Trim and Drag
by Lester Gilbert
Having supervised a number of student projects in the
Southampton towing tank, I thought it about time I ran a few experiments that I
was interested in. Graham Bantock had
earlier shown me his very neat “falling weight” device (FWD) for towing a hull,
so we arranged to spend two days on a summer project investigating hull drag.
Graham brought his FWD and two hulls, Fraktal and Ikon, and I brought my
Pikanto and some timber. Amongst a
number of other experiments, we wanted to know whether trimming an IOM nose-down
or nose-up made much difference to hull drag.
The results were surprising…!
FWDGraham’s FWD comprises an axle and a drum. A line is wrapped around the axle several times and attached to a weight. A second, towing, line is attached to the drum at one end, and to the hull at the far end of the towing tank at the other. Release the weight, and the drum rotates, reeling in the towing line and towing the hull. The weight needs to be reasonably substantial, the drum well balanced, and the axle bearings as friction-free as possible. The net result is a pulling force which is somewhere between 20 and 200 grams depending upon the weight used. Graham’s FWD has an ingenious “turbo” mode, where a second weight piggy-backs on the primary weight for a short time to provide extra drive at the start. The picture shows Graham adjusting the FWD which is mounted on some planks spanning the towing tank. When the drum reels in about 10 m of towing line, the weight drops around 0.25 m, giving a gearing of about 1:40, pretty much what you would expect from a drum with a diameter of about 200 mm and an axle diameter of about 5 mm.
Towing tankFor the towing experiments, the boat rig is replaced by a stub mast. This serves as a towing post, and also to secure a yard which, when extended over the side of the hull, allows weights to be attached so that the hull heels as required. We did not heel the hull for the experiment described here [see Canoe body yaw (AMYA MY #176) for the heeled experiments], but the picture shows the Ikon model heeled over and ready for the start of a towing run. While the University of Southampton’s “Lamont” towing tank is around 30 m long in total, we just used the central 12 m section. From release, the model accelerated to within 99% of its final speed over a distance of around 5 m, and it was then timed between two marks approximately 6 m apart. Using a weight on the FWD axle of around 0.8 kg gives a constant pull of around 20 grams. Plus the additional “turbo” boost of around 15 grams for the first metre or so, the model took about 17 seconds to reach the first mark, and then around 13 seconds to reach the second mark, moving at around 0.5 m/sec. I’m always surprised at how such a small driving force gets our models going – 20 grams force (4 sheets of paper, very light!) gets the boat up to 0.47 m/sec (a reasonable 1 mph!). If we were on the pond, this would be a light to medium-light airs day.
IOM trimThere is always an interesting debate about where to place the weight (battery pack, servos, and/or corrector weight) in a boat. If you are anticipating light airs, it is tempting to think the best place for the weight is in the bows, while if you expect heavy air, conventional wisdom says to put the weight in the stern. In the Ikon IOM which floats on a design LWL where the plumb bow and the transom stern both just touch the water plane, putting a 300 gm corrector weight in the bows puts the bows about 4 mm in, while putting the corrector into the stern puts the stern about 5 mm in and gives an overhanging bow of around 40 mm. We carried out a series of experimental runs with an Ikon where the weight was placed amidships, right on the bows, or right at the stern. The Ikon displaced exactly 4 kg, and we towed the hull with a relatively low weight on the FWD axle such that the hull reached a top speed of around 0.47 m/s.
Top speedThe figure shows the speed of the hull between the marks. In normal trim, this turned out to be around 0.471 m/s as the average over 5 runs, with a measurement error (standard error of the mean) of plus or minus 0.001 m/s or so. With the weight in the stern, speed dropped to an average of 0.466 m/s over 5 runs (error ±0.0002), while with the weight in the bows, speed was essentially unchanged at around 0.472 m/s over 5 runs (error ±0.0005). It is worth noting that measurement error is approximately 0.10% overall here. The speed difference between bows down and stern down is about 0.006 m/s, and that certainly doesn’t sound like much, about 1%. But if we imagine our boats have started equally well and are beating to the windward mark 100 m away, the slower boat will be about one boat-length adrift entering the zone… The surprise is that there is no difference in speed between bows-down trim and LWL design trim.
Acceleration timeMy job in the towing tank was to release the hull, move to the first mark, take a time reading, move to the second mark, and take the second time reading. While I was releasing the hull during these runs, I had the very marked impression that the Ikon was “sticky” when trimmed stern down, somehow slow to get going from being dead in the water. I was very interested to see the results of the timing from release to the first mark. In normal trim, this turned out to be around 17.3 s. When stern down, and bows down, this was around 17.4 s. The measurement error is higher with these measurements, around 0.35% overall, due to the difficulty of determining exactly when the hull first starts to move, so the results are not so clear-cut. But we can be quite sure that the LWL trim was quickest off the start, with both the bows and stern trims showing slower accelerations. The surprise was that the bows-down trim was as “sticky” as the stern-down trim to get going. This difference in acceleration is very small, a mere 0.1 sec, perhaps 1%, and on a real starting line we are unlikely to be moving at 0 mph at the gun anyway. If we imagine that we have not timed our run at the line perfectly and, in the light conditions we are investigating, find that we need to accelerate after the gun, then this slightly lower acceleration is likely to find us, after 10 seconds or so, perhaps 20 or 30 mm behind our better trimmed opponent who started in the same way with us; so not that much in it.
ConclusionsGraham notes that stern down trim is so much more than 1% worse in real life, and suggests that poorer balance may be the real culprit rather than the drag increase.
AcknowledgementsThese experiments would not have been possible without Graham Bantock’s enthusiasm and knowledge, or without the support of Prof Philip Wilson of the University of Southampton’s Department of Ship Science.
©2023 Lester Gilbert