Experiments

All you actually need to generate a supersonic flow is a nozzle, hole or orifice with an inside to outside pressure ratio of greater than 1.9. For example, between your compressor and the ambient air. In other words, if ambient air-pressure is 1 bar, then the compressor needs to generate a pressure of greater than 1.9 bar – and most small compressors can usually supply up to 7 to 10 bar of pressure.

Now sure, the flows you can generate have fairly low Mach numbers – but actually the flow inside the engine is also fairly low – because it has been slowed down by the intake. For example, the flow in the mixing area of a Mach 5 (free-stream) scramjet is only about Mach 2.4.

And yes, the flow is only over a small area (up to a couple of cm). But again, this is sufficient to test pellet dynamics, flow around wedges and similar important aspects of the engine. Because the Reynold’s number of these flows is so high, with the exception of the boundary layer, small objects behave similarly to large ones – so you can build your test objects on a smaller scale.

Compressor and extra air-receiver photo:

A setup like this will only give you a short burst of air before the pressure drops too low (a couple of seconds usually) and the larger the nozzle the shorter the burst. This means that experiments have to be short in duration and it’s often best to try and control the setup electronically if possible.

By careful design and production of the test area (the nozzle and its associated parts) using, for example, 3D printing, much more performance can be squeezed out of the system. The following file explains the options and a more technical view of these can be researched in Anderson’s books listed in the further reading page:

• Four options for supersonic wind tunnels.

It is also worth searching for “supersonic windtunnels” on Youtube and in Google to get an overview and design information.

This Excel spreadsheet allows you to find parameters for these.

And this document explains what’s in the spread sheet.

NASA have several pages about windtunnel nozzles and also some useful JAVA simulators for these. However, you need a JAVA enabled browser or Java IDE to view and use these (some modern browsers like Microsoft Edge won’t run JAVA for security reasons):

• Interactive Nozzle Simulator
• MOC Nozzle Simulator

Although you don’t need it (you can use a simple straight or parabolic curve – or copy an existing design profile), you can use the method of characteristics to design or more optimal or sophisticated diverging section if you wish to – how this works is explained in this document

Once you have your test setup or tunnel is up and running you can build a visualisation system or take measurements in the flow. These can range from simple shadowgraph systems to very complex optical setups.

Some more advanced design techniques are available here.