Browsing by Author "Sumali, Hartono"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Demonstration of active structural acoustic control of cylinders
    Sumali, Hartono (Virginia Tech, 1992)
    Active control is applied to reduce noise emission from a vibrating elastic cylinder by exerting forces on the cylinder that cancel the noise-generating vibration. This technique is called Active Structural Acoustic Control (ASAC) (Fuller, 1987). Sensors are implemented using piezoelectric film, and actuators are implemented using piezoceramic material. Both analog and digital noise cancellation control algorithms are used to reduce the noise emission from the cylinder. Two-cylinder boundary conditions are taken as case studies. The first boundary condition is the open cylinder case. The second boundary condition is where the cylinder has an end plate bolted to each end. Actuator placement and the sensor design are done by first obtaining the natural frequencies and mode shapes of the cylinder using both analytical and experimental methods. Modal sensors developed and tested in previous work (Lee, 1989) are applied. After preliminary control experiments with analog feedback loop show that control can be done with the sensors and the actuators, digital signal processing hardware programmed with the filtered-x Least-Mean-Square adaptive control algorithm is used to control the vibration of the cylinder. The excitation is single-tone on-resonance. Acoustic - testing demonstrates that ASAC reduces the sound pressure level generated by the vibrating cylinder by up to 29 dB in the reverberant field. Vibration measurement reveals that the reduction in sound emission from the cylinder is a result of reduction in vibration. The adaptive controller reduces the vibration level by up to 68 dB.
  • Thumbnail Image
    A New Adaptive Array of Vibration Sensors
    Sumali, Hartono (Virginia Tech, 1997-07-03)
    The sensing technique described in this dissertation produces modal coordinates for monitoring and active control of structural vibration. The sensor array is constructed from strain-sensing segments. The segment outputs are transformed into modal coordinates by a sensor gain matrix. An adaptive algorithm for computing the sensor gain matrix with minimal knowledge of the structure's modal properties is proposed. It is shown that the sensor gain matrix is the modal matrix of the segment output correlation matrix. This modal matrix is computed using new algorithms based on Jacobi rotations. The procedure is relatively simple and can be performed gradually to keep computation requirements low. The sensor system can also identify the mode shapes of the structure in real time using Lagrange polynomial interpolation formula. An experiment is done with an array of piezoelectric polyvinylidene fluoride (PVDF) film segments on a beam to obtain the segment outputs. The results from the experiment are used to verify a computer simulation routine. Then a series of simulations are done to test the adaptive modal sensing algorithms. Simulation results verify that the sensor gain matrix obtained by the adaptive algorithm transforms the segment outputs into modal coordinates.