit like this!

To refresh all of our memories about what the Cruiser teams can do (when they want to), I went back and looked at the 2013 and 2015 Cruiser average speeds, when elapsed time mattered in their scores. In 2013, the teams had ~16kWh batteries and could charge in TC, AS, and CP. That's a total of 64kWh available - well less than the 100kWh++ that teams were on track to use this year, had they finished - and three teams finished with average speeds >>80kph over the course of the whole event. But OK, they had three charge locations rather than two. So let's look at 2015, when the teams were limited to ~15kWh batteries and they were only allowed to charge off the grid once: in Alice Springs. That's only ~30kWh available over the entire race; less than any of the teams this year had in their battery packs on the start line! Kogakuin averaged 90.59kph over the whole event. They used less grid energy to go 3022km in 2015 than any team used to go 985km this year, and they were going a solid ~30% faster! Teams are complaining that Tennant Creek to Coober Pedy is too far to go without recharging? In 2015 they had to go 25% farther than that - twice! If your response is "Kogakuin's 2015 car was basically a cheater-challenger car disguised as a cruiser, they only had a single person in the car the whole way and it got a bad practicality score", well: Eindhoven averaged 86.62kph - just a smidge slower than Kogakuin - on the same amount of external energy, had two people in the car the whole way, and got the highest practicality score that year! It is totally possible to build a multi-seat solar car that goes fast and has a small battery and doesn't charge off the grid frequently.

Ok ok Eindhoven may be an outlier team, they've dominated every year, so let's look at Kogakuin again. Even if you don't like Kogakuin's strategy of favoring efficiency and basically neglecting practicality in their design, it was a good one that netted them 2nd place overall in 2015. And practicality becoming a multiplier in the 2019/23 score equation doesn't actually wreck this car; a highly-efficient car like that would still place extremely well under the current rules! Let's look at how it would do in 2023: If we took that car with its small 15kWh battery, gave it the same practicality score from 2015 (which would be the worst by a decent margin in 2023!), filled both seats (they only put one person inside in 2015, but we're going far slower + less distance between charge locations), and it only finished the first stage like all the other cars this year, its efficiency score would be high enough to offset the low practicality such that it would still finish 2nd overall in 2023. And by a pretty wide margin over the next car! Sunswift's Efficiency*Practicality score was 91.17, Kogakuin's theoretical 2015-car-in-2023 would have scored 67.97, and Minnesota's score was 22.38 (if un-minutes-late-penalized, it would have been 37.74).

And if Kogakuin actually finished the event, which that car could totally do at a 75kph pace even with the weather this year... It would have, y'know, won the 2023 Cruiser class, regardless of how far WSC nuked its practicality score?


Author photos of Kogakuin at WSC 2015

It would have done extremely well in 2017 and 2019 too, in years where it would not be the only finisher! Remember its performance in 2015: 90kph, one person, 30kWh. In 2017, if they stuck two people in the car and kept their total energy to under 24kWh - plausible considering that they'd only need to drive 65kph to meet the target time that year - they would have won! (it's also they likely didn't use a full 30kWh in 2015; that year charging was not metered and the score just assumed a full pack every time the car was connected to the grid. It's unlikely their battery was COMPLETELY empty when they charged in Alice Springs. Of course, charging was not metered in 2017 either, so they would have needed to shrink from a 15kWh pack to a 12kWh pack). In 2019, they would have needed two people in the car the whole way and to keep external energy under 30kWh to win, maybe more of a stretch at 75kph - but they'd have had the advantage of net metering that year, so it still seems plausible? And even if they couldn't pull off the win in either 2017 or 2019, the chasm in scores between 1st and 2nd in those two years was such that they'd have had 2nd place completely locked up.

Any team can look back at Kogakin's performance in 2015 and game out how they'd do under the current regs. So why aren't many (any?) teams building hyper-efficiency-focused Cruiser cars? Knowing everything we do about how the efficiency scores have played out vs the practicality scores, why are all the teams showing up with cars that appear to be emphasizing (imagined) practicality over (real) efficiency??? It's just inexplicable to me, it makes me want to tear my hair out.

Every now and then we get to see something good. For instance, Onda Solare's Emilia 4 at ASC 2018 was far more efficient than everything else that showed up, and ran away with the event. That car was extremely lightweight and had a clear focus on aerodynamics.

Emilia 4, via Facebook

See also Polytech Montreal's Esteban X at ASC 2022 - this car was extremely lightweight, it absolutely clowned on everyone's efficiency scores, and only lost due to getting DQ'd on a technicality (they got lost and took a longcut, inadvertently skiping an eensy portion of the route in the process). This car only had a 9.2kWh battery, and only slurped 3.3kWh off the grid over the course of the event. That's a real solar car!

Esteban X. Sharp trailing edges? On a Cruiser car? Yes, it's possible!
image source

I was also really interested by ATN's Pricilla at WSC 2023. Unfortunately suspension issues prevented them from finishing the first stage so we never got to see performance numbers, but it's still recognizably a solar car: Prioritizing flat array area over cosmetic curvaceousness, skinny wheel fairings, relatively slim trailing edges... This is the kind of car that I expect to see more of, but don't.

via Facebook

via Facebook

And of course, Eindhoven has dominated every prior WSC Cruiser event due to being wildly more efficient than everyone else. Yes, they've always topped the practicality scores, but never by very much - the vast chasm in the final scores between them and 2nd place has always been due to the difference in efficiency.

But this year, man... we really didn't get any super efficiency-focused cars in the Cruiser class. The batteries were all larger than before, the cars were all h e a v y, and they all had giant blunt rear ends that I'm still scratching my head over. A member of one team tried to argue with me that it was actually more aerodynamic, and, c'mon buddy. Sure, a Kammback is better than some other blunt shapes but it's still worse than a nice sharp airfoil; you don't see any Challenger teams making Kammbacks. Cessna or Cirrus aren't designing light aircraft with Kammbacks.

The only explanation I have for why the Cruiser teams seem to be lacking a focus on efficiency is a combination of A) teams look at the practicality multiplier in the 2019/2023 score and worry that WSC will put their finger on the scale at the finish line practicality judging to nuke their score to zero if WSC feels like their car wasn't in the "spirit" of the Cruiser event, so they focus on "soft" practicality considerations like "normal-car-like-aesthetics" more that the historic spread in Efficiency and Practicality scores implies that they should, and B) "you should use less energy than everyone else" is somehow a worse incentive to design an efficient car than "you need to go faster than everyone else".

Where do we go from here

Man, I dunno. I'm just monday-morning quarterbacking here, I fully admit that it's a lot easier to critique an event than it is to design a better one. And I don't have a lot of good answers here that won't make everyone mad in some way. The teams have repeatedly expressed the desire to keep the design space wide open in terms of batteries and number of occupants; I think this is one of those situations where people just don't know what's good for them. So if I was king:

My strongest opinion is that it should go back to an elapsed time event where the goal is to go faster: it's better for the event to encourage the teams to get as far away from failing as possible, it makes the event way more fun to watch and easier to understand as a spectator, and (I think) it actually made the teams care about building efficient cars in the early years of the Cruiser class in a way that the current target time fails to do. But if we re-introduce elapsed time variability, at least one of the other on-road factors has to go.

I'd scrap person-km altogether; just fix the number of people in the car and declare all seats have to be occupied at all times. The "allow teams to pick the number of seats" has never really been interesting to me from a competition design perspective. The team with the most seats occupied has always won the Cruiser class, the scoring equation for the past three events has very clearly favored "as many seats as possible", I don't really see much point in continuing to allow variability here going forward. And hey - if you start off with a full car in Darwin, but can't carry every person in the car all the way to Adelaide, that doesn't seem practical to me? You shouldn't require some of your passengers to hitchhike part of the way? The troll in me wants to say we should require three seats, because it's not immediately obvious what's the best way to arrange the seats and we'd probably get a fun variety of solutions.

Let's give them 8m² of array to play with, so they have a chance to be solar cars again. But remember: cars with 6m² of array (and one or two occupants) in 2015 were "too fast", so we'll need to be careful with how much external energy to allow.

