STATE-SWITCHED ABSORBER USED FOR VIBRATION CONTROL

Research conducted by:
Mark Horner Holdhusen
Advisor:
Dr. Kenneth A. Cunefare
 
A State-Switched Absorber (SSA) is a device capable of instantaneously changing its stiffness, thus it can switch between resonance frequencies, increasing its effective bandwidth as compared to classical tuned vibration absorbers for vibration control.

The research considered the experimental performance of the SSA for vibration suppression of an elastically mounted lumped mass base. State switching is achieved using magneto-rheological fluid to connect or disconnect a coil spring in parallel with other coil springs by applying or removing a magnetic field across of the MR fluid.  Experiments were performed over a range of forcing and tuning frequencies.  The SSA system, optimally tuned, outperformed the optimal classical TVA system for all combinations of forcing frequencies.

The research also considered the role of damping in the state-switching concept for a simple one-degree of freedom system and for a two-degree of freedom system.  Certain values of damping in the system improve performance, while other values hinder the performance of the state-switched absorber, as compared to classical absorbers. In general, a state-switched absorber with optimized tuning and damping is more effective at vibration suppression as compared to a classical vibration absorber with optimized tuning and damping.
 
 

Switches must occur at zero relative displacement across the spring. If switching occurs at a instant other than zero relative displacement, there is potential energy stored in the spring that is immediately released causing a shock or impulse response in the system. At each occurrence of zero relative displacement, the following maximum work extraction switching logic is used used to determine the next state of the system:
 
State-switching was implemented using MR fluid, which change from low viscosity with no magnetic flux across it to a high viscosity (nearly solid) with a large magnetic flux. A "switchable" spring was grabbed (flux on) and released (flux off) by the MR fluid. This changed the effective stiffness, and therefore the resonance of the absorber. The images below show experimental setup for the two-degree of freedom system tested.
mrssa 
 
The graph below shows the frequency response function for both state (flux on and flux off) of the state-switched absorber. There are two things to note from the graph. First, the resonance peaks are different for each state, implying that there is actually a change in stiffness and thus a change in natural frequency between the two states of the system. The second thing to notice is that the peaks when the flux is on are much sharper than with no flux indicating a change in damping between states. This is because with no flux the switchable spring flows through the MR fluid, whereas with the flux on the switchable spring is locked in the fluid, thus less damping is present with the flux on. This damping switching must be taken into account for any simulations.
 
 
The following graph shows the base response for one sample of an experiment run to validate the performance of the SSA at reducing vibrations of an elastically mounted base mass to which the absorber is attached. The first part of the time sample switching is not allowed, therefore, the system is acting a classical tuned vibration absorber. The last part of the time sample the system is allowed to switch and is acting as an SSA. Notice that for this specific forcing and tuning case the SSA reduces vibrations more than the TVA.
 
 
Similar experiments were for a range of tuning and forcing frequencies. The table below shows the optimized tuning parameters for both the SSA and the TVA and the corresponding base kinetic energies for each combinations of forcing frequencies. Notice that for each forcing case the energy ratio (SSA/TVA) is less than one, indicating that an optimized SSA performs better than an optimized TVA for each forcing case considered.
 
 


Damping Simulations
 
 

 
We have also looked at the role damping plays in the performance of the SSA. The first sytem looked at was a one-degree of freedom system where one two-state SSA was attached to a moving base and tuned to maximize the displacement to achieve a large energy dissipation. The SSA was compared to two similarly tuned TVAs attached to a similar base. The energy dissipation was mass normalized because the TVA system contains two masses as opposed to the one mass in the SSA system. The graph below shows the energy dissipation ratio (SSA/TVA). Notice that at low values of damping the ratio is highest, indicating that the SSA dissipates more energy than the TVA. Also, as the spacing in forcing frequency increases the ratio also increases, implying that the SSA is more effective the larger the bandwidth in the forcing. At higher levels of damping the TVA dissipates more energy than the SSA. This is because the switching law described above causes the SSA to hold in one state and not switch, thus the SSA is acting simply as a TVA at higher damping values.
 
 
The role of damping was also investigated for two-degree of freedom system where one two-state SSA was attached to an elastically mounted base mass. The system was compared to one TVA attached to a similar base. A direct search methods was employed to determine the optimal tuning for both the SSA and the TVA. The graph below shows the ratio of the optimized base kinetic energy (SSA/TVA) as a function damping for three different models of dmaping. As can be seen, the ratio is less than one for the entire range of damping considered, implying that the optimized SSA system has less base energy than the optimized TVA system for all dmaping considered. As in the one-degree of freedom system, the greatest relative performance occurs at low values of damping and increases as the spacing of forcing frequencies increase. Also, there is littlesignificant difference between the three damping models considered.
 
 
Currently, we are researching the performance of the state-switched absorber in continuous systems. I am optimizing the performance of the SSA using theoretical models that find the optimal tuning frequencies and location along a continuous beam. Once the theoretical optimization has been determined, an experimental study of the performance of the SSA on continuous beams will be performed. Continuous plates will be considered after the study of beams has concluded.
 
If you have any questions or comments, please contact Mark Holdhusen at the following email address: gte165r@prism.gatech.edu