Industry Sector RTD Thematic Area Date Deliverable Nr Land Transport & Aerospace Multi-Physics 13-Nov-01 Finite Element & Boundary Element Technology in Acoustics & Structural Dynamics : Current Status & Key Trends for the Future Ir. Peter SEGAERT LMS International - Leuven, Belgium Summary Overview of current FEM/BEM based simulation technology in Acoustics & Structural Dynamics, including obstacles to efficient use and key trends to satisfy needs of automotive & aerospace industries
1st FENET Workshop Wiesbaden [Germany] 13-14 November 2001 Finite Element & Boundary Element Technology in Acoustics & Structural Dynamics : Current Status & Key Trends for the Future Ir. Peter SEGAERT Product Support Manager - CAE VibroAcoustic Modeling Group LMS International [Leuven, Belgium] www.lmsintl.com
LMS & FENET Land Transport & Aerospace Industry Sector LMS develops, markets and supports CAE tools in the field of numerical vibro-acoustics based on FEM/BEM technology: Acoustics => LMS SYSNOISE Structural Dynamics => LMS SYSNOISE & LMS GATEWAY Durability & Fatigue => LMS FALANCS Motion => LMS DADS Acoustics, Durability & Motion => LMS VIRTUAL.LAB (next generation software) Main application fields of these tools are within R&D and Engineering departments of Automotive industry OEMs : car, truck & bus Automotive industry tier 1 component suppliers Aerospace industry : aircraft & satellite manufacturers General perspective from the software industry on Current practice of FEM/BEM vibro-acoustic codes in industry Obstacles to effective use of FEM/BEM technology Industry needs & key trends to satisfy these needs in the future
What is Numerical Vibro-Acoustics? Source vibrating body speaker Propagation sound path & absorption O airborne O structure-borne O mixed Receiver microphone ear
Why is Vibro-Acoustic Simulation becoming more important? Noise reduction and sound quality have become major issues Important contribution to perceived product quality Delivering low-noise level products creates a competitive advantage Government regulations constantly impose lower noise levels in the quest for a cleaner environment Analysis can be performed at several levels in the design cycle... Explaining and understanding the physics behind it Troubleshooting Evaluate design alternatives & reduce costly physical prototyping
Tools for Numerical Vibro-Acoustics (Semi-) Analytical Methods Closed form solutions, only for simple geometries Finite Element Method (LMS SYSNOISE/FEM) Volume discretization of fluid region Boundary Element Method (LMS SYSNOISE/BEM) Discretization of bounding surface only Statistical Energy Methods Energy exchanges between system components LMS SEADS Ray Tracing Methods Geometrical Acoustics LMS RAYNOISE
Example : SYSNOISE Modular Architecture FEM & BEM Methods Uncoupled & Coupled Transient Struct. FEM Struct. FEM Acoustic FEM I-FEM Harmonic Harmonic DBEM IBEM Struct. FEM Struct. FEM Transient Kernel
Numerical Vibro-Acoustics : Current Industrial Practice of FEM/BEM-based Codes Applied in R&D or Advanced Engineering depts : up-front troubleshooting what if experiments = virtual prototyping to reduce amount of physical prototypes gainining insight in the physics beyond analytical models In general, not applied in everyday use as mainstream design codes Typically used for : component-level analyses not for system-level analyses e.g. engine block radiation, oilpan radiation, exhaust radiation, but not complete powertrain, etc Typically lower frequency range or low modal density ranges (from a few hundred up to 2-3 KHz)
Numerical Vibro-Acoustics : Current Obstacles for More Efficient Use Large computational effort involved large matrices, especially in BEM (dense matrices), requiring large amounts of RAM each analysis frequency forms an independent problem => analysis over a large frequency range can take several days of solving time High-frequency limit Maximum analysis frequency linked to mesh refinement (λ/6 rule), leading to very large meshes for high frequencies Control of numerical error propagation at high frequencies Uncertainty with respect to impedance properties of acoustic damping materials Intake model Courtesy of Allied Signal for many materials, impedance is not known as a function of frequency
Numerical Vibro-Acoustics : Key Trends to satisfy Auto & Aerospace Industry Needs (1) Reducing computational efforts through novel techniques Acoustic Transfer Vector [ATV] concept LMS patent improving solver technology => importance of fundamental research at universities + dissemination by FENET Reducing computational efforts by exploiting networked computers as virtual parallel machines frequency level distribution for vibro-acoustic problems : multiple processors solve simultaneously for different frequencies => already implemented in LMS SYSNOISE introduction of domain decomposition techniques or distributed domain solvers : each processor resolves part of the solution domain, with appropriate continuity conditions at the sub-domain interfaces research work on improving parallel algorithms => role of fundamental research
Numerical Vibro-Acoustics : Key Trends to satisfy Auto & Aerospace Industry Needs (2) FE analyst user profile is changing over time! from FE analysis specialist to design engineer doing analysis => role of academia and organizations like NAFEMS & FENET to provide continuous education & grounding in best practices from central computer in batch to interactive UNIX workstation to Windowsbased PC platform user => role of industry to provide adapted software tools importance of proper initial training & continuous updating of knowledge cannot be overestimated => role of FENET! Most significant technical trend in numerical vibro-acoustics is the move towards aero-acoustics : flow-induced noise & fluid-structure interaction interfacing & coupling of Vibro-Acoustic simulation software with CFD software