Skip to content
Close
Talk to an expert
Japanese
Talk to an expert
Japanese

Accelerating nonlinear MEMS simulations with the harmonic balance method

See how harmonic balance is leveraged for solving nonlinear periodic problems in frequency domain for quicker, more precise results without transient analysis.

Watch recording
Download pdf summary

Everything you need to know about MultiphysicsAI

This episode contains everything you need to know about Quanscient MultiphysicsAI—straight from the CEO and co-founder Juha Riippi.

Watch the full episode
Read the key takeaways

Introduction to the guest

Juha Riippi, CEO and co-founder, has overseen the development of Quanscient’s cloud-based, natively coupled multiphysics platform and our work applying AI and machine learning to engineering simulation. His role gives him a unique perspective on how simulation is evolving—and what teams need next. 

In this episode...

In this episode, we cover the fundamental questions of Quanscient MultiphysicsAI: what is it, what does it do, who is it for, how does it differ from the other AI-driven simulation approaches out there, and how you can start using it today. Overall, this episode contains everything you need to know about the approach now and where it’s headed next.

Key takeaways

What Multiphysics AI is

Surrogate models trained on large volumes of Allsolve simulation data for design exploration, inverse problems, and optimization.
How the workflow changes

Shift from “design → validate a few sims” to “automate many sims → train model → search for designs that meet KPIs → validate finalists.”
Why it works

High simulation throughput provides the training data; simple ML architectures can perform well when fed strong, coupled physics data.
Always validate

AI is probabilistic. Use confidence metrics and then confirm shortlisted designs with deterministic simulations.
Example applications

Pareto fronts for sensitivity vs. bandwidth, piezoelectric material/property characterization, electric motor shape optimization, microspeaker distortion/output balancing, MEMS devices.
Current limits

Best for problems with a manageable number of inputs (≈ up to a dozen), clear KPIs, and accessible automated simulation pipelines.
Near-term roadmap

GUI flow to pick data, run training, and deploy surrogates inside Allsolve; pre-trained models as starting points.
Trust and bounds

Use models within the domain of the training data; treat extrapolation carefully and verify.
Quantum angle (forward-look)
Combine a little high-fidelity quantum simulation data with large classical datasets to improve training quality over time.
Who benefits today
Teams working on MEMS/ultrasound, microspeakers, motor design, chip packaging, RF/antenna, and similar optimization-heavy problems.
How to get started
Get in touch via us through our website and let's talk about how MultiphysicsAI could work for you.

Subtitles are available.

Listen to the full episode on YouTube

0:00 Introduction
1:03 What is MultiphysicsAI?
2:25 How does it work?
5:43 Applications and use cases
7:31 Difference from competitors
9:21 Concern about AI predictions
11:21 What will change going forward?
15:28 The role of quantum computing
18:29 Who can benefit of MultiphysicsAI today?
19:56 How to get started?
20:58 Final thoughts
22:15 Outro

harmonic-balance-2-1

Learn more about MultiphysicsAI

Read more and find all resources on our MultiphysicsAI webpage.
Open webpage
Resources

Learn more about the topics

Webpage

The official MultiphysicsAI webpage

Read more about the process and find all the resources on the MultiphysicsAI webpage. Use the contact formto start a conversation.

Read the webpage →

Launch event

Launch event recording and speaker slides

Replay the launch event where our CEO, Juha Riippi, and co-CTO, Dr. Andrew Tweedie, unveiled MultiphysicsAI to the world.

Open the event →

White paper

Accelerating engineering design with neural surrogates

Learn the 5 key stages for accelerating engineering work with neural surrogates.

Open the white paper →