Zenotech and AlgoLib are excited to kick off a ground-breaking collaboration on the application of AI to meshing, a pervasive and time-consuming engineering simulation problem. Funded by Innovate UK, the AI Mesh project brings together Zenotech’s HPC and CFD knowledge with AlgoLib’s AI and machine learning expertise to create the first AI-based mesh generator for industrial use.
Mesh generation is one of the most commonly referenced problems in computational engineering. It defines the discrete elements, lines and points around the geometry of a product (e.g car, aeroplane or turbine design) that are required before a CFD solver can be run. The quality of the mesh is the key to the quality of the CFD solution, but currently for any given simulation, the optimal mesh is only known once the simulation has been generated, leading to inefficiencies.
This prototype AI-based meshing system aims to increase automation and improve the efficiency of simulation processes. With no other AI-based mesh generators in industrial use, this new digital product is set to become a disruptor in the market. The AI mesh generator will automate mesh generation informed by feedback from the quality metrics of each mesh generated. The data collected for feedback will increase in value over time and with every application of the software.
The technology will be developed and critically applied in aerospace, automotive, civil and renewables engineering.
Zenotech director and co-founder, David Standingford said:
“I’ve been predicting that machine learning would offer the next leap in meshing technology and therefore it is particularly exciting to be working with AlgoLib on an AI generated solution to this endemic issue in computational engineering. We look forward to bringing Zenotech’s expertise in modelling and simulation to the development of this new AI mesh prototype. It has the potential to become a genuine disruptor in the market.”
James Sharpe, Zenotech’s big data and security lead commented:
“I’m excited to be driving AI Mesh forward as project lead for Zenotech. This prototype demonstrates the need for an innovative system architecture to make sure we identify areas where AI technology really can make a difference, rather than just applying generic AI algorithms to large amounts of standard CFD data. AI Mesh is all about generating truly useful technology and applying its benefits to industry.”
AlgoLib’s Robert Poole-Mckenzie, lead software developer for the project, stated:
“The partnership between AlgoLib and Zenotech provides an exciting and unique collaboration bringing companies with different backgrounds together to work on this AI Mesh project. Zenotech’s vast knowledge and expertise in the CFD sector, along with AlgoLib’s machine learning and data science proficiency is the perfect combination in which to advance CFD via AI technology.”
This 12-month project started in September 2018, so stay tuned via the Zenotech website for updates.