NAFEMS DACH 2020 Conference from October 13 – 14, 2020 (Virtual Conference)
The 5th German speaking NAFEMS Regional Conference features on engineering analysis, modelling and simulation and its impact on industry and beyond. The Conference offers a unique, independent, neutral, overlapping and comprehensive information and networking offer. (Source: NAFEMS)
This year the event will take place as a virtual conference. Participation is open for NAFEMS members and non-members. NAFEMS members participate free of charge as part of their membership.
Mvoid is participating in the event with a presentation on “Multidimensional Optimization of Loudspeakers Based on Multiphysical Models”.
Presenter: A. Svobodnik
Authors: T. Nizzoli, T. Gmeiner, A. Svobodnik (Mvoid Group), A. Saratov, P. Maury (DATADVANCE)
Day, time: October 14 at 12:00 CET
Panel session 6-1: Multiphysik/NVH
Attendees will be able to ask questions to A. Svobodik via chat window and to communicate directly with the lecturer by e-mail.
A virtual prototype is suitable for optimizing features of products and systems in an early development phase, hence offering a profound decision basis for improvements and necessary development steps. The optimum product often means finding the best balance between interacting technical parameters to meet divergent requirements as good as possible. Especially when a large number of variable parameters faces a multitude of objective functions, the methodology of virtual product development becomes particularly effective. In order to create a sufficient functional digital counterpart of an acoustic product it is necessary to apply various physical regimes, resulting in the combination of different simulation methods, mesh sizes, mathematical models and physical theories.
In order to explore the design options, optimization models based on predictive modelling, machine learning and neuronal networks are applied. These methods are also used to improve products in complex multidisciplinary contexts, subject to strong operational constraints and reliability requirements. Approximation models are generated for all relevant physical disciplines. These mathematical models vary the parameters from the virtual prototype and thus generate the surrogate model.
A. Svobodnik will demonstrate the methodology using a loudspeaker that operates in the non-linear regime and is optimized for multiple objective functions. A data link between the surrogate model and the virtual prototype was developed, automating the iteration process.
For multi-dimensional optimization of loudspeakers, Mvoid has integrated Datadvance’s pSeven software into the Mvoid methodology.
Link to NAFEMS 20 DACH
Datadvance is an independent software developer that offers its customers’ software solutions and consulting services for advanced data analysis, predictive modeling and design optimization.