Speaking of external energy, I'd really prefer to lock that down as well: no grid charging, no battery size variability, lock them into a fairly small battery, and the score is just something like "80*normalized elapsed time + 20*normalized practicality" or "normalized elapsed time * practicality, where practicality ranges from 1 to 0.5", or something similar. That would also let us eliminate the staged nature of the Cruiser event to give them more flexibility with their speeds over the course of the entire event. But, this might result in Cruisers beating Challengers to Adelaide, and I've been told that's a big no-no.

So one option would be to leave the battery size open and give 'em one grid charge location in Alice Springs: staging all the Cruisers up there on the third night means they finish about a day behind the lead Challengers no matter how fast the Cruisers manage to go, and I think teams have a better chance of success with two 3-day stages instead of three 2-day stages (the "fail out of the event" speeds are 60.9kph for 3 days from Darwin to Alice Springs and 61.1kph for 3 days from Alice Springs to Adelaide; far more forgiving than the current 72.4kph minimum speed for 2 days from Tennant Creek to Coober Pedy). In this case you would need some sort of exponential term applied to the energy usage part of the scoring equation to make burning energy to go faster have diminishing or even negative returns past a certain point; imho you'd want the "sweet spot" in the score to be an elapsed time around 75-85kph average speed. You don't want the ideal strategy to be to toe the speed limit regardless of energy usage (kind of like it almost was in 2015). But this may be a hard target to hit in the rules...

Or, just don't worry about that at all? Peg the battery size small enough that even with a grid recharge in Alice Springs, it's unlikely that anyone can sustain speeds that are toeing the speed limit? Eliminate "external energy" from the scoring entirely - like with the Challengers, just assume that it's in every team's best interests to max out the (small) battery size and also in their best interests to use the charge opportunity in Alice Springs to the fullest (This is what happened in 2013 and 2015: everyone's energy score was the same, ~modulo some eensy variations in pack construction). If a team messes up their strategy such that they haven't gotten close to the bottom of their pack by Alice Springs, or if they have electrical gremlins and can't charge all the way up to a full pack, well, that's their problem. Now the team's elapsed time to Adelaide is essentially a direct measure of the "efficiency" of the car and team. This option would be my personal preference.

Or if we must judge the cars on a variable amount of external energy, another fun idea would be to allow grid charging at multiple control stop locations... but grid charging can only be done on the clock between 8am and 5pm; no grid charging in the morning and evening. If you want to grid charge, you pull out of the control stop after serving your 30-minute time and pull into a metered grid charge area. This would force some hard decisions from teams about if time or energy is more valuable... but could also be confusing for spectators, confusing for team strategy, and could lead to some drama for officials if the grid charge equipment doesn't provide the charge rates that was promised to teams... Maybe table this idea for now.

Alright, I think I've typed too much about this already. See y'all when the WSC 2025 season starts heating up.

Sunday, November 19, 2023

WSC 2023 Challenger Wrapup

It's been a week TWO WEEKS THREE WEEKS since the awards ceremony, so let's wrap this up.

Chart via Scientific Gems

Similar concept to above, but inverted and with granular data rather than just control stop times
Adapted from work by Jelmer van de Wiel

Quick recap: Over the last two days of the event, there wasn't any movement among the top four. Twente closed the gap to Innoptus from 33 minutes at Erldunda in the middle of Day 3 down to 16 minutes at Coober Pedy on the morning of Day 4, and then 14 minutes at Glendambo... and then back up to 30 minutes at Port Augusta before the end of Day 4, and that was that. Speed limits drop soon after PA and catching up becomes extremely difficult, barring your competition breaking down. A second consecutive win for Innoptus proves that 2019 was not a fluke: The have what it takes to win even without their two biggest competitors wrecking out or burning down. And second place is still a great result for Twente, besting their third place finish from 2015!

As of the last update on this blog it still looked like anyone's race among the top 3, but afterward Delft slowed down dramatically relative to the top two; basically taking them out of contention for the win. They managed to stay on the podium and finish in 3rd, however. It's worth noting how big of an upset this is, historically: Other than coming in 2nd to Tokai in 2009 and 2011, and the DNF due to the fire in 2019 (after a super tight battle for the lead...), Delft has won every other WSC they've entered - 2001 through 2007 and 2013 through 2017, seven victories. At 2 hours and 18 minutes, this was their second biggest margin of loss: only 2009 was larger at 2 hours and 49 minutes. Talking with some team members after the race, it didn't sound like they had any critical issues this year - they just think that they've taken the catamaran design as far as it can go, and this year proved that the monohull design concept had more performance potential than they thought.

Behind Delft, Michigan was slowly catching up over the last two days, but they were just too far back to close the gap - Michigan was still an hour behind at Port Augusta, and had to be content with 4th place. Speaking with the team after the event, they were pretty sure they could have finished 3rd if they hadn't botched the hot lap - starting at the tail end of the grid put them 40 minutes back at the start line, and overtaking all of the teams certainly ate up more time. It's very easy to picture them being ahead of Delft at Port Augusta if they'd started up in the top 10. On the other hand, Michigan also had a lot more unplanned side-of-the-road pauses than Delft, so they can't blame their 4th place finish purely on starting position - Delft just ran a cleaner race! I wonder if this is a visible result of Delft finding races to attend in 2021 and 2022, whereas Michigan has not had to deal with the pressures of a competition since 2019.

Just Delft and Michigan, for clarity
You can see a lot more little stops peppered around on Michigan's line

Sonnenwagen rolled on the last day of the event, in a situation that sounds eerily similar to their 2019 accident: in a situation with high crosswinds from the right, the team passed by an oncoming road train - which blocked the crosswind momentarily - and then lost directional stability in the combination of the sudden return of the crosswind and the wake of the road train. Fortunately the driver was largely unhurt and able to walk away from the crash, and the car appeared to mostly suffer only cosmetic damage to the structure. The team claims that they could have continued, but it rolled entirely over and the regs this year state that if the car rolls it's automatically removed from the event.

Tokai finished 5th, around two and a half hours after Michigan. It might have been a tight competition for 4th with Michigan, but Tokai suffered some sort of suspension failure over a cattle grid in the late afternoon on Day 4. This put them on the side of the road for about two hours, and momentarily behind Top Dutch. But they were able to repair the car and regain their 5th place position quickly, finishing a comfortable hour or so ahead of Top Dutch's 6th place finish.

1st to 6th was a seven hour spread overall, and then we had a five hour gap back to 7th place, followed by less than a two hour spread between 7th and 11th! That's where the real action on the road was this year and I kind of wish I'd been back there to see more of it.

JU ended up on top of that group, prevailing in a tight multi-day battle with Kogakuin and Durham. 7th is also a best-ever finish for JU, ahead of their 8th place finish in 2017. Kogakuin finished in 8th, while in 9th was a surprise interloper - Western Sydney! WSU had been spending most of the race further back with Blue Sky, Goko, and αCentauri, but they got the car working the way they wanted to (or found the accelerator pedal?) sometime on the fourth day and made a breakaway ahead. At one point I was wondering if they'd be in contention for 7th! Even if they ultimately couldn't come out on top of this group of teams, it still had to be satisfying to finish on a strong note instead of in limp-mode. In 10th was ETS Eclipse from Canada, who also staged a late-race comeback, abet in not as dramatic fashion as WSU. And at the tail end of this group in 11th was Durham - completing WSC without trailing for the first time, on what I believe is their fifth attempt. A big congratulations to them!

It's really interesting to contrast the paths the teams took to this tight finishing group. Kogakuin started up front - on pace with the likes of Top Dutch, Tokai, and Michigan - until they progressively fell off-pace after Dunmarra. Meanwhile WSU was frankly on a not-finishing-the-race pace all the way through Alice Springs, and these two teams finished side-by-side 8 minutes apart.

Rounding out the finishers a little over an hour back from the 7th-11th group was the αCentauri team in 12th.

Of the teams that failed to finish, Sonnenwagen Aachen was the first to get past Port Augusta, so they received 13th place. 14th and 15th were Goko and Blue Sky - unfortunately they were about 60km shy of finishing the event by the 5pm cutoff on Day 6.

The last car still out driving at the end of Day 6 was Chalmers - they simply refused to put the car in the trailer until the race was really truly over. Earlier in the race they missed a control stop cutoff and were listed as "no longer competing" on the race map. Did they give up? No! They kept driving, and made the cutoff for the next control stop, and WSC changed their race map dot back to green! In the end they made it past Port Augusta and to within 260km of the finish. They actually made it past where Sonnenwagen wrecked - if WSC counted driving distance between control stops in the final rankings, Chalmers would have been ranked 15th, over Sonnenwagen in 16th. Never give up, never surrender! I would have nominated Chalmers for the spirit of the event award if I could!

No one behind Chalmers made it past the midpoint of the event. AUSRT called it between Barrow Creek and Alice Springs, Halmstad ended their race between Tennant Creek and Barrow Creek, Wakayama put it in the trailer between Dunmarra and Tennant Creek, ITU and ANU ended their race after Katherine, and Arrow and TUCN didn't even make it as far as the first control stop.

Overall this was a really clean year for the Challengers, which was a welcome change from 2019. Given that this year was the first run of the new 3-wheel regs, I doubt we'll see any huge shakeups in the formula for 2025, absent two possible areas: Active aerodynamic devices, and battery size.

Innoptus's big "secret weapon" - and winner of the innovation award - was their big retractable and steerable fin. They claimed it allowed them to reduce drag and improve the stability of the car in crosswinds. Post-race, the top teams seemed pretty unanimous - they don't want to see devices like this banned in 2025, they all want a chance to experiment with active aero devices themselves - but we'll see how the WSC technical staff feels about it.

On the battery front, WSC limits battery capacity by regulating the weight rather than by kWh, but they've reduced the allowed weight over the years as batteries have gotten better: we were allowed 35kg of Lithium-Ion batteries in 2003 and 2005, 30kg in 2007, 25kg in 2009, and 21kg in 2011 and 2013. These weights corresponded to about 5kWh of capacity. Then the weight was dropped to 20kg in 2015, and... that's where it's stayed. Lithium battery technology isn't advancing as fast as it was in the 2000s, but it is still advancing, and in 2023 the top team's battery capacity has crept past 8kWh and is knocking on 9kWh. 

Meanwhile, over that same timespan we've cut the array size way down. We've gone from "literally as many solar cells as you can fit on a car 1.8m wide x 5m long x 1.6m high, of any kind" in 2003 and 2005 (in practice ~8m² of dual- or triple-junction gallium), to 6m² of any array chemistry 2007-2009, to 6m² of silicon cells 2011-2013, to 4m² of silicon cells 2017-2023. Essentially, cars are making half as much array output as they did ~20 years ago - but they're also half the size, and much lighter and lower drag.

The long and short of it is that we've gone from the energy in the battery on the start line being about ~10% of the total available energy over the race 2003-2013, to being over 30% of the available energy in 2023. This doesn't feel great to some of the teams - it's supposed to be a solar car race, not an EV race. Given that the cars are also lighter and much lower drag than they used to be, that increase in battery capacity represents a lot more time/distance that teams can just punch through bad weather without slowing down. It's also changed strategy a lot - back in the day, we'd need to get pretty far down near the bottom of our ~5kWh battery packs every evening, so that we wouldn't "waste" any of the energy we'd gather from our ~1.8-2kW solar arrays by pointing them at the sun in the evening and following morning. Every day of the race was risky, misjudge the weather and you bounce off the bottom of your battery pack before the end of the day! But when the batteries are 8kWh and the arrays are sub-1kW? Teams will never get close to filling it in an evening and morning charge, they can hold a lot more energy in reserve.

The TL;DR is that everyone agrees that the batteries need to shrink, it's just a question of how much. The conservative answer is that we should just get back to ~5kWh or so - but this still leaves us with "proportionally" more battery storage relative to solar collector output and aero drag than we used to have. The aggressive answer is that we should shrink the batteries down to ~2.5-3kWh to be proportional again, but that might cripple the weaker/slower teams... it'll be interesting to see what WSC decides.

I had a ton of fun this year, and really reminded myself how much I love solar racing. Hopefully I'll be able to cover WSC 2025 more like I did 2015 and 2017.

Stay tuned for a recap and some thoughts on the Cruiser class.

Wednesday, October 25, 2023

WSC 2023: Day 4

Hello from windy Coober Pedy! 




Teams forming walls to shelter their cars from the wind in the Coober Pedy control stop

A brief update from Day 4: Great sun today, but worse headwinds than yesterday. The top five spots in the Challenger class seem pretty locked in and the leaders will finish early tomorrow. The Cruiser class is... a bit of a mess.

Chart via Scientific Gems

In the Challenger class, Innoptus remains in the lead and ended the day well south of Port Augusta. Twente managed to close the gap from 27 minutes at Erldunda yesterday, to 16 minutes at Coober Pedy early today, and 14 minutes at Glendambo afterward... but the gap had once again widened to 30 minutes again at Port Augusta and appeared largely unchanged through the remainder of the day. Barring an extended breakdown, it seems extremely unlikely that Twente will be able to take the lead now that both teams are deep into the area approaching Adelaide where traffic is heavier and overtaking is increasingly difficult.

Two hours behind the battle for the lead, Brunel remains ahead of Michigan. Michgian made up 13 minutes between Erldunda and Coober Pedy, and another 7 minutes between Coober Pedy and Glendambo, but are still a hair over an hour behind Brunel. Again, barring breakdowns, these two seem fairly fixed in position.

In 5th place is Sonnenwagen, about 45 minutes behind Michigan as of the Glendambo control stop and hours ahead of Tokai. Tokai was briefly in 5th place, having finally passed Sonnenwagen yesterday after a long chase, when they were stopped for about two hours early today due to suspension issues. Tokai fell back to 7th behind Top Dutch before getting back on the road, but they were able to regain 6th place soon after Coober Pedy.

About an hour and a half behind Top Dutch, JU is currently edging out Kogakuin and Durham in a 3-way battle for 8th place. This is a very tightly contested spot to watch tomorrow and into early Friday!

Eclipse is fairly firmly in 11th place with over an hour of space ahead and behind them as of the Erldunda control stop. Behind them are Western Sydney, who are finally making a bit of a run and have opened up a 40 minute gap over αCentauri, Goko, and Blue Sky. These are the last Challenger teams remaining on the course - Chalmers, Adelaide, Halmstead, and Wakayama all had to finally throw in the towel and put the cars on the trailers today.

In the Cruiser class... between the bad sun yesterday, the bad headwinds yesterday and worse ones today, and the fact that Tennant Creek to Coober Pedy was the longest and most challenging leg of their race, none of the Cruisers were able to drive all the way to Coober Pedy. Apollo was forced to drop yesterday, Solaride put it in the trailer early today, and then Ascend... and finally Minnesota and Sunswift threw in the towel at nearly the same time today and loaded the cars onto the trailers in Marla so that they could make it to Coober Pedy before the end of the day.

Now that every single Cruiser team has failed to complete the course, it's a bit unclear what happens now. In one of the various group chats, someone pointed out that a strict reading of the regs gives us a method to determine the winner, but that no one actually gets awarded the cup?


I guess Eindhoven gets to keep it for another two years despite deciding to do some solar offroading in Morocco this year rather than attending WSC?

Anyway, it appears that the winner will be determined by the performance of the cars that finished the Darwin to Tennant Creek leg, plus their practicality scores at the finish line. It's unclear what the Cruiser cars are all going to do for the rest of the event? They have a fresh grid charge and two days to drive down to Adelaide, but none of what they do from here on out counts in any way, so I guess they just chillax and party together for the rest of the drive? I have some thoughts about this competition format - that are basically unchanged since the scoring formula was revised for the 2017 event to use a target arrival time instead of rewarding faster elapsed times - but we'll save that discussion for after the event.

