How data will build the factories of the future
Mon 14 May 2018 | Leo Craig
With the continuing rise of robotics and automation on the factory floor, it’s not just UK manufacturing that needs to get to grips with these technological advances. Leo Craig, General Manager of Riello UPS, explains impact Industry 4.0 will have on the data centre and power protection sectors too.
‘Made Smarter’, the Government’s industry-led review of industrial digitalisation, claims Industry 4.0 could boost UK manufacturing by £455 billion and create 175,000 jobs whilst cutting CO2 emissions by 4.5%. On a global scale, Accenture research reveals the ‘Industrial Internet of Things’ (IIoT) could add more than $14 trillion by 2030, the equivalent of the current GDP of the UK, France, Germany, Italy, Spain, Canada, the Netherlands, and Belgium combined.
When you comprehend the sheer scale of those figures, it’s no surprise why the majority of UK manufacturers are keen to embrace the possibilities provided by increased automation, artificial intelligence, machine learning, and robotics.
The performance logs from a single works machine can generate around 5 gigabytes (GB) of data per week
The factory floor is home to hundreds of machines, from industrial plant and production lines, through to air conditioning units and the vital uninterruptible power supply (UPS) systems keeping the electricity flowing. Each of these devices is fitted with countless sensors that produce priceless data and enable them to interact with each other.
Combine this constant flow of data with intelligent, real-time analysis and insight, and the outcome is obvious. Reduced processing flaws, improved production quality, enhanced efficiency, optimised supply chains, better maintenance – working smarter offers some serious time and money savings.
As with any industrial revolution, there are questions and concerns. Nearly a third of UK jobs (30%) are said to be at risk from the ‘rise of the robots’, although the reality is likely to see many traditional roles displaced rather than mass unemployment.
There are big questions to answer for us in the datacentre and critical power protection industries too, namely how we contend with the enormous volumes of data that smart factory, IoT-connected devices will create. The performance logs from a single works machine can generate around 5 gigabytes (GB) of data per week, and a typical smart factory produces around 5 petabytes (PB) per week – that’s 5 million GB, the equivalent of more than 300,000 16 GB iPhones!
Big Data, big opportunities
For manufacturers to fully tap into the potential of Industry 4.0, they must combine data from their connected devices with AI, processing, and analytics. Whereas a traditional factory would produce goods, smart factory produces goods and data – the two must go hand-in-hand to realise the advances in productivity and efficiency.
These big data benefits are far-reaching across the factory floor and used in all areas of operations, from product quality and stock control, through to supply chain optimisation and improved health and safety. The data produced by a smart, connected UPS, can feed into AI-influenced decisions that impact on a plant’s power consumption, energy efficiency, and machinery maintenance regime.
As much as 30% of a manufacturer’s annual revenues can be lost through defects in the production process. By analysing real-time data from sensors on the production line, many quality issues are spotted and rectified as soon as they arise.
Tech giant Intel famously used data analytics to help predict equipment failure in one of their microchips. The outcomes were exceptional. A 50% reduction in maintenance time, 25% higher yields, and a 20% reduction in the cost of spare parts all added up to a saving of $3 million. The flip-side is that to produce this positive outcome required 5 terabytes (TB) of machine data per hour!
Another crucial area where data from IIoT devices is having a major impact is machinery maintenance. Industrial equipment tends to be serviced on a fixed schedule (i.e. monthly or yearly) regardless of its operating condition. Analysing performance statistics produced by the machines enables preventive maintenance to be carried out based on needs rather than time, leading to less wasted labour and reducing the risk of unexpected failure.
Whether it’s measuring damage or deterioration during transit to optimise packaging materials, studying consumer trends and buyer behaviour in real-time to inform production output, or monitoring KPIs such as staff absence or workplace injuries to improve in-house processes, the impact of connected devices and data analysis impacts every aspect of industrial life.
Manufacturers even glean valuable insight from sensors built-in to many products long after they’ve been sold to the end-user. Data demonstrates how an item is being used – is its performance impacted by the surrounding environment? What features are customers utilising the most? Precious insight that feeds back into the ongoing product development process.
The rise of the robots – or should that be the Co-Bots?
Of course, whenever there’s discussion about Industry 4.0 and smart factories, the talk soon turns to robots. Robotic process automation (RPA) has always played a huge part in driving forward manufacturing efficiencies.
