(Skip to the technical note)

There are a lot of advantages of using a seven-hole probe for aerodynamic measurement: they are reasonably inexpensive (compared to laser diagnostics), they can provide three components of velocity (if properly calibrated), and are based on comparatively simple technologies. They are also very robust, and experience minimal drift over long periods of time- making them ideal for air data systems and other control applications.


When using seven-hole probes for measurement, the difficulty often arises in obtaining three components of dimensional velocity from the pressures obtained. Many off-the-shelf products will do this for you, but it is nevertheless important for the system designer to be aware that a number of different techniques exist for this process, offering different trade-offs between uncertainty and required computer time.

How multi-hole probes work

If you take any bluff body subjected to flow, the pressure distribution around the body will be uniquely related to the magnitude and direction of incoming velocity. So long as you have more than three pressure sensors measuring within this distribution (you can only get as many independent outputs as you have sensors), and providing that you have some calibration data to compare to, you should be able to get a reasonable estimate of velocities from your pressure measurements. Multi-hole probes usually have either hemispherical or conical tips in which the holes are located, so the tips act as the bluff body and the holes provide access to the local pressure distribution. Five- and seven-hole probes are most common: five holes provide a convenient mapping to Cartesian coordinates, while seven holes pack the maximum possible number of holes (for best fidelity) into the available circular cross-section.

The catch

The difficulty arises in exactly how you get the velocity estimate from the pressure distribution. At higher angles, part of the probe tip will be in a the wake of the probe itself, and will not be able to return useful values (pressures in a wake don’t really change much). This causes two problems: first, if the pressures from certain sensors aren’t really changing with the conditions, they’re just contributing noise. Second, because holes will tend to switch fairly abruptly from reading meaningful pressures to being in wake flow, so it becomes very difficult to represent the responses as analytical functions. Often, a probe will have multiple calibration functions, and will switch between them depending on the flow direction: this can lead to some jarring discontinuities in the measurements if you’re not careful.


Selecting a data conversion technique

There is no one-size-fits-all algorithm for converting pressures to velocities. For a UAV air data system, for example, you need fairly high bandwidth and may not have much spare memory in your flight data computer: a reduced-order functional approximation or piecewise blended approach may be best. For less conventional probe geometries or “omni-probes”, a more generalized approach may be needed. For high-resolution, high-fidelity wind tunnel measurements, you may want the best accuracy money can buy- in which case a more sophisticated machine-learning approach may be best.

Choose wisely

Regardless of how good, how small or how precise the probe is, the measurements are only as good as the data conversion technique- so make sure that you are using one which is best for you. Our technical report explains how the entire data conversion process works, the different methods available, and how to implement these for your own specific applications.