Magnetorheological (MR) fluid is categorized as smart material whose rheological properties can be varied instantaneously under the application of an external magnetic field. Utilization of these smart multifunctional materials into the devices and structures provides a unique opportunity to develop adaptive devices/structures capable of changing their dynamic characteristics in response to wide range of external disturbances. MR fluid have been recently utilized in sandwich panels to provide variable stiffness and damping to effectively control vibrations. The main objective of the present dissertation is to investigate the sound transmission loss (STL) capability of sandwich panels treated with MR fluids at low frequencies. This dissertation contributes in three major parts. First the effect of applied magnetic field on the structural and acoustical behavior of MR fluid sandwich panels is experimentally investigated. An experimental test setup including two anechoic chambers and an electro-magnet has been designed and fabricated to experimentally investigate the effect of applied magnetic field on the STL and natural frequency of sandwich panels having different thicknesses of MR core layer. The magnetic flux density generated inside the electromagnet is simulated using magneto-static finite element analysis and validated with the measured magnetic flux density using Gaussmeter. The results from the magneto-static analysis is used to derive approximate polynomial functions to evaluate the magnetic flux density as a function of the plate’s radius and applied current. 
In the second part, the sound transmission behavior of MR based-circular sandwich panels is investigated through development of efficient numerical models. The forced vibration equations of motion of the circular sandwich panel fully treated with MR fluid core layer is first derived utilizing Ritz and finite element (FE) methods using circular and annular elements. The transverse velocity in the transmitted side of the panel is then calculated and utilized to obtain the sound radiated from the panel and subsequently the STL. The theoretical models are validated comparing the simulation results with those obtained experimentally. The developed models have been subsequently used to conduct parametric studies in order to investigate the effect of the applied magnetic field, the thickness of the face sheets and the thickness of the MR core layer on the first axisymmetric natural frequency and STL of the MR based-clamped circular panels.
The last part of the present study is devoted to the topology optimization of sandwich panels partially treated with MR fluid and silicone rubber core layer. The FE model of the sandwich panel partially treated with MR fluid and silicone rubber has been developed using circular and 4-node quadrilateral elements. The developed model is then utilized to investigate the vibroacoustic behavior of MR-based sandwich panels and to obtain their natural frequencies, loss factors and STL. Subsequently, systematic parametric studies on the effect of the position of the MR fluid and silicone rubber segments on the first axisymmetric natural frequency, corresponding loss factor and also STL are presented. It has been shown that the vibrational and acoustical behavior of the sandwich panel considerably changes by varying the location of the MR fluid treatment segments. A formal constrained and unconstrained design optimization strategy have been subsequently formulated to identify the optimal location of the MR fluid segments. Due to high computational cost associated with the FE model and considering that in each optimization iteration, FE model requires to be executed several times, approximate meta-models have been developed using random and D-optimal design points to conduct optimization problems efficiently without using the full FE model. The developed meta-models are then utilized to solve the topology optimization problems using the genetic algorithm (GA) and integer programing (IP) problems. The suitability of the identified optimal candidates are further evaluated using the developed finite element model to determine the true optimized topologies for the constrained and unconstrained problems.