Multiphysics vibroacoustic simulation has emerged as an integral element of electric vehicle (EV) powertrain development, providing quieter and more refined drivetrains with minimal physical prototyping requirements. An in-depth NVH analysis may involve various fields including electromagnetic motor forces, contact mechanics for gears and bearings, flexible multibody dynamics analysis, structural FEM modelling and acoustics.
High-fidelity time-domain models featuring flexible bodies and nonlinear contacts offer an acceptable balance between accuracy and computational cost, while frequency-domain methods provide faster evaluation for design screening and iteration, but can suffer from transient phenomena, tonal responses and nonlinearities that inhibit their use.
What is Vibro-Acoustic Simulation?
Vibro-acoustic simulation can ensure noise and vibration performance is built into products from their early design stages, prior to any physical prototypes being created. This helps lower development costs and decreases risk by early identification of performance issues – and prevent last-minute “band-aid” fixes being costly and time consuming fixes which delay schedule completion.
Vibro-acoustic simulation involves solving the governing equations of vibro-acoustics using techniques such as finite element analysis. These equations include stiffness and mass matrices for both solid and acoustic domains, source terms p, v, and q as well as coupling term Csa; all assembled into Equation 11 which is then solved frequency domain using Helmholtz equation.
This white paper explores the application of open-source FEniCSx to vibroacoustic problems, illustrating its ability to produce accurate and stable models of coupled fluid-structure systems containing nonconformal meshes at their interface. Furthermore, this approach provides an effective alternative to partitioned approaches when dealing with geometrically complex models.
How Can I Use Vibro-Acoustic Simulation?
Vibro-acoustic simulation can reduce development costs while ensuring noise and vibration performance is integrated into products from an early design stage, eliminating risks such as discovering noise/vibration issues too late in the cycle, where last-minute “band-aid” fixes can be costly, non-optimal and introduce schedule delays. Furthermore, engineers can identify dominant transmission paths between sources and receivers to minimize efforts required to mitigate non-dominant paths; as well as reduce transmission loss by optimizing structures, damping, coating etc.
To accurately simulate the acoustics of technical systems, various mathematical approaches must be employed. When dealing with subsystems with linear measures corresponding to wavelength, finite element modeling (FEM) and statistical energy analysis (SEA) are both suitable approaches; when dealing with mid-frequency problems however, hybrid FEM/SEA approaches often prove more suitable; both techniques complement each other well enough that when combined they capture all possible frequency ranges within complex technical systems.
Multiphysics vibro-acoustic simulation has emerged as an indispensable tool in designing quieter electric vehicle (EV) powertrains by uncovering otherwise hidden excitations. Yet its full predictive potential remains limited until more integration between simulation and experiment, more explicit treatment of uncertainty, greater transparency across tools, data sets, and simulation workflows, and improved standardization can be realized.
In an electric vehicle powertrain, gear mesh and bearing forces excite structural modes in the housing that radiate sound. Engineers use methodologies such as Equivalent Radiated Power (ERP), which estimates maximum possible noise from housing surface motion; or advanced workflows that integrate computed housed mode velocities directly into an acoustic solver to produce far-field SPL maps [5]. To predict this radiated noise accurately, engineers employ Equivalent Radiated Power (ERP), an estimation tool designed to approximate maximum possible noise from housing surface motion; or advanced workflows that take computed housed mode velocities directly into an acoustic solver for far-field SPL mapping [5].
Multiphysics models are complex and computationally intensive; to effectively deploy them requires adopting efficient modeling approaches and software frameworks capable of running on high-performance computing platforms. This paper investigates the application of open-source FEniCSx multiphysics modeling approach in practical vibroacoustic challenges, with an emphasis on its flexibility, efficiency and scalability for simulating coupled acoustic-structural models.
What are the Benefits of Vibro-Acoustic Simulation?