Tuesday, October 24, 2023

WSC 2023: end of Day 3

Hey folks - I'm chilling out at the Marla Traveller's Rest tonight at the conclusion of the third day of WSC 2023, and what a race it's been. For an at-a-glance race summary of the Cruiser Class, check out this chart from Tony at Scientific Gems:

Chart via Scientific Gems
"Two speeds are shown for each team:
the first is the average speed since Darwin;
the second (in brackets) is the average for the last leg."

Other good links: WSC's Tracker Map and a .kml file for Google Earth, a WSC Challenger Graph similar to Tony's (but inverted), and a WSC Cruiser Graph as well.

The first day this year was a great day for solar racing. The second day saw increasing amounts of smoke along the course dimming the skies (a few team members told me they spend the second night sleeping in N95s), and a front rolled north over our position on the second night bringing in headwinds, cooler weather, and patchy clouds for the third day that I am told will persist tomorrow.

I have largely been following along with the teams up at the very front, and the big three up there have been Innoptus (Leuven), Twente, and Brunel (Delft). Innoptus took an early lead on the first day, and between Dunmarra and Tennant Creek they had achieved a >30 minute lead - they departed the Tennant Creek, Barrow Creek, Alice Springs, and Erldunda control stops before Brunel and Twente arrived. They also managed to arrive at Alice Springs just before the end of the second day; the checkpoint there was not scheduled to open until the following morning and the race officials had to scramble to be ready in time for Innoptus's arrival! And speaking of Innoptus, a few people online have been speculating about how much use they are getting out of that retractible fin or if it's mostly just being used for photo shoots. All I can say is that we have overtaken them somewhere between 6 and 8 times, and the fin was extended on all but one occasion...

Further back, Brunel was in second place soon after the start, but began slowing down on the second day of the event and was passed early on by Twente. As we've faced headwinds on the third day of the race, Twente has closed the distance to Innoptus and are almost back to the point they'll be meeting at control stops again - it's very much still anyone's race!

Behind Brunel, Michigan has steadily been catching up. During qualifying, Michigan encountered issues on their hot lap and were forced to start at the very back of the starting grid. I'm told they overtook something like 25 teams on the first day alone... they spend large parts of the first day grouped up with Kogakuin and Top Dutch, and ended the first day at Dunmarra with those two and Tokai. They turned on the gas on the second day and made up quite a bit of time, and although they slowed down in the clouds and headwinds on the third day just like everyone else, they didn't slow down as much as Brunel and have closed from an hour and forty minute gap at Tennant Creek to just under an hour at Erldunda. Barring any mistakes or breakdowns by Brunel, in the distance remaining it will be a challenge for Michigan to close the remaining distance to overtake for 3rd place, but they have a chance...

About half an hour behind Michigan are Tokai in 5th and Sonnenwagen Aachen in 6th.

Tokai has been running a very steady race, they've been comfortably in 5th place for essentially the entire even thus far - momentarily falling to 6th when overtaken by Michigan, but climbing back to 5th after overtaking Sonnenwagen.

Sonnenwagen got the pole position at the start and hung onto 3rd place closely with Innoptus and Brunel up through Katherine, but have slowly and steadily fallen back - by Tennant Creek they had been overtaken by Twente and were in 4th, they had clearly fallen out of the lead group by Alice Springs, they fell to 5th by Erldunda, and were passed by Tokai before the end of the third day. 

Behind Sonnenwagen is a 2+ hour gap back to Top Dutch in 7th, another hour and a half or so back to Kogakuin, and the remaining cars are spread out behind. The current running order from 9th onward appears to be JU, Durham, Eclipse, WSU, αCentauri, Blue Sky, Goko, Chalmers, Adelaide, Halmstead, and Wakayama; all other teams appear to have stuck the car on the trailer. Halmstead and Wakayama do not appear to be on a pace to finish the race; we'll see if Chalmers and Adelaide can hang on. There may be some excitement in the back: WSU, αCentauri, Blue Sky, and Goko have been running extremely tightly together for the entire event and 12th appears to be the most tightly contested place in the Challenger class.

On the Cruiser side, right now it's Sunswift's competition to lose. Cruiser teams were supposed to arrive at Tennant Creek at 2:00pm on the second day, and if late they would start accruing a 0.99^(minutes_late) penalty multiplied against their score. Sunswift was the only team that arrived on time, Minnesota was next at 52 minutes late (penalty multiplier of 0.59), Apollo was 63 minutes late (multiplier of 0.53), Solaride was 65 minutes late (multiplier of 0.52), Ascend was 94 minutes late (multiplier of 0.39), SunShuttle was 128 minutes late (multiplier of 0.28), and the remainder appear to be out of the event. 

Via https://telemetry.worldsolarchallenge.org/contrib/vision/cruiser.html

It's especially heartbreaking for Minnesota, who were on a pace to finish before the target time... until they had some sort of issue and parked their car almost a full hour before the end of the first day, putting them off the back of the pack of Cruiser cars. They were the fastest Cruiser car on the second day, but it wasn't enough to make up for the time they'd lost.

It remains to be seen what the external energy usage and person-km for each team are, and of course practicality is a big black box at the finish line, but Sunswift has been carrying four people in the car for almost the entire even and I expect them to do very well on practicality, so it's hard to see how any other team can come back from those penalty multipliers. Of course Sunswift could also have some sort of issue and accrue a large penalty multiplier, but looking ahead to the second stage, Sunswift is the front-running Cruiser team at the end of the first day of the stage and appears to be on pace to hit Coober Pedy by the required 4:30pm time. Minnesota also remains on pace, while Solaride and Ascend slowly slip back, and Apollo apparently has suffered some sort of critical problem and will certainly miss the 5:00pm hard cutoff, leaving only four cruisers left in the running.

Ok - gotta hit the tent and get some sleep for the drive ahead tomorrow. I'll leave you with some photos:


A very confused road train driver trying to enter the Katherine weighstation

Sunset at Dunmarra


Fire and smoke between Barrow Creek and Ti Tree on the second day
We could feel the thermal radiation passing this fire from inside our car
Innoptus was ahead of us at this point, Twente and Brunel behind


Thursday, October 19, 2023

WSC 2023

 taps mic is this thing on?

Hello! Long time no see. Sorry about the extended absence... I intended to cover the 2019 World Solar Challenge the way I did in 2015 and 2017, but life got in the way. This year I'm headed back to outback... Even though I haven't been blogging in the lead up to the race, I wanted to get some thoughts down before the start.

So for starters: What are the significant changes to the regs since 2019?

1) Ground clearance/approach angles/etc. WSC continues to try to force teams to drive their cars on and off the highway and the start/end of the race day, rather getting lifted onto the highway and sitting stationary in the lane at the start of the day and parking on the highway to get lifted off to the shoulder at the end. Rather than just stating that "cars must be able to drive off a 50mm vertical drop" as in 2019, now we have a minimum 100mm ground clearance, minimum approach and departure angles, a minimum hump that cars must be able to drive over without high-centering, etc.

2) Wheel count. For the first time since 2011, WSC is allowing three-wheeled Challenger class cars (Cruisers still need 4 wheels). On the one hand, all the cars that I ever worked on were three-wheelers and I was disappointed when WSC started to require four wheels in 2013. On the other hand, given how many off-road excursions happened in 2019, I was very surprised to see this change... it's only going to make the modern skinny monohulls even less stable. WSC *has* added a static stability requirement - cars must be able to be tilted 45 degrees across any two adjacent tire contact patches - but I haven't seen any evidence of a tilt table at scrutineering...

3) No more gallium arrays; allowed solar collectors sizes are only called out for Silicon (unchanged at 4sqm). If a team want to propose using something different, "the materials should have low environmental toxicity; this precludes the use of GaAs, CdTe and CuInSe2".

