Tuesday, September 22, 2015

Cruiser Class Part 1: Background, and 2013 Reflections

Today, we're going to take a deep dive into the Cruiser Class, starting with a look back at 2013.

In addition to the Challenger class, the little single-seat race cars we all know and love, WSC announced a second class for 2013 - the Cruiser class. For the public perspective, the big news was that these cars would have more than a single seat (reg 2.25 specified "two or more seats", but interestingly, did not specify a maximum number of occupants). For the teams, however, the big news was that the way the winner would be determined was radically different.

With very few exceptions, solar car events are *races* - the fastest car to complete a fixed distance wins (a small number of events - mostly on closed course tracks - are over a fixed length of time, and the car that goes the furthest wins). The Cruiser class, however, would be scored on several aspects of performance, of which speed was only one portion. The four categories were:
  • Energy usage. In addition to 3x as large batteries as the Challenger class (up to 66kg of LiPo), Cruiser teams were able to charge off the grid at three locations - Tennant Creek, Alice Springs, and Coober Pedy. Each time the teams connected their cars to the grid, it would be assumed they had taken a full battery's worth of charge - WSC did not actually measure how much energy was taken off the grid.
  • Elapsed time. (Hopefully self explanatory)
  • Person-kilometers. The number of people in the solar car for each leg multiplied by the distance of each leg. Person-km only count for seats that are continuously occupied between control points.
  • Practicality. The only information given was "will be judged by an Expert Panel" and this category was otherwise a complete black box.
Energy usage was worth 1 point, elapsed time was worth 3 points, person-km was worth 0.3 points, and practicality was worth 1 point, for a total of 5.3 points. Percentage-wise:
  • 18.87% for energy usage
  • 56.60% for elapsed time
  • 5.66% for person-km
  • 18.87% for practicality
To determine a team's score in each category, their performance would be normalized relative to the highest performing team in that category - so a team's score partly relied on the actions of *other* teams, certainly an interesting wrinkle. Folks who are interested in reading the precise wording of the scoring rules should look at reg 6.95 from the 2013 WSC regulations.

When teams were discussing which class they wanted to design for, a few common objections to the Cruisers came up again and again. For starters, the size of the battery - and particularly the grid charging portion - didn't sit very well with some teams. If a team used the largest battery allowed and charged at all three grid charge locations, well over half of their energy (perhaps even over 2/3rds) would have come off the grid. That doesn't make for much of a solar car race.

The other common (and larger) complaint was the "practicality" portion of the score. It was such a large percent of the score, and teams were given absolutely no information about what WSC was looking for or how it would be judged. Given that most design features to enhance "practicality" would compromise the performance of the car, in order to win, it would be crucial to get the proper balance of performance and practicality. But there simply was not enough information given to make even the most basic design decisions. Unless you had a skilled telepath on your team, every design tradeoff would be a wild stab in the dark. Several teams were not willing to risk a 18 months their of time, effort, and money based on blind guesses about what the judges would want to see.

Many other design competitions use a computed score to determine a winner (The competitions in the SAE Collegiate Design Series spring to mind, as does the Solar Decathlon), but the rules and regulations always very clearly spell out precisely how the competitions will be judged and scored. In comparison, literally only giving a single sentence description of a factor worth 18.87% of a team's score - "Practicality will be judged by an Expert Panel" - was a sick joke.

Prior to the race, I did some rough performance calculations anyway, and decided to just ignore practicality altogether. After running through some sims, it became pretty clear that the highest performance designs would be aerodynamic two-seaters of some sort. The winning race strategy would be to drive with two people in the car at all times - the balance between scoring for person-km and elapsed time made the extra weight worth it - and to only use the middle charging location in Alice Springs. A maximum efficiency, no passengers, no-grid-charging run was plausible as a winning strategy, but allowed little to no margin for error, and using all three charging locations would cost more points than would be gained by shorter elapsed time. 3-seaters and more would simply be too un-aerodynamic to keep up without charging at all three charging locations, and the penalty for increased energy usage would more than negate the extra points in person-km. I figured (guessed) that the delta in practicality between +3-seaters and 2-seaters would not be big enough to make up for the performance difference.

Ultimately, eleven teams entered the Cruiser Class in 2013, and eight competed.

First, let's look at one of the teams that didn't make it to WSC:

University of Tehran

PGIII render from March 2013
(image source)
The University of Tehran first competed at the World Solar Rally in Taiwan in 2006, and brought Persian Gazelle II to WSC in 2011, finishing 31st out of 37 with about a third of the course completed. The team was constructing Persian Gazelle III to compete in the Cruiser class at WSC 2013.

The car appeared to be a highly-aerodynamic two seat car, with the occupants seated in-line. The driver bubble was so short that I assumed the rear passenger would be straddling the driver, bobsled style. This is almost exactly how I would design an upright-seating two seat solar car if practicality were not a concern. It's also how I would expect a team to design a car if they wanted to also be competitive at events that did not have a two-seat class.

In any event, the renderings that they posted to Facebook in March were the last we saw of them. They withdrew from the event without public explanation about a month before the event.

The team has since finished the car, and entered it into the WSC 2015 Cruiser class. I'm very excited to see the car in Australia.

Here are the teams that actually showed up in 2013:


PowerCore SunCruiser (image source)
Bochum had been building practical-appearing solar cars for several years, culminating in the two-seat SolarWorld GT at WSC in 2011, despite the fact that no two-seat class existed at the time. In 2013, they brought PowerCore SunCruiser, a 3-seat car. The driver was front and center, with passengers behind and to either side. 

Like many of the past Bochum cars, it had a 3sqm gallium-arsenide (GaAs) solar array. Given that silicon solar cell efficiency is creeping up to 25%, and GaAs cells top out around 30%, most teams go with the 6sqm silicon arrays. But on a multi-seat car, a smaller GaAs array allows much more flexibility with big windshields, doors, and curvier surfaces. Many Cruiser teams that went with silicon arrays were not able to get the full 6sqm on the exterior of the car, and had to pack some of the array onto auxiliary panels - packed inside the car while driving, and only utilized while statically charging.

It was quite a small car - only about 4.2m long and 1.6m wide.

Solar Team Eindhoven

Stella (image source)
Solar Team Eindhoven was a brand new team hailing from the Netherlands, the same country as Nuon and Twente. They were clearly shooting for maximizing practicality - they brought a 4-seat car named Stella, with a reasonably complete looking interior. The long upward-opening doors were a little janky, though - definitely awkward to get in and out of, especially the front seats.

The large flat top meant they were the only(?) Cruiser team to be able to fit a full 6sqm of cells onto the top surface of the car. They charged by rolling the car up onto ramps and popping up the rear hatch to normalize further.


eVe (image source)
UNSW's teasers prior to the event claimed they were building a "solar supercar", and the car certainly looked the part - if I had to pick a single one of these to show off in my driveway, this would be it. Their sporty side-by-side two-seater, eVe, utilized a tunnel underneath the car to reduce the aerodynamic drag. This is the sort of shape of car I was expecting most teams would bring - one that looks kind of like a normal car, with fancy aero on the underbody. However, UNSW compromised the performance of the car too much in the name of style for me to take it very seriously. Those fender flairs with the little 3-cell modules on the top of each were hilarious. The panel closest to the camera in this photo is all cells that were carried inside the car while driving; they probably could have put half of those cells on the top of the car if they'd made the fenders flush with the array, rather than protruding.

University of Minnesota

Daedalus (image source)
The University of Minnesota came back to WSC for the first time since 1999 with a very aerodynamic appearing two-seat Cruiser car called Daedalus. The two occupants were very cozy, and the seats were accessed by sliding the whole hatch forward. They were unable to fit the full array on the exterior of the car, and utilized two small auxiliary arrays in the morning and evening.

This car looked like it struck a very good balance between maximizing performance while slightly hedging bets against the practicality judging.

University of Calgary

Schulich Delta (image source)
The University of Calgary entered their two-seat Schulich Delta into the Cruiser Class. Notably, they placed some solar cells on the sides of the car, rather than carrying an auxiliary array inside the car.

TAFE SA Solar Spirit

Solar Spirit (image source)
The team from Technical and Further Education South Australia brought their car Solar Spirit. It was a two seat car, with the driver compartment accessed through a forward tilting hatch. The team had previously raced this car at WSC 2011, only completing 801km of the course.

Goko Industrial High School

Kaiton II (image source)
The team from Goko entered their car Kaiton II. Looking back at photos, I honestly can't recall wether it was a two seat or four seat car. Ton Boon has kindly pointed out that it was a two seat car.

Kaohsiung University of Applied Sciences

Apollo VII (image source)
The team from Taiwan brought Apollo VII, one of the strangest looking cars in the event. A lot of us were sure the rendering that was posted to their team profile page had to be a joke, but the car looked exactly like the render. When they crossed the line at the end of the race, the big wide wing was gone and I'm not sure what happened to it.


