Design of experiments plays a major role in wind tunnel balance calibration at AEDC

  • Published
  • By Philip Lorenz III
  • AEDC/PA
ARNOLD AIR FORCE BASE, Tenn. --  When a new six-component wind tunnel balance was fabricated at AEDC and required a full calibration for acceptance testing, AEDC Test Technology Branch project manager Carrie Reinholtz realized this project would provide a great opportunity to use an innovative approach called Design of Experiments (DOE) to calibrate the new balance.

DOE is exactly what its acronym states - experimental design. In other words, it's designing a way to collect data points to determine which variables affect a process, with statistical confidence and uncertainty. A DOE based approach is conducted in a randomized fashion in order to minimize errors. These errors include biases that may be encountered by measurement apparatus or user errors as just two examples.

David Yoder, one of the ATA project engineers working on calibrating the new balance, said, "For us, it was the first time that we had ever tried to use DOE in a balance calibration process.

Yoder compared and contrasted the two approaches taken for the recently completed balance acceptance testing.

"The traditional calibration practice at AEDC has involved dead weight loading using a calibration stand with a 'One Factor at a Time' approach," he said. "This traditional [calibration] approach takes usually around five days to complete, but with the Design of Experiments' load sequence, the dead weight randomized loadings took about two and half days.

Reinholtz, explaining the objective behind the two approaches, described what equipment needed calibrating.

"A wind tunnel balance is a device that measures the aerodynamic forces and moments acting upon a body, like an aircraft model or fuel tank mounted under an aircraft wing," she said. "These balances are used in wind tunnels like we have at AEDC and at other similar facilities around the country such as NASA Langley in Virginia," she said.

The calibration process requires hanging precise weights from the balance and resolving the measured forces and moments on the balance into the normal, side and roll components. As many as 1,000 weight configurations are used during a typical "one factor at a time" (OFAT) calibration sequence, resulting in the calibration process taking several days to complete. Reinholtz had taken Old Dominion University Professor Dr. Drew Landman's Design of Experiments (DOE) course and she had also taken several experimental design classes through the Air Force Institute of Technology's (AFIT) distance learning Test and Evaluation Certificate Program (TECP). She realized that the use of both DOE and OFAT for the balance's acceptance testing would be an excellent approach to assess the value of DOE for this type of application.

"This provided an excellent opportunity to introduce the use of DOE for balance calibrations," Reinholtz said. "Despite the given constraints, it ended up being a successful demonstration of where we could use DOE at AEDC."

In an American Institute of Aeronautics and Astronautics (AIAA) technical paper, coauthored by Dr. Landman and AEDC's Reinholtz, David Yoder, Paul Jalbert, and Dr. Doug Garrard, they wrote, "For more than a decade, DOE has been implemented in wind tunnel strain gage balance calibration processes [at places other than AEDC]. Calibration with DOE is a characterization process where data are collected and analyzed using statistical methods, allowing conclusions to be drawn with chosen levels of confidence and power."

Yoder added, "In this case we had the opportunity to design a load schedule [plan] that could be randomized within the constraints of the calibration hardware, so it was an opportunity to try Design of Experiment approach to meet our acceptance test requirement."

Explaining the DOE approach their team took, Yoder said, "We start out with measurement system uncertainty goals and the standard calibration approach gets there within 1,000 points and Design of Experiments got us there equivalently with 200 [points]."

Yoder emphasized that it is important to understand that a DOE approach does take some additional preparatory work to fully understand the requirements of an experiment. He feels that DOE methods could be used in all balance calibrations, but with this being AEDC's first attempt, additional studies could help reinforce this belief.

Yoder said using both an OFAT and complimentary DOE approach for the balance acceptance testing was a great learning experience.

"I think we got to the very best solution for this problem because of the collaborations with NASA and DOE experts, along with the program management allowing adequate planning time to fully understand the problem," he said.