4) Here's an interesting one: For the first time (as far as I can tell), WSC is requiring cars to have automatic battery protection systems, not just battery monitoring systems - peep the language changes in reg 2.5.9. I'm assuming Nuna 10 burning down on the finish line in 2019 might have had something to do with this? I guess it's one thing when non-competitive teams burned down outside of Darwin every other race, but a star team burning down in front of news vans outside Adelaide was different matter.

5) License Plate. No more shenanigans with it being inside the canopy and hard to see, NOTHING can be aft of the license plate within a defined cone. I was interested to see if any teams tried shenanigans here, never underestimate a determined aerodynamicist! But the rule is extremely clear and so far I haven't seen anything questionable. Most teams have a very clear location to mount the plate, as seen for example on the back of Twente's car below:

Photo: Ⓒ Solar Team Twente
(image source)

6) Occupant Space. This is the first year that WSC will be using a test fixture that they call "PVC Pat" to inspect the size of the interior.

7) Rear steer: forbidden this year. No using a rear steering system that only operates at extreme steering lock to meet the U-turn reg, no rear steering system to "crab" the car into crosswinds and reduce aerodynamic drag. I'm told that this is in a bid to reduce car speeds - apparently several teams were sailing VERY effectively in 2019. But notably movable aerodynamic devices are not banned...

8) 3.24.6: "If a solar car rolls onto its side or roof then the team must withdraw immediately." It'll be interesting to see how this reg intersects with the three-wheel reg potentially reducing the stability of the cars. Twente apparently already put the car on its side during pre-race testing.

OK regs talk over, let's talk about cars.

The main way these regs seem to be influencing the teams is: More monohull/bullet cars: fully 3/4 of the Challenger entries this year are monohulls. The modern monohull trend started in 2017: that was the first year of the 4sqm array formula, and we saw highly competitive monohulls finish in 2nd and 4th place overall. This result apparently inspired teams - about half the field was in monohulls in 2019, and the vast majority of the "good" teams were, with 7 out of the top 8 finishers being monohulls. But! Until Twente wrecked out and Vattenfall (formerly Nuon, now Brunel) burned down, the three fastest cars were all catamarans, and the overall winner was a catamaran from Agoria. Given that result, why are teams committing further to the monohull concept (including the winning team from 2019)? I think it's mostly down to the 3-wheel rule.

One of the major disadvantages of the 2017/19 4-wheel monohulls over their catamaran counterparts is that the wider monohull has a much larger surface area close to the road - the entire region bounded between the wheels is basically flat against the ground, in relatively dirty airflow. The body also has to stay wide all the way to the rear wheels and can only start to aerodynamically close behind them. The skinny catamaran pontoons can be much shorter front-to-back, minimizing the parts of the car that are down low. But with only three wheels, a monohull's body can start to sweep inward and upward immediately aft of the front pair of wheels, reducing the area of the car that has to be near the ground. Clearly teams think that this is a big enough change to tip the balance in favor of monohulls... Both Twente and Innoptus (formely Agoria, the 2019 winner) have switched from catamarans to monohulls for 2023.

The big outlier is Brunel (formerly Vattenfall, formerly Nuon): Nuna 12 is a three-wheeled "outrigger" catamaran. It has the driver against one side of the car with a wheel fore and aft of the driver as on previous four-wheel cars, but only a single wheel centrally located on the side opposite the driver. 

Photo: Ⓒ Brunel Solar Team
(image source)

I am genuinely excited about this car, I feel like the sport has been waiting for this design for over a decade. Nuon themselves thought about the idea when the change from laydown cars to upright seating happened with the WSC 2007 regs, you can see an outrigger catamaran concept peeking out in the back of this photo (they ended up choosing the middle concept, setting the tone for the 2007/09/11 generation of cars):

Photo: Ⓒ Brunel Solar Team
(image source)

Kansas State also had the same idea around the same time, and got as far as cutting up an old chassis to build a prototype to do some dynamic testing on. Here they are showing that it brakes straight with the driver's hands off the steering wheel:


And here's a video showing that it skids rather than flips in a high-speed turn:


And prior to the mid-2012 regs release mandating 4-wheel cars, I was seriously proposing to my old team that they field a design like this for WSC 2013 - and I know that I wasn't the only person doing do!

Brunel isn't the only team to look at this design for WSC 2023; Durham is also running an outrigger catamaran this year.

It's worth noting that this isn't Brunel's first rodeo with this car design: they built Nuna 11 as an outrigger catamaran for the canceled 2021 World Solar Challenge, and drove it to a 3rd place finish in Morocco in 2021 and a 1st place finish at Sasol in 2022. The car they are bringing to WSC this year is Nuna 12, their second-generation outrigger catamaran. Brunel clearly believes in the concept.

Of course, the same can be said for Twente and Innoptus - they clearly believe in monohulls now, despite being at the top of the field in 2019 with catamarans. That Moroccan race in 2021 that Brunel finished in 3rd with their first outrigger catamaran? Twente won that race with their first monohull car, and Innoptus came in second with their first monohull. Both of those teams have brought new second-generation monohulls to WSC this year as well.

It's going to be a really fascinating race between these teams this year. We also can't forget Michigan, one of the originators of the modern monohull concept back in 2017 and bringing another good looking monohull. Similar to the teams above, Michigan hasn't been resting over the WSC pandemic break - they also built a non-WSC car over the pandemic (although it didn't race anywhere), so the team has worked to keep their skills fresh.

Photo: Ⓒ University of Michigan Solar Car Team
(image source
)

Worth noting, since you're going to see it on a lot of the monohulls - see the raised nose? It's a very distinctive change from previous monohull designs, and one that is featured on a LOT of the cars this year. I was puzzling over why - maybe there's an aerodynamic reason, but how did every team land on the same aerodynamic feature at the same time? - but then I realized that this is a consequence of the 10° approach angle rule!

This post is getting longer than I planned, so very briefly: All the other teams from the top grouping in 2019 - Tokai, Top Dutch, Sonnenwagen, Kogakuin - are back, with monohulls. Kogakuin of course has some very interesting suspension concepts that I will hopefully be able to get pictures and post about. NITech and Antakari were unable to make it, but Blue Sky, Eclipse, and JU are all back as well. Notably Eclipse is the only WSC 2019 finisher that's coming back with a catamaran, and JU is bringing the only monohull that doesn't have the driver canopy protruding over the solar array.  WSU is also back with a monohull - hopefully they won't have a repeat of their very rough 2019 race.

On the Cruiser side of the competition, the big news is that neither Endhoven nor Bochum are attending this year! Eindhoven won all four previous Cruiser classes 2013-2019, and Bochum has historically been a strong competitor. Without the two of them, Minnesota, Sunswift, and Sophie are likely the three strongest competitors in the class this year, although I'm also keeping my eyes on Apollo. There's also a very interesting rule that the teams must upload telemetry about solar energy collected and energy used from the battery at each of the control stops, I'm very curious to see what that data looks like.

OK - time to stop typing. See ya later, folks.

Saturday, October 14, 2017

WSC 2017 Wrap-up

Hey solar car fans! Sorry the blog has been so quiet for the past two weeks - I've been incredibly busy, and it's been really hard to write this post from behind the Great Firewall of China.

Another WSC is wrapped up, so let's look at how things went down:

Challenger Class

The Challenger Class was really cool this year. Like I've written in other posts, I was extremely pleased when WSC shrank the array while increasing the allowed car size. The combination of those two changes opened up the design space considerably, and resulted in a much more varied field of cars.

I was really surprised that at the top of the field neither monohulls nor catamarans were dominant; the two monohulls from Michigan and Tokai were tightly intermingled with their three main catamaran competitors. It's clear that the teams haven't discovered a single dominant formula under the new regs, and as a result, I think we can expect an even wider variety of cars in 2019. Personally, I expect that monohulls may have an advantage in the future - teams have had 3 generations now to refine catamaran designs, while they have only made a first pass at designing monohulls.