Electric vehicles have revolutionized noise, vibration and harshness (NVH) requirements. Their absence exposes previously-masked drivetrain excitations that were once hidden by internal combustion engine noise; making vibro-acoustic simulation an integral component of developing electric powertrains. Vibro-acoustic modeling and prediction require multiphysics simulation, including electromagnetic analysis, gear contact mechanics, multibody dynamics simulations, structural FEM/BEM solvers as well as acoustic FEM/BEM solvers; this literature review comprehensively analyses these technologies in order to build one coherent framework linking end-to-end forces response responses with structure responses as well as radiated noise sources.
Vibro-acoustic simulation can significantly lower NVH development costs by enabling engineers to design for optimal acoustics early in product development, before physical prototypes are even created. Furthermore, this tool identifies dominant transmission paths between sources and receivers so designers can focus on mass-reducing non-dominant paths instead. Consequently, this can prevent last-minute fixes that require prototyping, rework or schedule delays – saving engineers both money and time in developing products with optimal acoustics from occurring last-minute band-aid fixes that require prototyping costs or rework or schedule delays.
Immersive virtual acoustics simulation can also make it simpler to demonstrate a product’s acoustic performance to stakeholders and customers, using 3D visual rendering and VR head tracking. By eliminating physical prototypes from simulation processes, simulation processes can be accelerated faster with less rework required between iterations rounds.
Vibro-acoustic simulation can also assist in predicting aerodynamic noise caused by high-speed rotating parts such as cooling fans and axle air ducts, such as cooling fans or axle air ducts. This is accomplished through coupling CFD with acoustics using techniques such as LES in Fluent or SNGR in Actran.
Vibro-acoustic simulation is made more understandable when we recognize that mechanical vibrations may either be deterministic or random depending on the size of their wavelengths in relation to subsystem dimensions. Therefore, for deterministic systems finite element methods should be employed while statistical energy analysis (SEA) should be utilized when dealing with random ones.
What are the Drawbacks of Vibro-Acoustic Simulation?
Noise and vibration simulation is an invaluable way to incorporate acoustic performance early in the design cycle, eliminating or minimizing the need for physical prototypes and costly tests. Furthermore, noise simulation allows you to identify any potential acoustic problems before production commences; saving both time and money.
Vibro-acoustic simulation is an engineering practice which uses models and computational methods to predict acoustic behavior of technical systems. This field encompasses both audible frequency range and all components within an overall system – as well as structural and fluid aspects of acoustics. As wave motion in structures and fluids varies greatly, acoustic modeling must employ different strategies for each subsystem type. Example: When subsystem dimensions are small relative to wavelength, a subsystem is known as deterministic and may be treated using finite element method (FEM). When these same dimensions increase significantly over time, however, the system becomes random and requires statistical energy analysis (SEA).
Advanced multiphysics workflows use FEM/BEM and SEA modeling methods to simulate structure-borne and airborne sound, with electric powertrain NVH using electromagnetic motor forces and gear meshing dynamics as excitation sources that generate structure-borne vibrations and radiate noise, with transmission path modeling methods used to predict actual sound pressure levels at microphones – giving engineers confidence in making design decisions with greater ease and accelerating development time. This end-to-end approach to eNVH empowers engineers with making design decisions with confidence while cutting development time in half!
Aeroacoustic simulation is another aspect of vibro-acoustics that uses Computational Fluid Dynamics (CFD) to model turbulent flow and extract sound radiation, which can then be used to predict noise in electric motors or axle air ducts for example. When combined with vibro-acoustics, aeroacoustic simulation can reduce dependence on wind tunnels – which are often too expensive and unavailable for testing new prototypes.
MSC Actran offers both SEA and aeroacoustic modules, enabling you to simulate the entire acoustic lifecycle using one user interface. Actran SNGR takes input from CFD flow data and calculates turbulent pressure distribution and noise radiation of complex, rotating systems; Actran Aeroacoustics then integrates those predictions with structural models and BEM solvers to predict coupling and radiation of soundwaves.
