The problem
Drag is hard to measure directly, so engineers infer it from pressure. But which measurement can you trust?
Wind tunnel testing · ME 347 · Cal Poly
Measured the drag on a cylinder two different ways.
In Cal Poly’s wind tunnel, my team of four measured air pressure around a cylinder and the wake behind it. One method matched published data and the other didn’t. Figuring out why was the real lesson.
- Role
- Testing & analysis
- Team
- 4 students
- Equipment
- Wind tunnel, Pitot-static probe
- Course
- ME 347, Cal Poly
What I did
Recorded surface pressure every 10° around the cylinder and surveyed the wake at two distances behind it.
The result
The pressure data matched known behavior. I traced the wake mismatch to specific measurement error sources.
19surface pressure points
13.75m/s wind speed
2independent drag methods
2wake survey stations
How I did it
- 01
Set up
Mounted a brass cylinder across the flow and located its stagnation point.
- 02
Measure
Recorded pressure every 10° from front to back.
- 03
Survey
Traversed a probe through the wake at two downstream distances.
- 04
Compare
Computed drag both ways and explained why they disagreed.
Skills I used
- Wind tunnel testing
- Pressure measurement
- Wake surveys
- Numerical integration
- Error analysis
Full engineering write-up
Worked with a team of four on this lab, and I measured how air moves around a brass cylinder and compared two independent drag estimates. The pressure result matched published behavior; the wake result did not, making measurement quality as important as the final number.
01
Test setup and method
One cylinder, two ways to measure drag
A 1.50 in brass cylinder was mounted perpendicular to the flow in the Cal Poly mechanical engineering wind tunnel. After locating the stagnation point, the team recorded surface pressure in 10-degree increments from 0 to 180 degrees.
A Pitot-static probe on an x-y traverse then sampled the wake at 17.5 in and 25.0 in downstream. Ambient conditions and free-stream dynamic pressure established air density, velocity, and a cylinder Reynolds number of 3.49 × 104.
1.50 inBrass cylinder diameter
13.75 m/sCalculated free-stream velocity
3.49 × 104Diameter-based Reynolds number
02
Surface pressure
The expected signature of separation
The stagnation point produced the largest pressure coefficient, Cp = 1.03. Pressure fell as the flow accelerated around the cylinder, reached Cp = -1.54 near 70 degrees, and remained negative across the rear surface—a broad low-pressure region consistent with separated flow.
Interpretation: the nearly flat, negative pressure on the back side indicates that the boundary layer separated and left a wide turbulent wake rather than recovering to the ideal-flow pressure distribution.
03
Wake development
Velocity deficit and downstream recovery
Both wake surveys showed their lowest normalized velocity near the cylinder centerline and recovery toward free-stream velocity away from the wake. The profile farther downstream appeared broader and smoother, which is consistent with turbulent mixing.
04
Drag comparison
A good result and a useful discrepancy
Numerically integrating the surface pressure distribution gave CD = 1.256, close to the expected value of approximately 1.2. Integrating the wake momentum deficit produced much lower values, so those results were treated as evidence of an incomplete or inconsistent wake survey—not as validated drag coefficients.
| Method | Location / source | Drag coefficient, CD | Assessment |
|---|---|---|---|
| Surface pressure | Cylinder surface | 1.256 | Agrees with reference |
| Wake momentum | x = 17.5 in | 0.179 | Underestimated |
| Wake momentum | x = 25.0 in | 0.031 | Underestimated |
| Reference behavior | ReD = 3.49 × 104 | ≈ 1.2 | Expected range |
Likely error sources: unsteady vortex shedding, probe misalignment, pressure-zero drift, an incomplete vertical integration range, and free-stream inconsistency. Several measured u/U values above 1 outside the wake support that diagnosis.
05
Engineering takeaway
Validate the method, not only the arithmetic
The experiment connected surface pressure, separation, wake formation, and pressure drag in one test. More importantly, comparing independent methods exposed a measurement problem that a single calculation would have hidden.
A stronger repeat test would re-zero the transducer before each traverse, verify probe alignment, extend the survey beyond the full velocity deficit, and repeat multiple passes to quantify the unsteadiness of the wake.