Few technology trends can truly be described as a paradigm shift – but the introduction of artificial intelligence (AI) is one of them.
Manufacturing needs to tackle its carbon footprint – and AI could help
“Manufacturing has a significant carbon footprint, and this is not compatible with our net zero aspirations,” said Dr Becky Bolton, knowledge transfer manager at Innovate UK Business Connect. “We also need to keep up with the technological advancements that are ever-prevalent in a digital world – but it’s not always easy for traditional manufacturers to adopt a new technology.”
AI could play a significant role in decarbonising manufacturing, claimed Şahin Çağlayan, co-founder and CEO of Faradai. “We need a dynamic, continuous decarbonisation intelligence,” he said. “Large language models, generative AI are popular in a lot of different verticals, but still in manufacturing, there are gaps and there are trust issues.”
There is a “missing link between carbon, energy, ESG (environmental, social and governance) and finance, data, operation, so it needs to be aligned all together and with a wider, comprehensive approach.”
Faradai aims to solve those challenges with its integrated intelligence platform, which it claims can cut carbon footprints by 20%. Approaches include “multi-objective optimisation”, which looks at the inputs and outputs of an entire production line to produce ‘heat maps’ of energy use and consumption of other resources. Results from the platform can then point out potential optimisations or changes to reduce consumption.
Extensive use of AI can result in increased emissions, however, so any deployment will need to be carefully considered to ensure it does not undermine its own purpose.
Caution needed
Neural networks can handle “very complex situations”, said past IMechE president Peter Flinn, spotting patterns in very large and complicated sets of data.
The downside is that the logic is difficult to trace, he said. “Hence it may not be repeatable – and therefore I think its usefulness, for example in safety-critical systems or situations, is going to take a bit of thinking about.
“And there are other concerns… I’m hearing people now expressing concern that putting organisations’ own data into large language models may compromise its safety and confidentiality.
“But my own view is that AI will definitely proceed to form an important part of engineering, albeit with some caution. So if you want to get to develop your competence in artificial intelligence, engineering is a very good field in which you can do that in a practical way.”
Flexibility is key
Many manufacturers know that AI will shape their future, said Dr Alina Ivanova, chief product officer and co-founder of Turation, but most are unsure of how and where to start.
“Today, manufacturing productivity is defined by the intelligence and flexibility of production,” she said. Low-mix, high-volume production already has minimal human involvement and high machine involvement, she said, but high-mix, low-volume production needs significant human contribution due to the complexity and adaptability required, tasks that are challenging to automate.
AI can help by aiding knowledge transfer and training new staff, mitigating variability and inconsistency in production. It could also help fully automate some tasks, Dr Ivanova claimed, but it would need to be flexible and adaptable to “unforeseen production scenarios”.
Turation hopes to enable that with its AI systems, which are trained to adapt to new use cases with little customer input, aiding small batch production. An adaptable system for advanced component manufacturing can detect anomalies and optimise operations through visual inspection, autonomously adjusting to disruption or to minimise production time.