fluids, structural mechanics & acoustcs research areas
As an integral part of Hydroacoustics, Structural Acoustics specializes in vibro-acoustic characterization and noise diagnostics (in air or water) of complex structures using both computational and experimental analysis. Industry standard and cutting edge in-house generated computational tools are used to accurately model how a structure responds to fluid forces influencing it and conversely how the structure generates a response in the surrounding fluid. Click HERE for information about ARL's reverberant tank facility.
These fluid-structure interactions manifest themselves in measurable quantities such as radiated sound power, surface vibration, surface pressures, loss factors, and modal response. The Structural Acoustics Department, working closely with the Flow Acoustics Department, also has the capability to fabricate, instrument, and measure complex structures and verify the results using computational analysis. Once the predominant structural-acoustic behavior is modeled, verified, and understood, Noise Control Engineering techniques are incorporated to tailor structures to control the responses.
This approach has been successfully utilized by the Structural Acoustics Department on a variety of applications including such items as propellers, plates, hulls, enclosures, and other complex structures to minimize vibro-acoustic response.
ARL Penn State performs experimental and theoretical research in the areas flow acoustics, structural acoustics, noise control, and active noise cancellation. Researchers in this area have access to excellent facilities that provide opportunity to research many areas of the physics of flow-generated acoustic forcing functions and the structural response to these forcing functions.
The research focus is primarily in the design, analysis, and testing of turbomachinery systems. Other typical systems that have been and are being considered are automotive HVAC, computer and electronic systems, vacuum cleaners, and diesel engines. The experimental effort is enhanced with a developing computational effort in the areas of structural response of propulsion systems, radiation efficiency, and flow interaction. The computational effort concurrently incorporates technology developed in other research areas at ARL Penn State, namely, fluid dynamics, computational mechanics, and manufacturing science.
Boundary Element Models
We model the sound radiated by vibrating structures with our ARL developed Boundary Element (BE) modeling technique.
ARL's approach converts surface vibrations into lumped volume velocities, leading to a computationally efficient approach which does not have the 'forbidden frequency' problems that plague many other boundary element approaches.
Our software allows us to compute radiated sound fields at any frequency, like the sound from this loudspeaker.
Much of our research is conducted on structures submerged in water, like marine propellers.
The mass loading and radiation damping that water imparts on a structure can be computed over a wide range of frequencies with unlimited frequency resolution.
Compute complex fluid-loaded modes of submerged structures are also computed by combining our BE fluid loading matrices with structural FE models, as in the submerged cylindrical shell modes shown here.
Computationally Efficient BE Modeling
ARL's boundary element approach does not require dense acoustic meshes. When coupling a structural Finite Element model to a BE model, we often condense structural mesh regions into lumped acoustic elements, which saves considerable computer time.
The loudspeaker example shows two meshes. The coarser mesh is suitable for low frequency analysis, and the denser mesh for mid-frequency analysis.
Computational Capabilities: Bearing Dynamic Modeling
Bearings exist in all equipment with rotating components (pumps, gearboxes, motors, generators, fans, etc.). Two of the most common types of bearings are fluid film (journal, tilting pad, etc.) and rolling element (ball and roller). The accuracy of dynamic models of machinery with bearings often is highly dependent upon the accuracy of the bearing models to couple the rotating and stationary components.
|Fluid Film Bearings
We model fluid film bearings by solving the Reynolds equation for hydrodynamic lubrication theory and estimated bearing stiffness and damping characteristics using a standard perturbation approach.
Our approach to appling the bearing coefficients involves a distribution of the coefficients to their physical location such that they produce a physically accurate interaction between rotating and stationary components. The general practice in industry is to represent the bearing coupling as a lumped effect and is applied at a single location (and often for a single frequency) in the model which can lead to large errors for some types of equipment.
The approach is described in ASME - IMECE 2003/NCA-43770 (ADD LINK HERE TO: Distributed Journal Bearing Dynamic Coefficients for Structural Finite Element Models. Proceedings of 2003 ASME International Mechanical Engineering Congress and Exposition, November 15-21, 2003, Washington D.C.). The distributed coefficients are used in finite element and substructure synthesis modeling to couple rotor and stator components of motors, gearboxes, propulsors, and other rotary equipment.
Rolling Element Bearings
Rolling element bearings are modeled with a Hertzian contact model to compute dynamic stiffness coefficients that couple linear and rotational relative motions. Our physics-based modeling approach enables bearing stiffness estimation for any bearing preload condition. The general industry practice ignores preload effects and applies a constant bearing stiffness that is not direction- or preload-dependent.
Bearing stiffness matrices are used in finite element and substructure synthesis modeling to couple rotor and stator components of motors, gearboxes, propulsors, and other rotary equipment.
We model complex propulsors and other structures with Finite Elements (FE), and compute the modes of our models with various commercial FE software packages, like NASTRAN and ABAQUS.
Normal modes are the basis of the noise and vibration analysis capability, combining them with Boundary Element (BE) models of surrounding acoustic spaces, like air and water.
The approach is extremely efficient, allowing for the modeling of very large systems across a wide range of frequencies with precise frequency resolution.
When driving FE models with dynamic forces, Structural Intensity can be computed using in-house software, which shows how vibrational energy travels through a structure.
Structural intensity is used to:
The structural intensity patterns in an automobile body model are shown here, for a vertical drive on one of the shock towers (where road noise enters a vehicle). Orange colors show high intensity, and the vectors show how structure-borne sound travels through the vehicle. An automobile company used our structural intensity capability to redesign the shock tower in one of their vehicles, reducing road noise.
Component Mode Synthesis
For complex structural systems, Component Mode Synthesis is often used to model assembled dynamic behavior.
A system is broken into logical components, like the motor housing and rotor shown on the left. In this example, the modes of the housing and rotor are computed separately, and coupled via impedances at the forward and aft bearings in the housing.
State of the art tools are used to compute Bearing Impedances of journal and rolling element bearings, and one of our Graduate Students is measuring the impedances distributed around the fluid film in a journal bearing.
Statistical Energy Analysis (SEA)
Mid-to-high frequency noise problems are modeled using SEA. SEA techniques have been successfully applied to many industrial and research vibration and acoustics applications including ground, air, space and marine vehicles, industrial machinery and enclosures, architectural acoustics and many others. SEA is a statistical technique for analyzing energy sharing and flow between groups of modes in structural and acoustic subsystems. SEA models can be quickly developed early in a system or product design to evaluate material and configuration trade offs when design changes can be most efficiently implemented.
Some of the examples of the use of SEA techniques:
Industrial machinery noise control enclosure studies using analytical SEA models
Structural acoustics response simulations of satellite structures due to lift off acoustic loading using a hybrid analytical / computational (finite element – boundary element) model: