Data at the ‘edge’ – harvested, stored or processed at an end point on the perimeter of the computer network – can drive real-time control over logistics and warehousing and optimisation of asset and inventory management, as well as minimising downtime.
However, perhaps 80% of manufacturers have not taken their first steps into the world of data enablement, according to Dan Mckiernan, President of Epicor‘s data capture acquisition eFlex.
Are you curious about how to deploy data within your organisation? Book your place at the Manufacturing Digitalisation Summit today.
“I would say the industry’s probably 20% into this evolution,” Mckiernan said. “With the shopfloor largely neglected (so far).”
Yet, data-driven intelligence can support event recognition and decision making, or help power flexible, responsive automation that hands fuller control to the operation, as required. That’s before talking about more ‘future forward’ approaches such as innovation with artificial intelligence (AI) or more advanced analytics.
From pandemic pressures to the future
Robert Crowther, Proprietor of Cheshire-based Devonshire Bakery, says the fifth generation firm is one of many that hadn’t had “a really good reason” to look at data until recently.
Pushed to try something new, partly due to the last three years’ pressures, the company has since learned that simply deploying sensors to monitor equipment, including the bakery’s key energy-hungry appliances, pinpointed where savings can be made. “Our biggest users of power are the ovens and the freezers,” he pointed out.
Devonshire Bakery also quickly discovered potential for fine-tuning its use of space heating and cooling, and learned that some refrigeration units could maintain a safe chilling temperature for several hours if switched off for a scheduled timeframe each day.
From the first month, savings were evident – helping guarantee company survival even as energy costs ballooned from £5,000/month to £12,500/month for similar consumption levels. Crowther added that more savings are expected over time.
“We’ve just been studying what alterations make the difference, and then we’ll look at whether the business could invest in much bigger changes,” Crowther said. “Data can give us confidence to make changes, and every little change can make a difference.”
eFlex’s Mckiernan said that, even with ERP already bedded in, many have a long way to go to realise these kinds of data-driven transformations, especially within smaller manufacturers.
An easy place to start that delivers results
A good starting point can be a paperless drive. Reducing paper use through digitisation can be far reaching and easy to make the case for early on. A common theme is that the order created on the ERP system is transferred to the floor on paper, in both small and large manufacturers.
“It’s distributed by some shop leader that has tribal knowledge and then it moves to the process. This is a very labour intensive, problematic way to attack manufacturing,” he said.
“So that is typically the point of entry and it builds on that.” From there, manufacturers can develop multi-media knowledge resources based on data, both from office and IoT enabled shopfloor systems, that helps production directly become more agile as well as enhancing decision making at all levels.
Anthony Walker, Strategic Manager, Faculty of Engineering and Technology at Liverpool John Moores University (LJMU), agreed. “The biggest challenge we’ve seen is people ask how they’re going to get ROI.”
That’s why their initiatives have increasingly looked at using demonstrations to help prove the business case and utility of a data project, themselves investing in “everything from simple IoT to visualisation” to help risk averse and smaller organisations see how to get started.
Demos of £50 sensors and a linked up, open-coded software running on a £500 unit for data acquisition can offer real-time info, representing a move to Industry 4.0 for low cost, showing how data can help decision making around process changes and the like.
“That’s been really successful,” Walker added. “Real-time connected data can revolutionise industry, it can reduce human error and allow for more informed decision making. With greater visibility into the metrics, a company can go forward.”
Eamonn O’Neill, Co-Founder, Director and CTO at cloud computing services consultancy Lemongrass, agreed that while some are further ahead on data enablement than others, data is still underused for the most part.
“It’s understandable. What’s particularly difficult about data projects within manufacturing companies is that you don’t know what the insights are going to be when you start,” O’Neill told The Manufacturer.
“In other words, you don’t know what value you’re going to get from your data until you do the analytics and machine learning work. This clearly dissuades a lot of companies that don’t want to take the risk.”
Fortunately, some companies jump first and when they get the benefits, others start to see the value and follow suit. He added that Lemongrass works with a global beverage company getting “huge value” through mixing internal and external data in a unique way that is delivering insight on production waste that they wouldn’t be able to get from anywhere else.
“Their strategy from the start was to figure out what the value was going to be once the data had been merged together,” he said. “This firm combined [production waste data] with weather data, which enabled them to create models showing that if it rained too heavily, production efforts would be wasted as rainfall affected the colour of the beverage making it unusable. That could save them millions.”
What about upcoming track and trace requirements?
Pharmaceuticals companies and similar might better meet their track and trace requirements with technology that can lower their data handling costs, for example, enabling them to go above and beyond current data analysis to track every movement of their product through the supply chain.
“They can even visualise these insights over time and produce animations of the flow of their product through the chain. All of a sudden, they have much greater insight into what’s happening minute-to-minute across global operations,” O’Neill explained.