On the array front, I was similarly impressed at how close the performance of silicon and multi-junction cars were. Yes, the three multi-junction arrays finished 1-2-3, but 2nd place was in question up until Glendambo, and 3rd place was a 3-way fight all the way to the finish line. I don't think multi-junction arrays had a clear advantage - the top three cars were from the 1st, 4th, and 5th place teams in 2015. I think it came down more to quality/experience of the teams than array chemistry - if you took a random team and gave them multi-junction cells but no other extra resources, I bet they'd finish in roughly the same position.

Alright, so how did the race itself go?


Well... not great, especially farther back in the pack. Weather played a major roll this year. You can see in the above chart from Scientific Gems that the weather between the Tennant Creek and Kulgera control stops just pounded the teams that were further back along the route. Principia, Goko, NWU, Aachen, ITU, and Kookmin all dropped out of the Challenger class along that portion of the route. Unfortunately, only 12 cars were able to complete the course this year.

Similar to last time, I've tried to compare each team's performance relative to 2015 by using Nuon's time in 2015 as a baseline to normalize their 2017 time against, and then scaling all of the 2017 finishing times accordingly. This is an attempt to account for regulation changes and weather differences - given how the weather disproportionately affected the back of the field this year, perhaps it's not as useful a metric, but there are still some interesting details.

Most teams had a lower performance this year relative to Nuon. WSU and JU were the only two finishing teams that managed to improve, by 1.8% and 1.1%, respectively - so for the second WSC in a row, WSU is the most improved team using this metric! Most of the rest of the finishers had single-digit decreases in performance. NIT was down 1.2%, Michigan and Punch were down 2.1%, Tokai was down 3.8%, Twente was down 6.2% (anything less than a nail-bitingly-close finish would have been a decrease for them), and Blue Sky was down 6.8%. The outlier is Stanford, who were a full 21.1% slower relative to Nuon than in 2015 - they seemed to be disproportionately affected by the weather between the Barrow Creek and Kulgera control stops.

I think I did pretty dang well with my Challenger predictions - I correctly called Nuon, Punch, and Michigan as the top 3, and Tokai and Twente in 4th and 5th! I did miss one Michigan prediction though - they finished in 2nd, which is exactly where I thought they would not be. I have to emphasize again how shocked I am that the monohulls were tightly intermingled with the catamarans; I absolutely did not expect that.

Monohulls in Coober Pedy on Oct 11: Michigan in 2nd, Tokai in 5th.
Only 16 minutes and 30 seconds separated them at this control stop.
Photo: MostDece

As of the Daly Waters control stop, the top 9 teams consisted of my "top five" + all three "challenging teams" + one of two "wild card teams", which I thought was pretty spectacular. The weather after that control stop shifted things around a bit - Stanford and NIT in particular seemed to have a rougher time with the weather. But in the end, out of the 13 teams I mentioned in my prediction post, NWU and ITU were the only two that weren't part of the 12 teams that finished. Antakari (the 10th place finisher) was the only finishing team that I didn't consider in my predictions as a likely top 10 team.

Some specific notes on individual teams:

Nuon: Great car, great strategy, great team. I heard that on the 2nd day of the race, their strategy team commandeered the media crew's satellite link, and their ability to acquire realtime or near-realtime weather data and ingest it into their strategy model was a huge advantage over the other teams at this edition of WSC. Nuna9 was also clearly a full step above the competition in terms of quality. The panel gaps were exemplary; the seam around where the array opens blew my mind every time I looked at the car in their finish line tent. The team really nailed every detail. There's a reason they've won 7 times now - which is half the WSCs ever at this point!

Aauugh so good!
Photo: MostDece

Michigan: The UMich team was finally able to break the 27-year-old 3rd place curse, finishing 2nd this year (or in other words, the top of the "Not Nuon" class)! Novum was a good looking little car and I think there were a lot of great ideas in it, but compared to their other close competitors, it was on the heavier side and there were some parts of the car that looked a little more roughly finished. I think the high performance of the car in spite of the higher weight and finish issues speaks to the aerodynamic advantages of their monohull shape over the catamarans. It also may reflect on the training and discipline of the team - Michigan did a ton of testing on public roads back in the USA and on the Stuart Highway prior to the race; the team seemed to be extremely well prepared and ran a very clean race.

Novum at Port Augusta on Oct 12.
Photo: MostDece

Punch: The team significantly improved their race operations for 2017. They did a much better job of keeping the car running smoothly on the road than in 2015, and this was a large contributor to their strong 3rd place finish this year. There were a lot of clever details in the car (including their steering system, which won the CSIRO Technical Innovation Award), but they were all solidly, functionally clever - nothing that was too clever for it's own good. A great car and a clean race overall.

Punch Two entering the Coober Pedy control stop on Oct 11.
Photo: MostDece

Tokai: Tokai appeared to be the most credible competitor to Nuon over the first half of the race. The two teams were neck-and-neck through Barrow Creek, and I'm not sure what happened to Tokai afterwards - they steadily slipped back over the next two control stops before recovering their previous average speed around Kulgera. I'm not sure if they outran their battery while pacing Nuon, or if they weren't as skillful/lucky as Nuon with the weather. Their inability to point the array as far to the side at control stops certainly cost them some energy... 4 of the 9 control stops occurred significantly far away from solar noon, and Tokai limited themselves to tipping the whole car a little bit rather than lifting and pointing the whole array as the rest of the top teams did. Regardless, they still managed to finish strongly.

Tokai clearly unable to point the array as far as far as Twente and Punch.
Taken at the Glendambo control stop on the morning of Oct 11.
Photo: MostDece

Tokai Challenger after departing Glendambo.
Photo: MostDece

Twente: Twente outperformed my expectations. I half expected them to be bumped out of the top five this time, but instead, the team was on track for a 3rd place finish throughout most of the race. They probably could have pulled it off... but they tried to play cute with a reflective ground sheet and ignored warnings on multiple occasions from officials before being slapped with a 30 minute penalty. You can usually get away with a lot at WSC, but ground sheets have been contentious in the past and WSC had very specifically called them out in the regs this year (reg 3.18.3). C'mon, Twente - you didn't even try to be sneaky (lol at this now-ironic caption), and whining about being caught and punished is poor form. Still, a 5th place finish is a major achievement, especially in a race as rough as this year's.

Red Shift under clouds north of Coober Pedy on the morning of Oct 11.
Photo: MostDece
Red Shift making tracks for the finish line after departing Port Augusta on Oct 12.
Photo: MostDece

WSU: I know I predicted they would do well, but I was still impressed at how well WSU stuck right with the top group for the first third of the race (which apparently caused some consternation among the teams behind them). However, the team made some miss-steps with weather strategy in the middle of the race. I don't think they outran their battery - if they had, they would have ended up with a dead pack under the clouds and would have suffered much more than they did. I suspect they thought the weather was going to be worse/unavoidable and began slowing down to conserve energy - and then were surprised when the weather wasn't quite as bad as they expected and the teams ahead of them were able to dash out ahead of the worst of it. One strategy mistake with a large weather system is all it takes... once they were back under the clouds; there was no way to drive fast enough to get out from under them and the team was forced to slow. The team has very few races under their belt when compared to the teams ahead of them; now they have some hard-earned experience with weather to apply to future races.

Unlimited 2.0 south of Port Augusta on Oct 13.
(image source)

Some other thoughts:

I've heard there was some drama centered around the close time of the Alice Springs control stop as a result of weather; apparently the control stop cutoff was extended, but this wasn't able to be communicated to all of the teams out along the route. I've also heard that there are two teams who definitely trailered forward along the route that are still currently listed as being in the Challenger Class with 3021 solar km credited on the website, and that some teams have filed formal protests over this. We'll have to wait and see how that shakes out; I haven't seen any official results PDF published yet and as far as I know the results are still provisional.

The control stops that I saw seemed to run much more smoothly this year***; the new "no touching the cars while in the control stop" regulation seemed to be a huge success. There were much fewer opportunities for shenanigans, and the stops appeared much less frenetic in general.

