Revolutionising manufacturing with artificial intelligence

Posted on 2 May 2019 by Maddy White

Artificial intelligence is the technology to transcend them all. It is set to transform design, manufacturing and after-market phases through the entire product life-cycle.

The time is right for manufacturing leaders to commit to AI beyond pilots and proof of concepts, and take full advantage of its promise to revolutionise their people, processes and machines.

The research projects taking at Factory 2050 produce real-world answers to today’s manufacturing problems – image courtesy of AMRC.
Factory 2050 offers real-world answers to today’s manufacturing problems – image courtesy of AMRC.

Eight out of ten (84%) senior manufacturing executives believe that digital technologies like artificial intelligence will drive innovation in design development and manufacturing processes, according to The Manufacturer’s Annual Manufacturing Report 2019.

It is clear that business leaders know the value of artificial intelligence, but are they moving fast enough to introduce it?

This formed the core of the latest in The Manufacturer Director’s Forum series of events. First, a group of manufacturing executives toured Factory 2050 — part of the Advanced Manufacturing Research Centre (AMRC) in Sheffield — before sitting down over dinner, co-hosted with IBM, to discuss the issue in-depth.

The executives, representing a broad cross-section of businesses, spent the evening discussing AI use cases, what they had seen at Factory 2050, and the possible applications of artificial intelligence in their businesses.

Sharpening the manufacturing edge with AI

The evening kicked off with Henry Anson, Managing Director of Hennik Group – publishers of The Manufacturer, discussing the importance of Factory 2050, a centre that produces the answers to many of the manufacturing sector’s problems.

 “It is a hub for cutting-edge innovation, and from the tour today it is clear that there are many opportunities artificial intelligence can offer the manufacturing industry.”

From retrofitted legacy equipment with sensors and advanced software to match 21st century manufacturing demands, to precise visual recognition tools that could aid medical professionals, Factory 2050 is a showcase which proves Industry 4.0 (and beyond) can be accessed by businesses of all sizes.

Factory 2050 is the place where businesses can come and interact with the technologies driving the 4th industrial revolution - image courtesy of AMRC

Sam Turner, CTO at the High Value Manufacturing Catapult, of which the AMRC is a part, continued the conversation by adding, “The integration of artificial intelligence into manufacturing will continue to accelerate. I think we will see a bigger role for AI in manufacturing design and in getting greater insights from the products already in operation.

“This is crucial in terms of product quality and in design stages. The first question to ask now is how accessible is AI to businesses like yours, and how accessible are the things you have seen today at Factory 2050?”

One delegate said, “There were many practical examples from today of technology from improving quality inspection to boosting workforce efficiency through AI-powered assistants. But we are running at 100mph to keep ourselves profitable and people are almost nervous to stop at times, I am interested to understand how we can make that AI engagement step change smoothly.”

Proving the value of artificial intelligence

A key issue discussed at the dinner was that making the business case for technology adoption remains a long-standing obstacle. And that even if the business case is accepted, and a pilot validates the value-add, projects regularly stay on the shelf.

“Often pilots are tested and advantages are proved, but concepts still aren’t taken through to scaled deployment,” Andrew Tyler, Industrial Director at IBM, said. “What are the barriers to that? It could be something to do with manufacturing firms’ ability to absorb change.”

Insights to data need to be available to the shopfloor - image courtes
Insights to data need to be available to the shopfloor – image courtesy of Depositphotos.

A senior executive at a global packaging company commented, “It is not just getting the cash, it is the resources too, to get people to spend the time to experiment and prove these concepts. I think it is my job to fight across the company to say, ‘look we need to do these things and actually implement them.’”

Sam Turner added, “The application of AI has started to progress to a practical level. What has become apparent is that in order to facilitate adoption of AI on the shopfloor, the insights from the data need to be made more accessible to the workers. That’s why the cognitive wrapper which allows conversation with the AI systems through chatbots and digital assistants are so powerful.”

The importance of all that data

The proliferation of data became a topic that resonated with the manufacturers around the table: the increasing volume, variety and aggregation of data and the challenge of knowing exactly how to exploit it.

“We seem to want to do something with the data but aren’t sure what. It is important to understand the objective and where the tangible benefit is. We can get data, but it’s what we want to do with it that counts,” one guest said.

This was confirmed by another delegate who added, “As an industry, I am not sure manufacturers understand the value of data, or the power of it. My concern, when we are talking about AI and data, is what’s practical on the shop floor? You can drown yourself in too much data, so it must deliver a real benefit.”

One use of data and artificial intelligence is in predictive maintenance. If manufacturers can track machine failures, and the conditions that cause them, by looking at the data, then they can become better at predictive maintenance, reduce downtime and increase profits.

On this note, a director at an industrial machinery manufacturer said, “How do we start to track those near misses, how do we start to understand those elements in a meaningful way, and then how do we deploy it to the shop floor so we minimise the friction of its adoption?”

Pilot purgatory to scaled deployment

The crucial question around the table remained: are the rich technologies in use at Factory 2050 and the benefits of adoption highlighted by the evening’s discussion within the grasp of all manufacturers?

smart factory - automation - image courtesy of Depositphotos
Are these rich technologies in use at Factory 2050 within the grasp of manufacturing – image courtesy of Depositphotos.

“Absolutely, there are applications and ideas that could be used in the firm I work at,” one delegate said. “The biggest barrier seems to be justifying it within our firms, to prove the value of these technologies and then actually implementing them.”

Sam Turner added, “I think there are relevant artificial intelligence applications for every manufacturer around the table; design, maintenance and quality inspection, and for small and large volume production.”

IBM’s Andrew Tyler rounded off the evening by saying, “The discussions today have confirmed that AI offers manufacturing huge opportunities, but we have to focus on a solution to how we move from pilot activity to scaled deployment that delivers the value from these technologies that we know are possible.

“Data is a very important facet for being able to take advantage of AI, especially considering its variety, velocity, volume and veracity coming off shopfloor machines.

“However, it is also about so much more than data. To truly embrace AI, manufacturers need to be prepared for a transformation encompassing their people, processes and machines. And if they’re willing to commit themselves to this journey with the right partners, they have a better chance of realising the huge potential from AI and cognitive technologies.”

Supercharge your AI data pipeline with IBM Spectrum Storage

The success of artificial intelligence depends on the ability to transform data into accurate models and insights with speed and efficiency. Address every stage of the AI data pipeline from ingest to insight with IBM.