In factories of the future, it won’t be an either-or choice between humans or robots – Co-bots and workers will be side-by-side on the plant floor
Robots have automated so many of the dull, manual, and even unsafe tasks that we humans used to perform. They perform these tasks better, quicker, and more accurately than we ever could. And unlike us, robots don’t get tired towards the end of a shift, so productivity levels remain consistently high.
Here in the UK, we lag behind many of our competitors. International Federation of Robotics statistics shows that there are only 33 robots per 10,000 employees here, well behind Japan (213 per 10,000), Germany (170 per 10,000) and even Sweden (154 per 10,000). And with the Boston Consulting Group predicting a quarter of manufacturing tasks will be carried out by a robot by 2025 (up from 10% at present) that prevailing mindset, and those ratios, need to change.
Advances in computing power and networking technologies mean today’s robots are far ‘smarter’ than their predecessors, not just in the functions they can perform but how they can adapt on the go. Whereas early robots simply carried out the same function again and again, modern robots can adjust their movements in real-time, learn lessons, and even collaborate with each other – all thanks to the quantities of data they and their fellow machines produce.
In factories of the future, it won’t be an either-or choice between humans or robots. Co-bots and workers will be side-by-side on the plant floor. Indeed, by 2020 The Manufacturer predicts 60% of human factory workers will be working alongside automated assistance technologies such as robotics or AI.
What are the Industry 4.0 implications for data centres?
At one time, datacentres were purely means for storage of data. But we all know the industrial internet of things makes everything far more dynamic and fluid. In smart factories across the country, data is being captured, analysed, and processed – all in real-time.
All this additional processing requires extra storage capacity and extra power. It does, of course, provide the data centre industry with fantastic opportunities, but at the same time raises legitimate questions for managers and operators to consider if they are to fully capitalise.
Is your datacentre capable of handling the 5 GB of data that a single smart machine will create each week? Probably. But what about 100 machines? Or 1,000 machines? Or even more, when you think how many devices and sensors can be housed on the factory floor? And how will you balance the need for additional power to keep up with processing demands without requiring a huge expansion in footprint? The classic conundrum of ‘doing more with less’.
Smaller micro datacentres utilising a modular approach are making it possible for data processing facilities to be based either on-site or as close to the location as possible. This edge computing is essential as it gives manufacturers the capabilities to run real-time analytics, rather than vast volumes of data needing to be shipped all the way to the cloud and back for processing.
Modular datacentres give operators the scope to ‘pay as you grow’ as and when the time comes for expansion. The rise of modular UPS provides similar benefits in terms of power protection requirements too. And with all the additional revenues a datacentre could make from Industry 4.0, the need for a reliable and robust continuous supply of electricity becomes even more imperative.
Transformerless modular UPSs deliver higher power density in less space, run far more efficiently at all power loads so waste less energy, and also don’t need as much energy-intensive air conditioning to keep them cool.
Any data centre manager planning to take advantage of manufacturers’ growing data demands would be wise to review their current power protection capabilities. If their UPS units are older, bigger, and less efficient models, upgrading to modern, modular versions would be a prudent move to ensure you’re fully prepared for the demands smart factories will bring.
Of course, for certain manufacturers, particularly those with multiple sites spread across several locations, there’ll still be a requirement for some sort of centralised data storage capability, whether that’s an onsite datacentre or via the cloud.
Increased automation will also leave us questioning many of the fundamental factors usually required for running a factory or datacentre. If processes are automated to the stage where facilities are practically unmanned, do they need lighting or heating? This opens up opportunities for significant energy and cost savings.
Today’s consumers are more demanding than ever. They prefer personalised or custom products to standard ones, and they want them almost instantaneously. Of course, they also expect these unique products to be accessible and as cheap as mass-produced ones.
The rise of machine learning and Industry 4.0 has the potential to kick-start an ‘Uber-isation’ of manufacturing, where custom products are made on-demand locally, rather than being shipped throughout the world.
It’s clear datacentres and manufacturers will need to work together to harness the full potential of the IoT and the fourth industrial revolution. Check back next week when we’ll take a look at how data, robotics, and AI is the 21st century’s prescription that will help keep the NHS and the country’s healthcare service fighting fit.