After decades of false starts, Artificial Intelligence (AI) is finally pervasive in our lives, and our businesses. Kiran Krishnamurthy, AI Domain Specialist, CFMS, explains how this new wave of technology can improve workflows, save time and cost, and help you leverage your in-house data.
As manufacturing becomes more about brains and less about muscles, AI is transforming high-value manufacturing and high-value design industries with its data-driven decision-making capabilities. When Professor Andrew Ng — former Baidu Chief Scientist, Coursera co-founder, and Stanford Adjunct Professor — gave a talk on why ‘AI is the new Electricity’, he emphasised the transformative nature of the technology. And as this recent wave of AI begins to cause a cultural shift from ‘know-it-all’ to ‘learn-it-all’ across many diverse industries like aerospace and healthcare, the reality is that enterprises that don’t go through this transformation may perish.
Of course, embracing AI is not as simple as deploying the relevant technologies, or adapting the working culture – not that either of those tasks can truly be defined as simple. An initial challenge is the gap that exists between the worlds of high-value manufacturing and design, and AI. It is this gap that prevents the seamless transition to, and application of, AI technologies. At this point it’s important to provide a distinction between evolution, and change. Many companies will change over time, but evolution requires the adoption of new technologies and processes in a more intrinsic way.
The key to competitiveness
Within the high-value manufacturing and design space, evolution is to a certain extent occurring naturally. Both efficiency and effectiveness of traditional design and manufacturing processes has plateaued, and as pointed out in McKinsey’s Industry 4.0 report, traditional productivity levers within Manufacturing (e.g. lean methodologies, and outsourcing, etc.) are widely exhausted. This means that time-to-market and customer responsiveness are now key to competitiveness. So where does AI fit in?
To unlock the value of data, companies must be able to collect, analyse and act on traceable, validated information. Within the current landscape that information exists in a one-way flow of information that leads from design through manufacturing and to maintenance. To be able to almost instantly learn from the data generated along the product lifecycle, and feed that information back through the design and development process in a circular manner, would have significant benefits.
AI provides a natural fit to the sensor-filled manufacturing industry, and industrial AI enables the value in data from equipment and sensors to be harnessed, enabling intelligent predictions, and the automation of operational decision-making. In essence, a digital twin mirrors the entire production process including machines, lines, and plants, and serves as the foundational layer for enabling AI-driven automated decision making on the factory floor.
The integration of predictive machine learning
By combining early stage manufacturing analysis activity with a type of AI model known as machine learning, which leverages statistical techniques to ‘learn’ and make predictions, companies can benefit from accelerated time to market and process maturity. This has been demonstrated by CLAMPS (Computer Learning in Automated Manufacturing Processes), a collaborative project between the Centre for Modelling & Simulation (CFMS) and the National Composites Centre (NCC). The latter brings composites manufacturing domain knowledge to the project, while the former is an independent, not-for-profit specialist in high value design and manufacturing capability, underpinned by advanced modelling and simulation, high-performance computing, and artificial intelligence capabilities. Together, both organisations aim to demonstrate that the integration of predictive machine learning will drive improvements and minimise variability within a composites manufacturing process.
Highlighting the digitalisation and automation steps necessary for incorporation into a liquid composite moulding process, the project will result in a physical demonstrator which will automatically adjust the manufacturing process parameters to ensure consistently high-quality parts. This will reduce the need for rework, scrap or repair, ultimately saving costs at the NCC. Because the project deals with preliminary manufacturing, it didn’t have the high volume manufacturing data necessary to make best use of AI. In order to meet the data requirement, the ability to detect and control defect formation will be based on the output of more than 15,000 virtual manufacturing simulations of a liquid composite resin infusion process. The simulations themselves are good but computationally heavy and so a machine learning algorithm has been used to process the data and classify the decisions made during the infusion process in order to minimise the formation of porosity and dry-spot defects.
In the physical manufacturing process, the flow of resin is monitored using intelligently positioned sensors, enabling a real-time understanding of the flow inside a closed mould. The machine learning model will be used to predict and mitigate the formation of defects by selectively opening and closing inlet and outlet valves to influence the flow of resin. Instead of tens of thousands of simulations, which can take up to 48 hours to complete, predictions as to the quality of the resin coverage can be made by the machine learning model within a fraction of a second. This pilot project will act as a proof-of-principle for the application of machine learning to control a composites manufacturing process, with the potential of being upscaled to complex composite part geometries for industries including aerospace, automotive and renewables.
Liberating engineers; driving innovation
AI-driven automation will enable the high-value manufacturing and design industry to reach a higher level of accuracy and productivity – a level that is currently beyond human ability. By liberating engineers from somewhat mundane repetitive activities, AI will free more time for innovative and creative tasks. In order to accelerate and facilitate the adoption of these technologies, CFMS is using its position as an independent, not-for-profit at the heart of the design and manufacturing community.
The collaborative and innovative nature of AI is intrinsic to CFMS, which fosters collaborative innovation within many industries, between enterprises of varying sizes, and within academia. CFMS is able to offer a range of skills, expertise and access to technologies that help enterprises go through the current journey of technological transformation in a seamless way; to first help them identify opportunities for AI-related topologies within their own businesses, and to do so with minimum impact and maximum benefit. With minimal capital investment, there’s no need to change existing set-ups> Now is the time to let clever algorithms harvest business data and unlock its value in order to accelerate learning, improve processes, and reap the benefits of AI. If you are interested in discovering how AI technologies can work for you, contact CFMS.