Sharon L. Padula, Daniel L. Palumbo* and Rex. K. Kincaid**

Multidisciplinary Optimization Branch, NASA Langley Research Center

*Structural Acoustics Branch, NASA Langley Research Center

**Department of Computer Science, College of William and Mary

June, 1996

RTR 505-63-36-06

Research Objective. Active structural acoustic control of aircraft interior noise requires arrays of piezoelectric actuators and microphone sensors. The goals of this research are to assess the importance of sensor/actuator location, to develop combinatorial optimization routines for selecting the best locations and to test the optimized arrays in the Acoustics and Dynamics Laboratory.

Approach. Combinations of 4 out of 8 actuators and 8 out of 462 sensors were tested on a composite cylinder which simulates a commuter class aircraft fuselage (see photo). As in previous Structural Acoustics Branch tests, the noise source was a loudspeaker and noise control was implemented at several discrete frequencies. Noise with and without active control was measured and the difference was reported in decibels (dB). Prior to testing, the MDO Branch predicted the noise reduction potential for all possible combinations of 4 actuators assuming a perfect control system. Then for the best and the worst actuator array, the performance of the control system was simulated for 1000 randomly selected sets of sensors. A histogram of these results is shown for the 275 Hz noise source. ( See bar chart.) The predictions range from a +3 dB noise increase to a -5 dB noise decrease depending on the sensor/actuator configuration. Since all possible combinations of sensors and actuators could not be investigated, a modified tabu search optimization method was developed to select the optimum sensor locations for the best and worst actuator locations.

Accomplishment Description. The measured difference between the best and worst case of actuator locations was slightly less than the predicted difference (see table). Furthermore, the best case actuators with optimized sensor locations performed noticeably better than sensor/actuator locations selected by traditional methods which were used in a previous test. These trends were consistent for the three different noise source frequencies tested. This is an experimental verification of the predicted importance of sensor/actuator location.

Significance. Active noise control of low-frequency harmonic noise sources (e.g., propeller noise) is more attractive than passive acoustic baffling because it has the potential for significantly greater noise reduction for constant or reduced weight. These test results suggest that optimizing sensor/actuator locations is practical and yields added noise reduction without added complexity or cost.

Future Plans. New laboratory tests are starting immediately with an increased number of potential actuator locations. Flight tests are also being considered. The tabu search optimization method will be extended to allow a much larger combinatorial design space and to consider the predicted uncertainty in sensor measurements as it affects the choice of sensor location.

Figure: Optimized Sensor/Actuator Arrays for Active Noise Control