Smart factories - why manufacturing big change is difficult

Profile picture for user cmiddleton By Chris Middleton November 21, 2019
Summary:
A new report from Capgemini promises major gains from smart manufacturing – but also major obstacles.

real-time-manufacturing

Among the many industries that stand to be transformed by robotics, automation, and Artificial Intelligence (AI), one stands out. While energy, oil & gas, nuclear decommissioning, engineering, mining, agriculture, transport, and the supply chain are among the sectors being impacted by Industry 4.0 technologies, manufacturing is the one most often in the media spotlight, partly because of the decades-long adoption of robots in the automotive, electronics, and metals sections of that industry.

For example, US car makers reportedly purchased 60,000 industrial robots in 2016 alone, and yet human employment has increased, including for skilled engineers. As explored in previous diginomica reports, the widely held belief that ‘robots in’ equals ‘humans’ out’ is not borne out by many statistics to date, with human unemployment low or falling in nearly all of the top 10 most automated, robot-dense nations on the planet, such as South Korea, the US, and Japan.

But automation and industrial robots are just one element of smart manufacturing. Sensors, Big Data analytics, Enterprise Asset Management, and the broad range of connected devices that make up the Industrial Internet of Things (IIoT) are the critical components in a market that could add between $1.5 trillion and $2.2 trillion to the world economy over the next five years, according to Capgemini.

‘Could’ being the operative word. Just 14% of the consultancy giant’s research sample of 1,000 companies worldwide describe their forays into smart manufacturing as successful, with nearly two-thirds of organisations saying that they are struggling to scale their smart factory programmes. The findings have been published in a new report, entitled ‘Smart Factories at Scale’.

In other words, that potential $2.2 trillion global uplift – which could be achieved via increased productivity, improved product quality, and bigger market shares from optimising operations – is merely an aspiration for most companies. Complexity, legacy technology, security fears, lack of skills, and intransigence all conspire to make transforming a traditional sector a more frustrating experience than the marketers would have us believe.

One-third of the factories surveyed have been transformed into smart facilities, and those companies plan to establish 40% more over the next five years, increasing annual investments in connected technologies by 1.7 times compared with the previous three years. 5G is set to become an important enabler of these initiatives.

Robotics remains a key component of industrial architectures. However, the main areas of investment for at-scale smart factory deployments are the IoT and manufacturing intelligence, in order to support data-driven operations as well as remote and mobile capabilities.

There are three elements in the promised performance leap that comes from smart factories, says Capgemini:

  • efficiency by design, via virtual design and simulation prior to installation.
  • effectiveness in operation, using advanced analytics to create a closed loop en route to self-optimised work.
  • deployment at scale.

Once manufacturers have reached this advanced stage – a virtuous circle of efficiency – the biggest challenge is scale. Put another way, some of these initiatives work well in testbeds, but rolling them out across an organisation is a much bigger challenge. The main obstacles to achieving scale are related to IT and operational technology convergence, says the report, and the need for employees to develop hybrid capabilities and soft skills:

To ensure digital continuity and enable collaboration, effective IT-OT convergence will be critical, including digital platform deployment and integration, data readiness, and cybersecurity. Agnostic and secure multilayer architectures will allow a progressive convergence.

Developing hybrid job profiles, such as engineering/manufacturing, manufacturing/maintenance, and safety/security will be essential, as will skills such as problem-solving and collaboration. However, organisations are not investing fast enough to fill their skills gaps, warns the report.

Meanwhile, less than 50% of organisations have adequate data availability and cybersecurity measures in place, while nearly one-quarter of manufacturers have experienced a cyber-attack on their connected systems in the last year.

Health analogies 

In this sense, manufacturing has much in common with healthcare, another industry where specialist legacy equipment was never designed to be connected to the internet and can't simply be stripped out and replaced overnight. As a result, it becomes a security weak spot and target.

Nigel Thomas is Head of the MAALs (manufacturing, aerospace & defence, automotive, and life sciences) unit at Capgemini, which provides technology and transformational services to clients such as Rolls Royce, BAe Systems, and Airbus. He says:

Healthcare is a good comparison. It’s a huge challenge to work out whether or not you should be connecting these devices to each other or to a data platform, and invariably that means the cloud. And if you do that, then your strategic asset – your data – becomes at risk.

Nevertheless, he remains bullish about the sector, despite the mixed bag of success, failure, and challenges that the report represents:

In the two years since we last looked at smart factories, the projected spend is up by about 50%, as respondents see significant growth in their investment in this area.

For me, the interest is in the development of the investment plans and the maturity – particularly among those that we identified as front runners. The 10% of respondents who were mature two years ago are now running away with the concept of smart factories and their ability to deliver the projected value.

These front runners have successfully digitised their entire industrial systems, says the report.:

This elite group of companies make significant investments in the foundations (digital platforms, IT-OT convergence, talent, governance), and balance efficiency by design and effectiveness in operations, leveraging the power of data and collaboration.

