An AI specialist has developed a cloud-based enterprise platform that generates actionable insight from complex manufacturing process analysis to empower engineers to improve productivity and accelerate net zero goals.
Intellium AI, based in Bristol, partnered with GKN Aerospace, a leading manufacturer of aerostructures and engine systems, based in Bristol, as part of Project Butterfly, a consortium of high value and fast-moving consumer goods (FMCG) companies, technology solution providers and research organisations, funded by Made Smarter Innovation’s Sustainable Smart Factory programme.
The R&D project resulted in BoostBot, an innovative Generative AI enterprise platform, which streamlines and automates various tasks, translating complex manufacturing data into conversational English to help operators identify opportunities to optimise processes, backed up data-driven insights.
Chris Needham, Innovation Lead in the Made Smarter innovation challenge said: “Applying effective digital technologies can have a substantial impact on the manufacturing sector, bringing outdated, inefficient and unproductive products and processes up the standards needed for a net zero industry of the future.
“Intellium AI has clearly demonstrated industry insight and innovation can leverage data and make a significant difference to manufacturing business, while developing a skilled, digital-savvy and engaged workforce.”
The Inspiration
If UK manufacturing is to do its bit to achieve the country’s ambitious net zero targets by 2040, then it requires innovation and collaboration. This means pushing the boundaries of how we apply digital technologies and tools to outdated, inefficient and unproductive products and processes to drive down energy consumption and emissions, as well as optimising material consumption.
AI and Machine Learning software has emerged as a potential solution to help companies meet sustainability targets by optimising resource use, analysing data to understand relationship between machine settings, part dimensions, and energy used, supporting net zero without compromising quality or speed.
But adopting and exploiting the full potential of AI to gain a competitive edge is difficult given the shortage of industry relevant AI skills, perceived complexity, not knowing where to start and AI hype.
Kiran Krishnamurthy, CEO of Intellium, worked at Airbus UK for 18 years helping transform its manufacturing operations with AI.
“In my experience, the trickiest bit of any digital transformation project is upskilling the workforce and changing the culture,” he said. “Engineers and operators on a factory shop floor want answers to questions, not another big confusing dashboard they have to interpret.”
The Innovation
With that in mind, Intellium AI worked with GKN Aerospace to develop and test a novel AI solution which uses advanced predictive modelling, intelligent process optimisation, real-time monitoring and what-if analysis to identify opportunities to reduce the company’s resource and energy consumption, while empowering and upskilling its workforce.
The first step was to capture, cleanse, analyse and visualise the key real-time and historical data. Then by applying sophisticated machine learning techniques, Intellium AI was able to learn and then make predictions of what could happen, but also provide prescriptive action of what needs to be done to ensure it does.
Intellium AI then developed an interface, a web-based tool to enable an engineer or operator to ask a question in plain English, and receive an accurate, verifiable data-supported answer with a list of actions to give them the insights they require, without writing a single line of code.
Features of BoostBot include: Audit trail to provide complete transparency of AI’s decision making process. This feature is key to developing user trust in AI systems. There could be regulations in the future requiring such features to aid human understanding of AI solutions. Additional Explainable AI features include ‘Answer Assurance’ enabling ‘fact checking’ of AI’s response, further aiding in building trust in BoostBot.
BoostBot was tested on two GKN Aerospace’s use cases: aircraft window coating and ribs cycle time analysis.
To ensure aircraft pilots have clear visibility in various atmospheric conditions, cockpit windows require a thin layer of gold coating to enable demisting through controlled heating. However, the process struggled to ensure coverage and thickness of the coating was consistent.
Aircraft wing ribs, critical components which ensure the structural integrity and aerodynamic efficiency, are manufactured through a subtractive process, removing material from metal billets, However, GKN experienced huge variations in manufacturing cycle times and due to complex manufacturing process it was unclear as to why. Meanwhile, inconsistent production schedules resulted in difficulties in capacity planning and meeting delivery timelines.
The Impact
By applying AI to the aircraft window coating process, Intellium AI was able to identify key process parameters driving coating consistency, increase resource efficiency in the quality inspection process, reduce rejection rates and improve overall product quality. The solution reduced CO2e emissions from the manufacturing process by 5%, the equivalent of 20,000kg per year.
For the ribs cycle time analysis, AI identified key process parameters driving milling cycle times and identified the setup parameters and machines causing the variations. The result was that GKN was able to increase its overall REEE in the process by 5%, reducing CO2e by 13,717kg per year.
BoostBot is now being used by GKN for the two processes and talks are underway about developing the software for other manufacturing processes.
Intellium AI is now demonstrating its solution to other companies in aerospace, and other sectors including naval shipyards.
As a business, Intellium AI has expanded its team by two to 15 as part of its growth strategy.
Kiran said: “BoostBot puts AI in the hands of engineers and operators as another tool in their toolkit, effectively turning them into citizen data scientists.
“Our solution accelerates the adoption of AI across a range of businesses, from large OEMs to SMEs, by delivering value through democratised analysis of IoT data related to cycle times and manufacturing process parameters. This approach is representative of processes typical in both large and small-scale manufacturing environments, ensuring that AI is accessible and impactful regardless of business size. The benefits are extensive, including accelerated product design cycles, faster manufacturing maturity, reduced manufacturing costs, lower operational expenses, increased productivity, and quicker time to market.
“It has been a fantastic experience to collaborate with GKN Aerospace on this project and for everyone to experience such positive outcomes. It has clearly demonstrated what can be achieved by developing cutting-edge technology with people at the heart of the solution. By marrying up data-led and people-led digital transformation UK manufacturing can tackle many of its core challenges such as net zero, productivity and the digital skills challenge.”
