Subaru, Honda and Neural Concept Showcase AI-Powered Engineering Innovations at Cybernet Seminar Tokyo
27th October: This week we attended the Cybernet Seminar in Tokyo along with 250 other experts to explore how the automotive industry can leverage computer-aided engineering (CAE) and artificial intelligence (AI) to enhance and streamline design processes for customers.
The demand of AI for CAE was clear - this innovation will reduce product development times and costs, and as such, will be key to reducing manual errors while being able to develop better quality products at scale.
Subaru introduced "Prediction of formability of press parts for automobiles using Neural Concept Shape", while Honda presented "Introducing Honda's automobile development process that utilizes simulation in the upstream development process and the potential of surrogate AI".
Introducing Engineering Intelligence
Our CEO, Pierre Baqué and Laurent D’Alvise, Commercial Director spoke at the seminar about how Neural Concept uses Engineering Intelligence to develop more efficient, higher performance products for automotive OEMs and suppliers.
At Neural Concept, we use non-parametric, physics-based surrogates which run on deep learning, 3D generation, and large-scale language models. By learning from these designs, our Neural Concept Shape (NCS) platform has a clear impact on any optimisation capabilities within the design process.
This fundamentals of deep learning and 3D generation modelling has caused a paradigm shift, where we are now seeing significant reduction of development times. As a company, we are now looking to take this shift to an unprecedented new level. Neural Concept Shape is a software platform that allows advanced engineering organisations to use these aforementioned functions to guide designers faster towards radically greater concepts..
Pierre’s presentation spoke in detail about a series of practical success stories that affect businesses within the automotive industry; be it external aerodynamics to heat exchanger design, and subframe component optimisation, there are many instances where embracing Engineering Intelligence is working wonders for many of our automotive OEMs and suppliers.
Demand for the industry
The automotive industry right now is in a position where the demand for development and innovation is ever-increasing, but in the same breath, development time is required to be shorter and more efficient – thus creating a situation where the need to utilise the latest digital engineering technology is of paramount importance.
It was a huge privilege to listen to a lecture hosted by Mr. Sakatu Suke from Subaru Corporation, who have been using Neural Concept Shape to predict the formability of press parts for automobiles.
He spoke of how higher quality and shorter delivery times are now being required in development and manufacturing of automotive press parts, and how NCS can create a Surrogate AI which learns from previous Finite Element Analysis (FEM) and Computer-Aided Design (CAD) shapes.
In doing so, these FEM and CAD shapes can predict new design shapes at high speed, and part of Mr Suke’s speech explained how our NCS platform was used to predict press moulding in comparison to the FEM method.
Analysis and design using computer-aided engineering has been widely carried out, but more recently, methods which combine AI technologies – such as surrogate AI – have been gaining traction, and this is expected to further improve quality, shorten development lead time and reduce costs.
The advantage of utilising a surrogate AI, as we’ve seen with Honda’s automobile development, is that you can create a model based off of a simulation and then apply this during the design or engineering phase of development. Surrogate AI obtains its training data by probing the simulation outputs at several intelligently chosen places within the design parameter space. Each of these undergoes a comprehensive simulation to compute the associated simulation outcome.
Discover the conference's full agenda.
A special thank you to our partner Cybernet for hosting this amazing event.
And so, what about our NCS platform? Our methods of Engineering Intelligence allow us to develop with greater efficiency and performance – but let’s delve a little deeper into what NCS can really do.
Neural Concept Shape, ultimately, places deep learning at the core of an engineering toolkit, as it is a modulable platform which is built on core deep learning algorithms, tailored for hybrid geometry generative models and physics surrogates.
Engineering teams are aware of the importance of Engineering Intelligence, but many lack the capabilities to make their core technology perform on these complex processes to be able to deploy it at scale.
Our open-platform NCS simplifies the process, and by understanding and integrating Engineering Intelligence, organisations can not only meet the demands of today, but also shape the future of design engineering tomorrow.
If you would like to find out how Neural Concept Shape can work for your company, click here to watch our video demonstration, or alternatively book in a meeting with Lorenz Frey, our Technical Business Developer by clicking the link below: