Coastdown Testing at the University of Illinois.
Obective
Confirm that the fairing geometry identified during wind tunnel testing reduces drag on a full-scale chair
Approach
Measure the deceleration of a racing wheelchair as it coasts, with and without a fairing, and compare to determine whether the chair decelerates more quickly with or without the fairing
Test Setup
A partial fairing designed using wind tunnel testing, shown in Fig. 1, was fabricated from carbon fiber. Coastdown tests were conducted on the chair, with and without the fairing installed, in which timing blocks were used to measure the time it took the chair to coast through seven 25-ft. intervals. The timing blocks were constructed from IR detectors which output a voltage sampled at 2.5 kHz. Laser pointers on the opposite side of the coast track were directed at the IR detectors. When the wheelchair coasted down the track, it broke the beam of each laser pointer in sequence and the time at which each beam was broken was recorded. These data were used to construct a velocity profile of the chair over the length of the coast, and several repeat runs were performed. The track was constructed in the West Hall of Memorial Stadium at the University of Illinois, which sheltered the track from wind. To account for the possibility of varying run conditions, some of the runs without the fairing were conducted before the fairing was installed and some were conducted after it was removed.
Fig. 1 Full-scale carbon fiber fairing prototype used in the coastdown test. Fairing was based on a design found during a previous series of wind tunnel tests to reduce drag by nearly 12%.
Fig. 2 Diagram of coastdown track constructed in Memorial Stadium at the University of Illinois.
Test Equipment
Coastdown Track: University of Illinois Memorial Stadium West Hall
Track Length: approx. 450 ft.
Acceleration length: 235 ft.
Coastdown length: 175 ft.
Braking zone length: 40 ft.
Max Speed before coast: 17 mph
Instrumentation: 8 timing blocks spaced 25 ft. apart over track length, uncertainty +/-0.0002 s
Results
For a perfectly flat track, it would be expected that the coast down curve be parabolic. A parabola consists of three terms: a constant, a linear term, and a quadratic term. The constant is proportional to the initial velocity of the chair, the linear term is proportional to the rolling resistance of the chair, and the quadratic term is proportional to the drag on the chair. So in theory, absolute values of rolling and air resistance can be obtained from accurate measurements of a coast.
In practice, it is extremely difficult to separate air resistance from rolling resistance for something like a wheelchair, although rolling resistance can be obtained from low speed tests in which air drag is negligible. The problem is that it is difficult to push a chair fast enough for drag to be much larger than rolling resistance. This means the quadratic term is very small compared to the linear term. An example of this is shown in Fig. 3; the blue curve shows the theoretical coastdown velocity profile of a wheelchair. Separating the air resistance from the rolling resistance in this theoretical profile would require the very slight curvature of the blue curve to be measured, which is difficult to do experimentally.
To complicate matters further, small fluctuations were present in the grade of the coastdown track. Although the track appeared level, measurements of the chair's coast revealed that the chair did not have a nearly constant deceleration. Some sample velocity profiles are shown in Fig. 3, and it is clear that the chair decelerated more quickly over some stretches of the track than others. Simulations showed that changes in grade of 0.2% and less would cause fluctuations similar to those observed in the data. In general, these fluctuations occured at the same locations in different runs, although some variation among different velocity profiles was present and most likely due to steering inputs to keep the chair straight. The changes in grade could be partially accounted for by running the coast in the opposite direction, but this was not done due to time constraints. However, the slope of the course did not change for different runs, so comparisons among runs could be made. While absolute values of CdA could not be measured accurately, it could still be determined whether or not the fairing was effective at reducing drag.
Fig. 3 Comparisons of theoretical wheelchair coastdown velocity profile with samples of experimental data obtained with and without the fairing. The theoretical data is based on wind tunnel data for CdA, but rolling resistance has been adjusted to help account for the varying slope of the coastdown track.
To mitigate the effects of the changes in grade over the track, the average deceleration between the last and first intervals was calculated for each run. This effectively averaged out the changes in slope and made comparisons among different runs possible. Paired t-tests comparing coasts with and without the fairing showed the runs with the fairing to have a smaller average deceleration than runs without the fairing at a 98% confidence level, based on 8 repeat runs in each configuration. Since the runs were conducted by the same athlete in the same chair, the rolling resistance and weight were virtually identical between the two sets of runs. Additionally, the average velocity for each set of runs was almost equal (actually slightly higher in the cases in which a fairing was installed). Since all other parameters were equal, these results strongly suggest that the fairing did reduce wheelchair drag.
Although it is difficult to determine by exactly how much the fairing reduced drag, it is possible to estimate the performance enhancement the fairing may afford. The average deceleration with the fairing was 7% less than without. This translates to a difference in power of about 5 W at a cruising speed of 16 mph, and the fairing would have a larger effect at higher velocities.
Rolling resistance is more easily measured than aerodynamic drag because it does not vary significantly with velocity and is relatively large even at low speeds. In addition to the higher speed coastdown tests described above, low speed tests were conducted to measure rolling resistance. The wheelchair coasted down from slightly less than 5 mph several times over a 75 ft. section of the track, and measurements of the coast were obtained using the same timing blocks that were used in the higher speed coasts. This was done several times in each direction to cancel the effects of changes in slope over the track. Note that it was much easier to run the low speed test in both directions since very little run-up was required to get the chair up to 5 mph.
At these low speeds, aerodynamic drag makes up less than 20% of the resistive force on the wheelchair. While this is small enough that rolling resistance is easily measured, it is large enough to have an effect on those measurements. Therefore, the data were corrected for drag by using CdA data obtained from wind tunnel testing. With this correction, the rolling resistance coefficient (Crr) was calculated to be about 0.0034 in the more downhill direction, and about 0.0046 in the more uphill direction (the terms uphill and downhill may be a little misleading here, as the grade was less than 0.2%). Since both rolling resistance and grade are linear, these two measurements can be averaged to eliminate the effect of grade, so the true Crr = 0.0040.
Conclusions
Coastdown testing confirmed earlier wind tunnel test results and showed the fairing to reduce drag compared to the unfaired wheelchair. Absolute values of CdA could not be obtained from this testing due to variations in grade along the coastdown track. An absolute value for Crr was obtained by coasting in both directions -- Crr was found to be about 0.0040.
Given these results, the fairing is recommended for use in race events. The fairing can reduce athlete input power required by about 5 W at long-distance race cruise speeds, which can result in a significant integrated time savings.
