Thursday, September 7, 2017

WSC 2017 Challenger Summary (Part 1)

Let's do a summary of the 2017 Challenger Class! First, a quick primer on historical solar car shapes.


There are innumerable factors to account for in solar car design, but the one that most constrains and defines the shape of the car is how to package the driver and wheels in the most aerodynamic way possible. Before the mid-2000s, there weren't a whole lot of rules constraining this (other than "no head-first driver seating"). The favored way to place the driver was lying on their back, with their head poking through the top of the array just high enough to meet driver vision rules. As no part of the driver stuck out below the car, driver position relative to the wheels was mostly unconstrained.

Cars like Nuon's Nuna3, Michigan's Momentum, Ashiya's Sky Ace Tiga, MIT's Tesseract, and Minnesota's Borealis III represented the best of the last evolution of the laydown cars.

Michigan's Momentum at NASC 2005

In 2007, the World Solar Challenge changed things up significantly: They required the drivers to sit upright, which suddenly put large new aerodynamics constraints on the car. As a result, there was a lot of experimentation in 2007. Nuon's winning formula that year was to place the driver centered vertically in the array airfoil, with their head sticking out the top and their rear hanging out below, and their legs lifted up into the array to minimize the size of the lower protrusion. The driver was shifted to the rear of the car so that the single rear wheel could be aerodynamically masked by the driver, and for center-of-gravity reasons, this requires the battery to be shifted very far forward - which results in a very long nose sticking forward of the front wheels. This style of car became ubiquitous in 2009 and 2011 - the years of Tokai's two victories (and the only races after 2001 that were not won by Nuon).

The 2011 Tokai Challenger on the way to winning WSC
(image source)

In 2013, the rules changed to require 4-wheeled cars, and required that the driver's heels be the lowest part of their body - no more lifting the feet up into the solar array. Some teams tried to adapt their 2009-2011 designs to the new rules, but the rear wheels couldn't be masked by the lower portion of the driver, and the new seating position made the lower driver protrusion much larger and longer. The top two finishers (Nuon and Tokai) did something different: they placed the driver far offset to one side, within a long monolithic fairing containing a front and rear wheel. By combining three protruding objects into one smooth fairing, frontal area and aerodynamic drag were minimized. In 2015, fully two thirds of the field built these "asymmetric catamarans".

Nuna 8 at WSC 2015
Photo Credit: MostDece

(Aside, an old-man grumble: I've seem a few folks going "Wow! So many cars that aren't doing the traditional catamaran design this year!" A design that was introduced two races ago and only became dominant at the last event hardly seems traditional to me. The only design tradition at WSC is that everything always changes)

In the background of all of these driver and wheel layout changes, the array shape has been stuck in a rectangular planform for a loooong time. From before 2001 through 2005, teams were pretty much forced to fill the entire 5m x 1.8m planform in order to fit the entire array. Teams were given a teensy bit of room to experiment  with slightly rounded shapes from 2007-2011 when the array was shrunk to 6sqm, but designs were jammed back into the rectangle in 2013 when the maximum length shrank to 4.5m.

The 2017 Challenger Field

Alright, now that we've done a rundown of the past decade of WSC solar car design, let's look at this year's field. The big rule changes for 2016 are:
  • Smaller array: 4sqm of silicon cells instead of 6sqm.
  • Proportionally more multi-junction array allowed: The old rules allowed 3sqm of multi-junction (50% the allowed area of silicon), but the new rules allow 2.64sqm (66% the allowed area of silicon). This makes multi-junction arrays a competitive option again for the first time since 2007.
  • In contrast with the smaller array, the 2017 regs specify a larger car bounding box: Up to 5m long x 2.2m wide, vs 4.5m x 1.8m in 2013 and 2015.
By the numbers, the field doesn't have much more variety than in 2015 - back then, 19/30 cars were asymmetric catamarans (that post includes Antakari, who did not make it to WSC), compared to 19/26 cars doing catamaran cars this year. However, in 2015, the top end of the field was homogenous - only the 8th-place and 14th-place finishers out of the top 15 were not catamarans. This year, the variation is all at the top of the field: two of the top five from 2015 are doing narrow monohull cars, and three are doing multi-junction arrays. Only one of the top five from last WSC is building a silicon-array catamaran that resembles the dominant 2015 formula.

This year, I'd organize the cars into three general groups:

1) Asymmetric catamarans. 19 out of 26 challenger cars are in this category.
  • One (Durham) is a returning car from 2015 - the rest are built new for 2017
  • Only Goko and Durham are 3-fairing designs, with individual front and rear wheel fairings on the side opposite the driver.
  • Only Punch and Nuon aren't using silicon arrays: Both are using 2.64sqm multijunction arrays.

2) Monohull designs. Common features of this group: They're very long, and they're as narrow as possible. The wheels are contained within the body and the array is largely aft of the driver. Four cars are in this category: Michigan, Tokai, Kogakuin, and CUER. Michigan is the third car that we know is using a smaller multijunction array, and presumably CUER is as well.

I expect someone will try to credit CUER with inspiring these designs, but the other three team's basic design strategy is fundamentally different from CUER's concept. CUER's 2013 concept involved sacrificing power generation while driving for the sake of aerodynamic efficiency, and making up for it with a huge concentrator array while statically charging. It was a clever idea that might have been highly competitive under the 2013 regs (if the team could have kept the rubber on the road), but became a much less viable concept as WSC reduced the potential advantages of concentrators in 2015 and again in 2017.

However, the array sizes are small enough this year that the other three monohulls aren't making an aerodyanamics-vs-array-power tradeoff - they fit the entire allowed array, and they do it directly on the exterior of the car - not under a hot, reflecting dome. They're not making an array tradeoff vs the catamaran cars; it's just straight up a question of if their monohull designs are more aerodynamic than the competition or not. This is a much more compelling concept than CUER's, and one that has only been made possible by the 2017 regulations.

3) Unique layouts. There are three teams doing one-off designs that don't fit into a clean category with any of the other cars.
  • MDH - The driver and wheels are laid out like a 2013-2015 five-fairing car, but with a twist: there's nothing on either side of the driver pod. The array is split into two halves, entirely ahead and behind the driver. I'd include it with the five-fairing cars, except there aren't any others this year! The five-fairing car has gone extinct in 2017.
  • NWU - It's sort of monohull-ish with a minimal, aerodynamic driver pod, but a has lot of different features that I think puts it into it's own category: The wheels are splayed wide on outriggers rather than contained within the driver pod, and the array section is short and wide rather than long and narrow. Finally, the array not tightly integrated with the driver pod; it's clearly a separate body. You could say that this is true of Kogakuin as well, but I'd argue that Kogakuin is much more visually similar to Michigan and Tokai than NWU. The fact that Kogakuin's array is elevated off the body doesn't change the fact that the car's profile is long and narrow, whereas NWU is the widest car at the event. Sometimes you just have to make judgement calls with solar car taxonomy...
  • ANU - This car is just kind strange. The latest CAD they teased us with is closer to a normal asymmetric catamaran than the earlier screenshot, but it still is a 4-fairing car with a split driver-side fairing in addition to the individual wheel fairings on the opposite side. Definitely haven't seen that before.

Driver Side

Digging deeper into Challenger field, first let's look at which side the teams put the driver on (compare to 2015).

Out of 26 challenger cars, 20 are asymmetric:
  • 17 have monthlithic fairings on both sides. Goko and Durham split non-driver fairings; ANU has split fairings on BOTH sides.
  • 15 are driver on the right, 5 on the left
  • 9 have the exact same layout as Nuon's winning 2013 and 2015 cars: monolithic fairings, driver on the right. Twente has switched from 3-fairing to 2-fairing, now mimicking Nuon's layout.
Some more comparisons on driver side:
  • Of the 12 teams from countries where normal cars have the driver on the left, 1 team built a solar car that kept the driver on the left (Blue Sky), and 11 teams switched to the right side.
  • Of the 8 teams from countries where normal cars have the driver on the right, 4 teams built solar cars that kept the driver on the right, and 4 teams switched to the left (Adelaide U, UiTM, Durham, and Nagoya).
Of the left-hand-drive teams, Blue Sky, Adelaide, and UiTM all did it in 2015 as well. Durham is bringing their old car, so they don't have a choice, and Nagoya had the driver in the center last time.

So like in 2015, there's a strong trend among the asymmetric cars to put the driver on the right, and it's not just an artifact of what country the teams come from - teams from countries with left-hand-drive cars overwhelming switched to put the driver on the right side. Also like in 2015, 2-fairing designs are much preferred over 3-fairing designs among the asymmetric cars.