Out of the eight cars that competed, there was a single four-seat car, a single three-seat car, and six two-seat cars. Only two of the cars appeared to emphasize aerodynamics in any major way - most were very high frontal area "solar bus" or "normal passenger car" style vehicles. It's clear that most teams that did enter the event had decided to emphasize practicality over performance, rather than design to win the scoring equation.


Only four teams in the 2013 Cruiser class managed to finish the race: Eindhoven, Bochum, UNSW, and Minnesota. Final results for those four teams were as follows:

  • Eindhoven: 1st place, 97.5% overall
    • Elapsed Time: 40 hours, 14 minutes
    • Person-km: 9093 km, avg 3.01 people in the car
    • External Energy Usage: 64.0 kWh
    • Practicality: 88.3%
  • Bochum: 2nd place, 93.95% overall
    • Elapsed Time: 41 hours, 38 minutes
    • Person-km: 6484 km, avg 2.15 people in the car
    • External Energy Usage: 63.5 kWh
    • Practicality: 87.0%
  • UNSW: 3rd place. 92.18% overall
    • Elapsed Time: 38 hours, 35 minutes
    • Person-km: 3022 km, avg 1 person in the car
    • External Energy Usage: 64.0 kWh
    • Practicality: 70.7%
  • Minnesota: 4th place, 79.18% overall
    • Elapsed Time: 51 hours, 41 minutes
    • Person-km: 5454 km, avg 1.8 people in the car
    • External Energy Usage: 64.0 kWh
    • Practicality: 69.3%

It's hard to draw too many conclusions about the various car designs due to the small sample size, and extenuating circumstances that affected each team. Minnesota was crippled by malfunctioning motors and barely managed to finish the race at all (sacrificing many power FETs in their motor controllers in the process). UNSW's car was finished at the last moment and had host of minor electrical bugs and other issues. Notably, the passenger door would not latch, so they had to tape the door shut. It is expressly against the rules to tape an occupant exit hatch shut, so they were not allowed to put a person in the passenger seat. They had to drive with only a single person in the car for the entire race.

Interestingly, the practicality scoring was fairly tight. Eindhoven scored the highest with 88.3%, and Minnesota scored the lowest of the finishers with 69.3% (KUAS was the only team to score lower, with 50.3%). But remember, scores are normalized relative to the highest scoring team in that category! So Eindhoven effectively got 100% for the category, and Minnesota scored 78.5% relative to Eindhoven. This is not a huge spread, considering the visibly apparent differences in the practicality of their two cars! Reports that I read said that the "expert panel" consisted of several journalists for car magazines, among others. Jeroen reported that the appeared to be judging the cars on cargo space, the number of seats, and parallel parking, among other things.

Knowing what we now know about how the practicality scoring went down, let's speculate a bit. Playing with the numbers, it looks like if the more aerodynamic cars - Daedalus and eVe - had worked properly, one of them probably could have won it. If either car had worked right, they *should* have been able to only use one of the three charging locations - which would have put a huge damper on the score of Stella and SunCruiser. To win using one charging location and two people in the car at all times, they would have had to average about 70kph - a little slower than Bochum, but a little faster than Team Arrow, the 7th place Challenger team. Given that they would have had 6x the amount of energy charged off the grid than Arrow did, that certainly should have been achievable for slippery two-seaters.

Ultimately, the 2013 Cruiser Class came down to team and car preparation more than it came down to the design of the cars. We'll see what lessons the teams, new and old, learned when they compete this year.

For 2015, the Cruiser Class is largely the same, although WSC has slightly addressed the two major pain point that we discussed earlier in this post. They have reduced the external grid charging locations to only one, in Alice Springs; this means at least half of a team's energy will have to come from the solar array. The cars will be scored in the same four categories, although the balance is different - Energy is now 15% (down from 18.87%), elapsed time is 70% (up from 56.60%), person-km is 5% (down from 5.66%), and practicality is 10% (down from 18.87%). They've also come to their senses and clarified practicality at least a little in the rules - although they don't commit to anything, they mention that the judges may be looking for "...ease of access and egress, ease of operation, reliability, versatility, style, suitability for declared purpose...". They outright state that being registered as a street legal vehicle in the team's home country will be "highly regarded".

We'll discuss the 2015 Cruiser field in depth in another post.

1 comment:

  1. At the time I did this chart of the results for the top four teams. For each team, the four bars on the right add up to give the total-score bar on the left.