Aerospace testing service organizations must continually incorporate new technologies into their testing facilities to keep pace with the increasing demands of their customers. Aerospace Wind Tunnel Facilities can run in the $100s of millions to build. Key to keeping wind tunnel services profitable is providing offerings that provide customers with deeper insights faster reducing customer R&D costs and improve product time-to-market. 

SUMMARY

A world renown wind tunnel testing services facility implemented a new edge-based computing system with Viviota software to gain significant efficiencies and quality improvements in high-volume image enhancement processing procedures. With this solution the facility can support research of next-generation algorithms for flow visualization and can now double its camera capacity as planned.

• Improved processor core usage by 22X
• Reduced overall system power consumption by 20X
• Processed 4X as many images per second
• Supported 10X more camera feeds

NEW SERVICE OFFERINGS

Wind tunnels collect data from many different sensor sources during operations. This includes sensors for controlling environment settings such as airflow and temperature. In aerospace, applications sensor instrumentation is highly specialized and difficult to deal with because sensors must endure extreme conditions in wind tunnels.

Patterns in airflow are visualized using highly sensitive imaging sensors. This flow visualization process helps engineers and scientists analyze test models. This information is critical to designing airplanes and spacecraft, and for comparing with data collected from computer simulations. To revolutionize aerodynamic science in wind tunnels, engineers are using highspeed infrared (IR) cameras for flow visualization of aircraft and spacecraft models. Introducing camera sensors has multiple benefits. Camera sensors are:

• Easier to maintain and calibrate because they can be mounted
 outside the wind tunnel
• Passive sensors so they don’t affect the results of the tests 

CHALLENGE - HIGH-SPEED IMAGE DATA ENHANCEMENTS

The image sensor data collected during wind tunnel operations is enormous. An IR camera produces terabytes of image data for a single test. The raw image data must be refined for it to be useful. To do this, the engineers need an edge-based system capable of enhancing images on-the-fly.

Highly sensitive IR cameras are used to capture high-resolution images at high rates of speed. This results in several terabytes of image data for each hour of testing. To capture information from different surfaces of the model, multiple cameras are used. Although this creates a massive amount of data to store, the more immediate challenge is processing the raw image data so it can be useful.

Flow visualization requires image data that reveals temperature changes on surface regions. The raw image data is insufficient and complex image processing must be done to improve quality. For this the engineers tried to implement a well-known contrast enhancement algorithm, but it ran far too slow to keep up with incoming camera data. What was needed was a new edge-based computing system capable of processing images in preparation for storage and post-processing. The wind tunnel facility turned to Viviota to solve this edge computing challenge.

OBJECTIVE - A SYSTEM FOR PROCESSING HIGH-SPEED CAMERA DATA

The wind tunnel facility added industrial quality IR cameras to operations. The image data is collected and integrated with existing sensor data such that all sources are synchronized for use by in-house data processing software. The test models must have proper thermal properties so that the information extracted from the processed images is reliable. With the existing image processing software being too slow, it was unclear whether the current enhancement algorithm was the right choice for flow visualization. The solution must enable exploration of alternative algorithms as well as meet timing constraints.

Ultimately, this system must process camera data under the following conditions:

• Execute from a compute system capable of supporting 4-8 camera feeds simultaneously
• Process image data at a rate greater than 100 frames per second for each camera in operation
• Run on a system that is designed for edge environments (e.g. low power and rugged)
• Integrate with existing engineering software and hardware infrastructure
• Eliminate the need for embedded or specialized equipment
• Provide an upgrade path for more cameras with high resolution and high frame rates

SOLUTION - VIVIOTA'S ADVANCED CONTRAST ENHANCEMENT ANALYTICS ON HPE EDGELINE

Viviota deployed a real-time ready contrast enhancement library that ran faster than the existing image processing software. The software runs on four HPE ProLiant Server cartridges in an HPE Edgeline EL4000 chassis. The HPE Edgeline Series is designed for edge applications, in the engineering test environment where engineering work gets done. It uses low-power Xeon server processors in a rugged industrial chassis, enables high-speed storage with onboard NVMe, and integrates with existing IT infrastructure.

• Deployed a key image contrast enhancement algorithm
• Integrated a high-performance edge system with scalability 

RESULTS - EXPANDED SERVICE OFFERINGS

The Aerospace Wind Tunnel Facility is now able to offer expanded services supporting research of next generation algorithms for visualization. The facility has plans to double its camera capacity.

• Improved processor core usage by 22X
• Reduced overall system power consumption by 20X
• Processed 4X as many images per second
• Supported 10X more camera feeds