Car Dimensions and Proportions

An area that we see a lot of variation this year that we haven't in the past is the shape of the array and the general proportions of the car. It used to be taken as a given that the car design would end up within a centimeter or two of the maximum length and width, but the smaller arrays + the larger bounding box blow the design space wide open. It shows in the 2017 field: we have narrow cars, wide cars, long cars, short cars, rectangular cars, curvy surfboard-shaped cars... and whatever wins this year, there are just so many different things you can do with this set of rules that I doubt the cars will converge as quickly on a winning formula as they did in 2009 and 2015. See below for a table of all of the car dimensions that I have:

Team Layout Array Length (m) Width (m) Aspect Ratio
3: Nuon Catamaran GaAs 3.27 1.55 2.11
32: Principia Catamaran Si 3.38 1.60 2.11
8: Punch Catamaran GaAs 3.5 1.6 2.19
71: ITU Catamaran Si 4.0 1.8 2.22
7: Adelaide U Catamaran Si 4.1 1.7 2.41
38: NWU Outrigger Si 4.98 2.05 2.43
16: Stanford Catamaran Si 3.91 1.57 2.49
22: MDH 5-Fairing Si 4.5 1.8 2.50
70: Sonnenwagen Catamaran Si 4.3 1.7 2.53
46: JU Catamaran Si 4.28 1.63 2.63
21: Twente Catamaran Si 4.27 1.38 3.09
15: WSU Catamaran Si 4.58 1.40 3.27
25: NIT Catamaran Si 4.98 1.49 3.34
10: Tokai Monohull Si 4.98 1.20 4.15
88: Kogakuin Monohull Si 4.98 1.05 4.75
2: Michigan Monohull GaAs 5 1 5.00

A typical car from 2013 or 2015 (constrained to 4.5m x 1.8m) would have an aspect ratio of around 2.5, whereas an older car constrained to the 5m x 1.8m planform would have an aspect ratio closer to 2.78. A lot of the catamaran cars are fairly close to these values, but there are some significant outliers. NIT, WSU, and Twente are all building cars that are proportionally longer and narrower, whereas ITU, Punch, Principia, and Nuon are building cars that are proportionally shorter and wider.

In terms of absolute size, note that despite being diminutive in length, the two GaAs catamarans are actually wider than the long & narrow Si catamarans. It's interesting to see where the different designs diverge, and where they cluster...

Array Shape and Placement

There's another place the narrower catamarans stand out this year: front nose overhang. Most of the cars this year have relatively small front and rear overhangs, but WSU, Twente, and Stanford all have relatively long front overhangs. It looks to me like all three have done everything they can to keep the wheelbase and catamaran fairings as short as possible. For what it's worth, these cars kind of end up looking a little like 2/3rd scale versions of last year's winner, Nuna 8.

There is a lot of variation when it comes to the shape of the array - not all of the cars are simple rectangles this year. Some of the cars narrow a little at the front or rear, but a few teams have taken this to the extreme and built "surfboard" shaped arrays: Nuon, Stanford, NIT, and JU. NuonNIT, and JU in particular have shaped their array such that no part of it sticks out to the left or right of the wheel fairings. A sharp junction between bodies is a tricky shape to design and can easily generate a large vortex if done poorly, and those three teams have simply eliminated a huge length of possible junction drag. It's super clean, and I like it a lot.

A horseshoe vortex spun off the front of a bad fairing junction
Source: The Leading Edge, 1999
NIT's very curvy "surfboard" array doesn't stick out sideways past the fairings

WSU has similarly eliminated this junction on the rear half of their car, but can't on the front due to the long, overhanging nose. Stanford appears to have also done something similar on the rear, but perhaps not as cleanly. Punch has done this on the front, but not the rear - the back of their array is rectangular, and sticks out to either side in the familiar way.

Rear wheel fairing/array junction on WSU's car; Nuon's is similar
(image source)

Nuon's car, in particular, really stands out to me. They've eliminated the front, side, and rear overhangs on the array. The car is formed by a single geometric sweep, starting at the base of one fairing, curving sideways into the array, and then curving down through the other wheel fairing and back to the ground. The team has completely removed any and all unnecessary protrusions and corresponding junctions from the car.

Photo: Jorrit Lousberg
(image source)

So, there's our 2017 Challenger field: Smaller cars in general. Still mostly asymmetric catamarans, but a lot more variety within that group. The five-fairing category has basically gone extinct, and some of the top teams are experimenting with long, narrow monohulls. It's going to be a really exciting race this year!

Come back next week and we'll talk about which specific cars we think will be the top contenders.


  1. Im concerned about Twente, looks like the didn't really move forward body wise while everyone else took a massive leap.

    1. Yep, see today's post: