Using parallel processing to speed-up rotorcraft air vehicle flight test analysis

Abstract

Rotorcraft air vehicle flight testing is required to determine how well the aircraft will perform in selected operational scenarios. The actual air vehicle flight test program may require extensive planning, instrumentation, flight testing, data analysis, and reporting. Considerable emphasis has been placed on working to reduce the cost and cycle time associated with conventional flight testing. One option with the potential to help reduce the cost and cycle time of air vehicle flight testing involves using credible simulation. Although numerous simulation models exist, very few have the capability to duplicate the flight test techniques taught at the US Navy Test Pilot School (USNTPS). With dedicated research and development, these classic flight test techniques have been implemented into a generic analytic model structure. The analytic model structure can be operated on a personal computer with a Linux operating system. The time required to run a complete analytic rotorcraft air vehicle flight test program can be overnight or much longer depending on the parameter variations selected for the specific computer run. A goal is to be able to use the analytic options to show how close the projected flight test matrix flight conditions approach flight limits. Another goal is to be able to show the sensitivity of selected flight test parameters while the test plan is being developed. A recent High Performance Computing (HPC) portfolio on Collaborative Simulation and Testing (CST) was established to help integrate analytic options to enhance flight testing. The CST-03 topic focuses on developing parallel processing options to greatly speed up the analytic ability to predict rotorcraft performance, stability, and control.