Objective: To produce a near-real-time phased-array acoustic measurement and visualization system for wind-tunnel testing by combining the skills of the DARWIN and HPCC projects, and to apply the system to the analysis of the acoustic environment around a DC-10 model with extended flaps and landing gear in the Ames 40x80 wind tunnel.
Approach: Instrumentation, data collection, and storage computer systems are combined with the HPCC IBM SP-2 so produce a heterogeneous distributed computing system. The Parallel Virtual Machine (PVM) software provides data communications between machines and within the SP-2. Acoustic information is collected by an array of 40 microphones, and is stored in memory on the instrumentation computer This digitized data is routed to the SP-2 for phased array processing. A surface of points are "scanned" to determine the strength of noise sources at each location. Sound pressure levels on this surface are visualized in the FAST visualization system. A graphical user interface provides an easy-to-use data entry environment.
Accomplishments: Prototype software is complete linking the NPRIME data collection system, the graphical interface, and the IBM SP-2. Calibration tests were carried out during late July and early August. A survey of the acoustic environment around a DC-10 model in the 40x80 wind tunnel was completed September 1, 1995.
Significance:Recent improvements in engine noise has increased the relative contribution of the airframe to the total noise produced by aircraft during landing. Tighter airport noise regulations may limit the markets of U.S. transport aircraft manufacturers. Prior analysis procedures completed the analysis of a few frequencies overnight. The new system provides analysis for dozens of frequencies in less than 5 minutes (between test points). During the DC-10 test, several previously unknown noise sources were identified. McDonnell-Douglas and other participants in the Advanced Subsonic Transport (AST) program are pleased with the results.
Status/Plans: Analysis of the DC-10 data is continuing. Improvements in parallelism and solution efficiency should allow the visualization of hundreds of frequencies in near-real-time. With a greater number of microphones, greater resolution will become possible at higher frequencies, and volume (as opposed to surface) rendering will be practical. This will require even more computational horsepower to meet the near-real-time requirement.
Point(s) of Contact:
Merritt H. SmithDARWIN Web Page at NASA Ames Research Center: http://racimac.arc.nasa.gov/