Leading Artificially Intelligent Engineering (AIE) software company Monolith AI and Imperial College London will use a £500,000 Innovate UK grant to launch a project to build a new type of artificial intelligence solution that will be able to assess if metal components are manufacturable and even explain why they might not be manufacturable.
While CAE simulations have dramatically improved how the manufacturability of industrial products like for example the door of a car is assessed, the final decision on whether this door is actually manufacturable is still limited to the assessment of a small number of domain experts who need to assess the simulation results.
We know from medical image processing used, for example, to diagnose lung cancer, that machine learning models can provide a highly accurate assessment for new patients when trained with enough data. However, knowing the final prediction is not enough. Doctors need to understand why and where the AI algorithm detected cancer. The same is true for engineering applications.
“Knowing that a door is not manufacturable is not enough. You need to understand why, and even more importantly how you could change the design and operating conditions to make it manufacturable”, says project leader Dr Joël Henry.
Therefore, the goal is to build a new version of explainable AI that will provide clear feedback to engineers on how it arrived at its conclusions, removing the ‘black box’ dilemma.
Monolith AI and Imperial College London want to streamline the manufacturing process and provide a new competitive advantage to high volume manufacturers by using AI to learn from what could be manufactured in the past and predict what is best for new components, enabling engineers to build expert simulations based on repetitive tasks and historic data.
The platform will be developed over the course of 18 months, and will be evaluated by multiple industry partners. If successful, the project is set to revolutionise Computer-Aided Engineering (CAE) within the manufacturing industry, allowing engineers to run complex manufacturability assessment in seconds compared to weeks freeing up engineering skills for the most complex, value-added tasks.
Dr Richard Ahlfeld, CEO and founder of Monolith AI, comments: “CAE has done a fantastic job advancing component manufacturing, but there are still many areas where physical simulations still cannot capture the true complexity of components. Large engineering companies collect a lot of data when assessing manufacturability and our goal is to make that data work to their advantage. This latest funding will allow us to explore this possibility and drive not only the automotive industry, but other sectors forward.”