The only thing I didn't like was the extra stress put on the drivers... The "driver dash" to the timing tent in socks or bare feet across hot gravel parking lots looked particularly painful. I remember being a solar car driver; I was in no condition to run after climbing out. At least the clouds meant temperatures were low this year; so there weren't any heatstroked drivers passing out and going head first into the dirt/gravel/pavement on their sprint to the timing tent. What little drama I heard about mostly involved the control stop crews not being on-the-ball checking the teams into the stops, sooo... why not just eliminate that and the driver sprint altogether? They allowed the driver to stand next to the car at the end of the control stop rather than sprinting to the car from the timing tent, so perhaps have the team's observer simply note what time the the driver jumps out and sticks their hands in the air? Food for thought.

Punch's driver grimaces as he hops across the sharp gravel in the Alice Springs control stop. The blue shirt control stop crew inexplicably directed him to a parking spot as far away from the timing tent as he could possibly be, despite being the only car at the control stop.
Photo: MostDece

One aspect of the rules that I hadn't thought about prior to the event was that since only the driver is allowed to reconfigure the car upon control stop arrival and departure, the drivers had to load/unload their own ballast and water, connect up their own radios, strap themselves in, etc. I saw a lot of control stop departures delayed due to drivers having issues prepping to leave. I think the "No touching the solar car for the duration of the control stop" regulations contributed much more to the smooth control stop operation than the "only the driver can reconfigure car" regulations; even allowing a single person to help the driver strap in before departure might be an improvement. 5/6-point harnesses are pretty hard to put on by yourself and I'm pretty sure I saw at least one driver take off without bothering to belt in. Nuon had a major advantage here as they used a standard 3-point automotive seat belt, which was much faster to fasten. I'm not sure how I feel about that; automotive seatbelts are designed to work in concert with crumple zones and airbags, and I'm doubtful of how well one would restrain a solar car driver in a rollover. But if WSC keeps these rules for 2019, expect to see a lot more 3-point harnesses in the Challenger cars.

Anyway, like I said above, the control stop regs and operations this year were a huge improvement, so good job WSC. I'm nitpicking because that's just what I do.

***Caveat: I only personally witnessed the race from Alice Springs onward, where there were many fewer cars in a control stop at one time. I heard earlier stops (particularly Katherine) were complete zoos.


Cruiser Class

I was pleased with consistent updates on WSC's website. Official realtime coverage of the race has been lacking in past years, but this year's daily Cruiser summaries were nicely done.

The scoring formula seemed a lot more interesting this year, but I'm still scratching my head at the competition overall. WSC has been encouraging the teams to build ever larger and heavier cars while simultaneously reducing the allowed size of the solar array; the regs (written and unwritten) incentivize teams to build 4+ seat cars that can fit large amounts of cargo inside. Even without the weather, it seemed like this year was going to end up as a battery-electric car competition - most of the cars this year were going to end up with less than a third of their energy coming from the solar array. If WSC keeps following the trajectory that they've been on with the Cruiser class, I'm not sure what the point of the solar array will be. There was that interesting tidbit about having to interact with smart microgrids that was mentioned offhand at the awards... hmmmm.

Alright, time to talk about how the Cruiser competition actually went. Let's start off with some tables - we'll be referring to these below.


Pink background indicates invalidated efficiency scores due to finishing after the cutoff window

Briefly, the point I made in my Cruiser post before the race was right - the teams with more seats had many more strategy options, which was important for the success of both Eindhoven and Bochum.

The teams may have planned to charge off the grid at only a few select locations, but the weather threw a wrench in everyone's plans. In the Cruiser class, every team ended up charging off the grid every single night. Eindhoven was able to use the number of people in the car to modulate their performance - you can see in the efficiency table how they had the car packed with 5 people consistently up through Alice Springs, and then reduced to a single person in the car through Port Augusta in order to keep up the pace through the clouds. Similarly, when Bochum had motor issues, they were able to limp along with fewer people rather than be dead in the water entirely. Meanwhile, the 2-seat cars like Minnesota and IVE didn't have as many options in the cloudy weather and were unable to drop enough weight to keep up.

I tried to hammer this point in 2015, and still was true in 2017: under the Cruiser regs, aero efficiency is still king! Most teams built vaguely "car like" looking vehicles with relatively poor aerodynamic performance; they couldn't compete AT ALL with Eindhoven. The Dutch team's efficiency score was so embarrassingly higher than the rest of the field that Eindhoven could have scored a flat zero on practicality and still won the overall competition by a huge margin. I don't understand why teams keep fielding cars that are so woefully unsuitable for placing well in the competition; it's like teams are designing their cars as if the completion is scored 80/20 practicality/efficiency, rather than the other way around. I simply don't get it. In 2013 no one knew quite what to expect from the Cruiser class, but by 2017 I think it's pretty inexcusable. Eindhoven's victories result in great press releases for the Netherlands and WSC, but it makes for a boring and frustrating competition to spectate when you know exactly who is going to win and why for at least a month in advance.

So, how did I do with my Cruiser predictions? Well, not that great. I correctly called Eindhoven as the winner (but that was frankly a no-brainer) and Bochum as the second place team. After that, however, I didn't do so well - PrISUm, UNSW, and Lodz all were forced onto the trailer fairly early in the competition. PrISUm had battery issues on the first two days that left them running on empty when the clouds and rain hit on the third day, forcing them to trailer the car before Barrow Creek. UNSW withdrew somewhere around Tennant Creek or Barrow Creek due to safety concerns over their suspension issues. I'm unsure about what happened to Lodz, but given the small size of their battery (second smallest in the Cruiser class), I expect that they were unable to collect enough energy to keep up under the clouds. My consolation prize is that the three 2-seat cars I mentioned further down in my predictions ended up being the three remaining contenders. Arrow was the least efficient of the three cars, but their larger battery allowed them to pull more energy from the grid and keep their speed up through the clouds and rain. Minnesota had the smallest battery in the Cruiser class and IVE had a small battery as well; neither were able to maintain the minimum Cruiser pace despite charging every night and driving with a single person in the car. The past couple of WSC events have been pretty dry overall, so I guess some of the Cruiser teams forgot that sometimes WSC is cloudy and rainy...

My notes about the "cheater" Cruiser being the winning strategy were probably right. A team with performance comparable to WSU could have slapped a second canopy and an extra 1sqm of array on their car; with their standard-size Challenger battery and a single grid recharge, they would have posted an efficiency score so high that they could receive a flat zero on Practicality and still beat Eindhoven. Sure, Einghoven was so far ahead of everyone else that they probably ran a pretty conservative strategy over the later half of course, and could have posted a higher efficiency score if they were feeling pressure... But I still remain convinced that minimizing energy usage with a no-holds-barred aerodynamic vehicle could have yielded an extremely competitive entry.

Eindhoven: As in previous years, Eindhoven focused on packing a large number of people into the most aerodynamic shape possible, rather than building a vehicle that resembles a normal car. The efficiency of Stella Vie in terms of person-km/energy used was almost embarrassingly better than the other teams - more than double the efficiency of the next two cars, and more than triple the efficiency of the two cars after that. Not only did they tie for the 3rd smallest battery in the Cruiser class, but they were able to use it to haul an average of 3.38 people in the car across 3021 km. The three other cars with sub-10kWh batteries weren't able to keep up, even with only a single person in the car.

Eindhoven will simply remain untouchable unless other teams start focusing on aerodynamic performance. No matter how the Cruiser regs evolve, I find it hard to imagine a credible event that doesn't place a majority of the score on vehicle efficiency in some way - and as far as cars are concerned, efficiency = aerodynamics. Period.

But that's not to say that Eindhoven neglected the practicality portion of the competition, far from it: they also scored the highest on practicality. They received the highest score from the judges, and were the only team other than Bochum to ace the exercises. WSC 2017 was Eindhoven's most impressive performance to date.

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

Bochum: The Germans were the most credible competitors for Eindhoven this year, but they were plagued by motor controller issues. They went through a total of five WaveSculptor22 motor controllers over the course of the event. They were only able to carry a single occupant over the first leg of the race due to motor controller issues, and were the last Cruiser car to make it to Katherine. Bochum were able to carry a full four people over the Daly Waters to Tennant Creek leg, but the car refused to drive out of Tennant Creek until they removed half the people from the car. It was a similar story in Alice Springs - they tried to depart with two people, but the car just didn't want to go until they removed some weight. The car seemed to run better over the second half of the race, but they still burned out another motor controller when they arrived at the parade grounds in Adelaide.

Bochum's post-race WaveSculptor graveyard
Photo: MostDece

Bochum's car also had some surprisingly crude aerodynamic modifications on the rear end. I'd seen these in photos after their wind tunnel trip, but I assumed that they'd have a more refined version installed for the race...

Crude rear wheel strakes on Blue.Cruiser
Photo: MostDece

Duckbill spoiler
Photo: MostDece

Bochum's interior was extremely nice, containing a lot of sustainable and recyclable materials - sadly I didn't manage to get any good photos. The team scored very well in the Practicality portion of the event, getting the 2nd highest Judge's score overall (including the highest score in the "Desireability" section) and acing the exercises.

Arrow: Arrow's car was a little rough around the edges, but it had it where it counted: the ability to actually drive. The team made steady progress through the outback, which was aided by the second largest battery in the Cruiser class this year. I doubt the team was ever short on energy, but their large battery also resulted in the lowest efficiency score of the five teams that drove the entire distance. As it turned out, that didn't matter much this year - so few cars finished that simply completing the course was good enough to get them onto the podium. They also scored very well (4th) in the Judge's portion of Practicality, but did abysmally badly in the Exercise portion. They only received 3 out of a possible 9 points there, resulting in the worse exercise score. Apparently they had difficulties with the 3-point turn? Not sure what went down there.

There were a lot of neat features - particularly, their gorilla glass encapsulated array, which they delighted in doing things to that you would never, ever do to a normal solar car array.

Arrow STF heading south out of Alice Springs on Oct 10
Photo: MostDece

Some gripes:

I didn't like the sharp time cutoff in the Cruiser regs this year - it basically meant that if teams had a single issue and fell off pace, they were out of the race. There were effectively zero ways to mitigate risk without drastically reducing the efficiency score. Yes, there's a sharp cutoff at the end for Challenger class as well, but Challenger cars are incentivized to be as fast as possible - as far away from the cutoff as possible. Because slower speeds are more energy efficient, the Cruisers are incentivized to be as slow and as close to the cutoff as possible, which basically encourages bad outcomes. It's a recipe for last-minute problems completely sinking a team. For example, if Eindhoven had experienced their midconsole structural failure at the start of the 6th day instead of the end of the 5th day, they might have simply been dead in the water.

This year, only 3 of the 13 cars managed to finish the course before the cutoff. IVE was just barely on-pace to finish, but got stuck in unexpected (to them) Adelaide traffic and was a bare 18 minutes late to the marshaling point. Minnesota was just a bit too slow over the entire route and arrived 54 minutes late. If those two teams had arrived on time, Minnesota and IVE would have been the 3rd and 4th place Cruiser teams overall. But because they missed they arrival cutoff, WSC has given them a score of zero for efficiency - in other words, they aren't credited at all for any distance driven. Strictly speaking, this follows the letter of the regulations... But by not recognizing Minnesota's and IVE's performance on the road in any way, WSC has bumped those two teams back to 8th and 11th place, respectively - far behind teams that completed less than a third of the route (and in some cases, none of the route). I would have understood if WSC had penalized the teams such that they placed behind Arrow - the last team to complete the course on time - but the current results seem extremely harsh for IVE and Minnesota. Perhaps for 2019 WSC should implement a target time rather than a cutoff time, with a sliding scale of penalties for missing the target.

Practicality scoring seemed much more haphazard than last year, which was surprising to me given the increased emphasis on practicality in this year's scoring equation. Instead of a big show in the center of the square, it was crammed off in a back corner. I was talking to a Challenger team and keeping my eye on the middle of the square while waiting for it to start, and completely missed the start of the judging as a result. This may have been due to the large volume of Challenger cars slowed by the weather and crossing the finish later, but parts of the judging still seemed like it was organized at the last minute and being figured out on the fly, rather than planned out ahead of time.

The teams were allowed a spotter to direct the driver through the parallel parking test, so I'm not quite sure what was the point of that test. Didn't WSC already test the turning radius of the cars up in Darwin? If they're not testing the car's rear and side vision by allowing a spotter, isn't this basically a test of the driver (and the spotter), and not of the car itself? And these two driver skill exercises accounted for a full 1/3rd of the practicality score!

"Storage" wasn't a varied set of items in a range of sizes like in 2015 - it was a stroller, a baby car seat, and an entire city bicycle that appeared to have been grabbed right off the street that afternoon. The baby car seat seemed to be picked to specifically penalize teams that used racing bucket seats, for some reason that eludes me. The bicycle had nutted axles instead of a QR, so the front wheel could not be removed, and WSC forbade teams from dropping the seat or rotating the handlebars... I haven't owned a single car in my life that I could have fit that bike inside! I've always needed to be able to remove the front wheel to fit a bike inside anything smaller than a minivan or SUV. Is WSC accusing my trusty old Camry of being an impractical car? It had a positively cavernous trunk, but without fold down rear seats I'm pretty sure that I couldn't have fit a bicycle inside even with both wheels removed...

Teams also didn't appear to be scored on how much time it took to pack the items inside. Iowa simply opened the rear hatch and chucked the bike inside the large cargo area, while some teams used several team members and many tens of minutes to carefully ease the bike through the roof into the passenger compartment - including requiring the driver to contort around parts of the bicycle intruding into their seating area. I could have taken Iowa's car and driven all the way from my house to my favorite mountain bike trailhead before Bochum finished packing the bicycle in their car, but WSC gave both teams the exact same score for this exercise. Madness.

Bochum laboriously, carefully filling their passenger compartment with a bicycle.
(image source)

A WSC official explained to me that they were trying to pick items for the storage judging that would differentiate the cars rather than award participation points to everyone, but the limited selection of large items really dismayed the teams that had attempted to design efficient commuter and around-the-town cars with lots of easy and convenient storage for daily items like groceries, but no bulk storage.

WSC seems really enamored with "family cars" like Eindhoven's entries over the past 3 years (perhaps not without reason, given the team's performance), and I'm not sure why they even bother to allow 2-seat cars in the Cruiser class anymore. Not only did this year's Efficiency scoring equation favor cars with more seats, but the Practicality scoring seemed to heavily favor the cars with more seats as well - the 4- and 5-seat cars uniformly placed ahead of the 2-seat cars in Practicality judging. A team relayed to me over beers that they were explicitly told at Practicality judging that points were being taken off for their lack of a rear row of seats. The regs state that teams will be judged on "suitability for the declared purpose" (4.4.13), but instead, every team was judged on how well they had built a family minivan. If WSC wants to see 4-seat cars, just get it over with - require 4 seats!

Personally, this sort of scope creep is really frustrating to me in the real automotive world as well - no one sells any genuinely compact trucks in the USA anymore, for example. I neither want nor need a giant Family Truckster, but it's becoming increasingly hard to buy simple and small cars that aren't painfully decontented and boring econoboxes. My daily-driver for the past 7 years has been a 2-seat car; it doesn't matter because the vast majority of the time I am the sole occupant. It's super easy to park and it gets great fuel mileage. I have a hitch rack for two bikes, so I can easily load it up for a weekend adventure with a friend. It's been to both the Atlantic and Pacific Oceans. But since I can't fit a stroller in the trunk or an entire bicycle inside the car, WSC would tell me it's "not practical"...

Ultimately, all of the practicality shenanigans and gripes didn't matter - the few cars that finished had such wildly disparate energy efficiency scores that practicality judging was a sideshow rather than anything that affected the finishing order.

Sorry for ending this post on an extended gripe - there was a lot of cool stuff in the Cruiser class this year, but there are just a lot of frustrating aspects as well.


Thanks for all of the fun memories, and I'll see you all in Darwin in 2019!