Organisations should identify where they are lagging behind and learn from the best practices of the front runners. Smart factories are a critical part of ‘the intelligent industry’. Therefore, realising the complete potential of smart factories will be the key to reaping the benefits of intelligent industry.

To unlock the promise of the smart factory, organisations need to design and implement a strong governance programme and develop a culture of data-driven operations.”

So why are some companies not exploring smart manufacturing in greater depth, beyond the triple-S challenges of scale, skills, and security? Thomas says:

What’s constraining adoption is a good question. The legacy technology base – the machines and the manufacturing infrastructure – is one factor. There are many manufacturers who look at their existing operations and see they’ve been using the same machines in some cases for decades. So they think, ‘Why do I need to connect my machinery to anything?’ Workforces have been using the same machines for however long and they know intuitively how to get the best out of them, and maximise uptime and throughput.

That’s a significant challenge, dealing with the legacy; the brownfield approach, if you like. It’s easier in concept to take a greenfield view and say, OK, I’m going to start on something new from scratch, like a new model car, or an engine on an aircraft, or a production line in life sciences. The brownfield nature of much manufacturing is more of a challenge.

And then you have real challenges in terms of how you collate data from devices and link it all the way through to the analytics needed to produce the insights. How do you select the digital platform, and how do you deploy and support it? That’s a fairly big challenge for most organisations.

And then, of course, you get to the biggest strategic asset, the data itself. And there you have issues about data quality, trust in that data, and availability, and surfacing the data in a reliable manner to the right place at the right time so you can do something insightful with it. And underpinning all of that are skills. That’s a big challenge too: the growth of our own digital skill sets. But all of these are inhibitors of adoption rather than barriers.

Presumably this is where Capgemini comes in for its clients. But it’s not always the case that greenfield smart factory programmes are a comparative walk in the park for manufacturers. In the same week that the report was published, sports footwear and clothing manufacturer Adidas announced that it is to shut down its cutting-edge Speedfactories programme.

The two prototype, largely automated facilities have been located in Ansbach in Germany and Atlanta in the US over the past two years. They have been part of an experimental move by Adidas to shift manufacturing away from a monolithic, global, outsourced model – such as making a million units in China and shipping them across the globe – and towards a more personalised, automated, and localised (PAL) model of micro-manufacturing shoes to order much closer to the customer. The closure will see Adidas redeploy its technologies to factories in Asia.

Whether the announcement was down to the technology not keeping pace with the company’s ambition, the old economics of global supply chains being too attractive to abandon, or a lack of customer demand isn’t clear. But it’s a disappointing outcome for proponents of a smarter, more efficient, more local, and more sustainable supply chain. Thomas observes:

Nike have been doing something similar in terms of allowing you to get your foot scanned and build your own customised shoe, and so on, at slightly incremental costs above mass production. There is certainly a manufacturing theme here – a nirvana in manufacturing of a batch size of one, or a lot size of one, and the ability at almost zero cost in terms of change to be able to produce hyper-customised products.

That’s still where manufacturing wants to go. It wants to have that flexibility and get rid of long batch runs, which inherently mean you have to have more inventory on the shelves, more costs per unit produced, and you have to run production for so long to get the economies of scale. So it’s a shame. It may just be a case of timing. A lot of this is about when we’re ready to consume these sorts of things.

Indeed. Another challenge facing smart manufacturing is the ‘PR problem’ faced by robotics, AI, and smart technologies: a hostile media that – wrongly, in most cases – equates innovation with humans being sidelined by evil, job-stealing machines.

Encouragingly, Thomas rejects the idea that manufacturers are at all fazed by this. He says:

Everybody in the industry is adopting a robot policy, whether its software robots or physical robots and ‘cobots’ [collaborative robots] to provide that ‘third hand’ when you are doing something complex on the production line, because everyone is driving towards more productivity.

Productivity is the single biggest problem – for UK manufacturing in particular. I don’t think inside businesses there is any scaremongering about robots; that isn’t the case. Jobs may change, which is why the report talks about the need for developing both hybrid and soft skills.

As you start to link technologies on the shop floor, you need those hybrid skills. They don’t grow on trees, they need time, so there is an element of education and training involved.

What we do with a lot of our clients is help them to grow their own skills locally and co-invest with our customers to create local courses or enhance apprenticeship schemes. In this way, the workforce coming in is already equipped with more data-oriented, change management, and problem-solving skills, which perhaps they didn’t need before when they were just working on a production line.

My take

This is an interesting takeaway from the discussion. Despite the labels ‘smart factories’ or ‘smart manufacturing’, it seems that the key driver for this sector is not greater insight from data, but increased productivity – in other words, doing more with less.

Thomas says:

Well, you’ve got to find the right business driver, the right business case. And I think that it is about data and insights, but it’s also about why are you doing it. So with just 14% declaring success from their smart factory initiatives, have you got your business case sorted out?”