Improvement of Wind Tunnel Flow Conditions:
Design and manufacture a flow conditioner structure seeking to improve the flow conditions inside a wind tunnel test section.
This This project was done for the Motor & Sport Institute (MSI) of Madrid where they teach a Master in Science of Motorsports Engineering. As part of the program, they teach external aerodynamics and CFD specializing in racing cars.
The MSI has a medium size wind tunnel in their facilities which is used to perform practical tests as part of the educational program as well as for commercial usage. It is important that the wind tunnel performs adequately and therefore that all sensors are working correctly and that the flow conditions inside the test section are within adequate levels.
Conditions Assessment
As part of the project the flow conditions inside the test section were checked and characterized using existing flow measurement equipment and in some cases by manufacturing new equipment. The main factors that were checked were turbulence intensity and flow uniformity.
In order to measure these parameters, the built-in pitot tube was used to calculate the turbulence intensity and a metallic rake was designed and manufactured with 16 3D printed pressure probes. It is known that measuring the turbulence intensity with a pitot tube is not ideal, however due to cost restrictions it was decided to use it knowing that some degree of error would exist.
Functional rake assembly inside wind tunnel.
Turbulence intensity Vs. Velocity at pitot tube.
Initial findings show that the turbulence intensity was higher than desired being close to 3% at high speeds. For the types of tests that will be performed in this wind tunnel it is advised that the turbulence intensity remains under 2 – 2.5%.
Rake results also showed a poor flow distribution inside the test section, with very poor performance at low speeds and settling from 20m/sec onwards. The graphs show a color map of normalized velocity for ease of comparison. The relative error between the mean rake velocity and the highest point of delta to that mean value is shown in the plot. The delta error settles to roughly 4% after 25m/sec. The results suggest that flow is separating from the inner corner of the elbow ahead of the test section.
Normalized velocity distribution at testing section (Left: 20% motor power; Right: 80% motor power)
Relative error between mean rake velocity and highest velocity delta at rake
Problem Mitigation
In order to mitigate the poor flow conditions inside the wind tunnel it was decided to introduce a new segment prior to the test section which would contain a number of flow conditioners. Due to the size of the room containing the wind tunnel the length for the new segment would be limited to 150mm. In this length we were able to fit one aluminum honeycomb and two high density stainless steel meshes. All three items are removable in order to test their performance and also for educational purposes for the students of the MSI.
Due to the lack of space, the design of the new segment had to be done with high precision manufacturing and all the flow conditioners were bonded to metal frames in order to maximize packaging space.
CAD design of new wind tunnel segment containing removable flow conditioner.
In order to ensure reliability of the mesh assembly, tensile tests were performed to ensure the chosen adhesive would work reliably and provide enough bonding force. The test was performed in comparison to using spot-welding. The results show that the bi-component adhesive can withstand more force that the spot-welding option. In both cases were the bi-component was used the failure point was the mesh and not the adhesive.
Tensile strength test.
Tensile test results and failure modes.
Analytical results of tensile strength tests with mesh20.
The final assembly is shown in the images below.
Final assembly of flow conditioners segment.
Flow Conditioner Implementation Results
When implementing the flow conditioners results show an improvement in both turbulence intensity and flow distribution inside the wind tunnel test section. The turbulence intensity significantly reduces reaching roughly 1% from below 20m/sec.
Turbulence intensity Vs. Velocity at pitot tube with different flow conditioner configurations.
The flow distribution also improves across all velocity range, especially at high speeds.
Normalized velocity distribution at testing section post flow conditioners.
20% motor power.
Normalized velocity distribution at testing section post flow conditioners.
80% motor power.
The improvements achieved by the implementation of flow conditioners have been substantial enough to be able to perform tests with good level of accuracy and repeatability.