‘Boring’ and non-consumer facing or social data can hold more value than expected, yet can be neglected. Low-level data (perhaps stored in the cloud) can be collected and structured in a readable way, then acted on with analytics or even machine learning – running models with data to tease out competitive advantage.
“It of course varies for different companies and there’s no guarantee, but this is a pattern we’re seeing repeatedly,” O’Neill said. “Investment is required, which makes taking that leap a little riskier. It’s only when they see competitors do it and get that advantage that it becomes essential and urgent.”
Speakers at manufacturing’s InterAct forum this year highlighted, meanwhile, that transformation based on smarter innovation is needed if the UK industry and its people are to thrive.
Data becomes central to successful business
Michael Ford, Senior Director of Emerging Industry Strategy at Industry 4.0 focused Aegis Software, added that the key is how you drive efficiencies, productivity and profits with all the data already collected with manufacturing execution systems (MES) and more. “That can be the nub of many things,” he stated. “But it’s not easy at all.”
Nowadays machine connectivity is available but the data communicated needs to make sense. One problem is that more standardisation is needed. Ford estimated that total potential costs could be in the billions of pounds for developing “basic smart factory” connectivity across the industry.
“It needs to happen step-by-step. With three machines, you can have three software systems that want to make use of the same data, which makes for multiple different connections and software customisations,” Ford added. “This is the scale of the barrier to data acquisition.”
Companies such as Aegis are developing tools and plug-and-play universal Internet of Things (IoT) standards aimed at surmounting these barriers, communicating with ERP, management systems, monitoring and more. However, full dissemination of fully interoperable tech through the manufacturing customer base, so they can enjoy the benefits, will take time.
Data points also need to be valid in the context in which they are used. “So for example, if the machine reports product completed, that’s interesting, but what does that mean?” Ford added. Does it also mean new materials need to be provided just in time? Does it mean production counts go up? Is the object or part in question good – or bad and should be scrapped? What does it mean with regards to reports or cycle times? Right now, some of those questions often go unanswered.
“That’s something that people forget. Data is going into these cloud-based data lakes and are ruining the environment,” he pointed out. “We found that 80% of the data in the cloud (hosted on the internet) is absolutely useless, because it hasn’t got context.”
Fixing all of that can take a long time and require considerable expertise – even before you start thinking about AI or machine learning and analytics enablement to make sense of the data available, as well as traceability. You need a digital record of something that’s happened, as opposed to a disparate data point.
“And for small and medium enterprise companies, they don’t have millions of dollars to throw at the problem,” Ford pointed out. “This is the big barrier. If you buy a machine and it’s making products, the ROI is very tangible. But with software, if you don’t use it, there is zero return. If you use it badly, it can even make things worse,” he warned.
Tom Richter, Global Head of Discrete & Process Manufacturing Verticals for Digital Industries, Nokia Enterprise, noted that the need for comprehensive data schemes vary per manufacturer. “It really depends on what they do and how they do it,” he said. Additionally, legacy equipment presents a barrier, while veteran workers with much experience sometimes don’t feel a need to know more, because they have such a good understanding of the machine and its requirements.
Yet that brings up another, increasingly worrisome, challenge – how do you transfer knowledge and retain that value as workers join and leave an operation over time?
Data for knowledge transfer and teamwork
And that’s where bidirectional use of data, communicated to and from the shopfloor, even with remote teams working on things like table inspections of aeroplane parts, via universal connectivity, really demonstrates value, Richter pointed out.
“Yet there’s a long way (for manufacturers) to go to make this all tangible,” he agreed, noting that preconfigured solutions, automation, dashboards and plug-and-play functionality can speed up change, becoming more essential with upcoming legislation – including around supply chain due diligence.
“This requires you to get the data otherwise you will be out of business,” Richter warned. “But you need to ensure quality and efficiency while you produce continuously and support new techniques and new generation workers who expect digital enablement.”
Epicor UKI Vice President Mark Hughes said its acquisition of shopfloor data capture focused eFlex, reflects these multiple drivers. “The challenge with the shopfloor is what you do with that data,” he agreed. “Another is the people.” Skills are an issue but another that might not always spring to mind is that sometimes workers perceive data collection and monitoring as corporate surveillance or threatening their roles.
Yet as well as collecting data, IoT-enabled systems can deliver work instructions that can boost flexibility, making it easier for staff to transfer to and from different roles or parts of a factory, helping drive improved job satisfaction and innovation. “When customers have gone for flexible workstations with adjustable heights, they get a kit of parts delivered at the right height.
They’ve got a screen telling them what they need to do, how to assemble it, what to check etc,” said Hughes. “You can make anything that arrives on this workstation based upon the instructions delivered, validate and check it and so on. Data can really drive agile, flexible manufacturing by delivering information to the operator.”
For more stories on Industrial Data & AI click here.
