Search

Upcoming Conferences

Wireless comms in manufacturing: improving safety and efficiency

Manufacturers are increasingly looking to a range of communication technologies to improve operational efficiency. Sam Fenwick learns how they also help protect both men and machines

AdobeStock_98183151.jpeg

Manufacturing is a tough business. The ability to ship containers full of mass-produced goods all over the world means that factory owners are competing in an international marketplace. This puts such pressure on margins that companies either adopt a laser-like focus on incremental efficiency improvements or go out of business. At the same time manufacturers have a duty of care to their employees, who work with heavy machinery and alongside industrial processes that use chemicals and extremes of temperature and pressure.

Wireless communications can help boost efficiency through speeding up response times and removing the delays associated with trying to get through to engineers on cell phones. However, according to Klaus Allion, managing director of ANT Telecom: “Some manufacturing companies are still working in very antiquated ways”, with employees not having access to communications devices. 

Integrated communication systems, such as those supplied by ANT, allow users to press an emergency button that sends a text message to multiple potential first aiders, allows them to acknowledge it and say if they will respond. If no-one can the message is escalated and the system supplier can provide location data to other staff or the emergency services, if required. 

Allion highlights the importance of this with a recent case at a paper mill “where a diabetic worker suddenly became unconscious and fell. He couldn’t press a button. In that case we automatically detected the fall and started [the] automatic process… By giving an approximate location he could be found relatively quickly, first aid could be administered, and an ambulance was immediately called. Apparently if that hadn’t have happened that quickly it would have been much worse for him and he probably wouldn’t be able to work anymore.”

He adds that one company was using different lone worker processes at its various locations. The Health and Safety Executive pointed out that as one of the solutions was likely to be safer than the others why not apply it to all the sites?

This alerting method isn’t limited to ensuring workers’ safety either. One of the key drivers for good communications in manufacturing is the need to cut the time it takes to recover from outages. While downtime costs may not be in the millions of pounds per hour they can still make unplanned outages incredibly painful for manufacturers. To this end, having a predefined protocol that automatically alerts operators and managers and eliminates the need to scrabble through address books makes a lot of sense. 

“When we take those cases we ask if it’s more efficient to automatically also send machine alerts to potential responders, who can then say that they are taking responsibility.

“[We] provide them with ideally a more detailed error description than [a verbal one], to equip them immediately with the right information,” says Allion. He adds that this can be applied to machine breakdowns and instances where predefined operating thresholds are breached, so engineers “can immediately attend to the machine and change the relevant parameters to make sure that either it’s completely resolved or not too much product is wasted”.

Such alerts “can be integrated with whatever communications technology the customer has”.

AdobeStock_57528140.jpg

Tips for manufacturers
What advice does Klaus Allion, managing director of ANT Telecom, have for manufacturers? “First look at [your] work and safety processes… to fully understand them and see where effi ciencies can be gained so you know what you want to do. Also assess what you might want to do in a year or two’s time, and then look at all the available technology, ideally across a number of plants set in maybe slightly diff erent environments and see what the best solution would be.”

He adds that not every plant is the same in terms of cellular coverage so one MNO might not provide adequate coverage to all the plants a manufacturer operates. “Not every plant is the same, so you need to look at your requirements reasonably diligently before you [decide] where to invest your money.” The same holds true for devices: “Look at where you use the equipment in a litt le bit more detail and make sure you are not going to provide your staff with equipment that you’re going to replace in half a year’s time; either because it’s broken or because no-one is using it.”

Machine learning and simulation
There’s growing use of machine learning to provide users with greater visibility of industrial processes. Andrea Carcano, co-founder of and chief product officer at Nozomi Networks, says that this has several benefits. From a cybersecurity perspective it allows a system to interpret commands and understand if someone’s attempting to disrupt the manufacturing process. A classic example of this type of attack is Stuxnet – a worm identified in 2010 – which is believed to have altered the behaviour of uranium enrichment centrifuges to slow down Iran’s nuclear development programme. 

In manufacturing there’s a danger that malicious changes to automated processes might not be detected until the affected products enter use, potentially creating huge reputational damage. 

Carcano adds that operational benefits from using machine learning to understand industrial processes include faster troubleshooting, which reduces the need for emergency maintenance when misconfiguration damages a component.

One of the big ideas in this realm is the ‘digital twin’ concept; in which processes and products are simulated using data feeds from sensors on the field equipment, giving engineers much greater insight into what is going on internally to better optimise performance and maintenance in real time. This approach is being pursued by companies such as GE, Ansys and Siemens.

While sensor readings alone may be enough to determine that there is a problem, the digital twin allows users to see the effects changing various parameters would have without altering the machine’s physical process. GE demonstrated a digital twin in combination with an augmented reality headset and voice recognition at its Minds + Machines conference to show wear and tear in a steam turbine, where the trouble spot is, and two options to address the issue based on the turbine’s history, with the idea that in the field it would be possible to overlay the digital simulation on its physical counterpart. 

The greater visibility and understanding provided by digital twins might also lead to new business models, such as expensive machinery being supplied as an asset under a service model (machines as a service). This is because it could give manufacturers the information on availability, costs and yields that is needed for drawing up contracts and make it easier to address issues remotely. 

The international nature of some large companies is also creating a desire to manage assets across continents.
In November 2016 Nestlé announced it is using Telefónica Business Solutions to provide IoT communications for its coffee machines in more than 50 countries, which allows the firm to control the machines’ parameters remotely to optimise the end user experience. 

Those looking to embrace the full potential of the Industrial Internet of Things (IIOT) shouldn’t forget that there are costs as well as savings. A Cisco Jasper whitepaper published in April 2016 puts the price of running 100,000 industrial monitoring and heavy equipment devices at $1.25 per device per month, and adds that payments to MNOs can account for 33 to 50 per cent of the operating expense of IIOT connectivity. 

5G for factories
What role will 5G play in industrial IoT and how engaged are manufacturing end users with the 5G standardisation process?

Volker Held, head of 5G market development at Nokia, sees two major benefits – reducing fragmentation and the amount of cabling. He highlights the large number of different processes and applications that take place in a typical factory and their varying requirements – some are very time-critical while others may be less so. 

“Historically they’re using different communications standards, I would call it a family of standards with a lot of sub-standards. If a system is labelled RFID you cannot be sure that two systems labelled RFID can talk to each other. So there’s a [lot of fragmentation]. With 5G there’s a chance to get rid of that because 5G can adapt to the applications’ different requirements in a manufacturing setting. You would dedicate a specific network slice or even a group of network slices to a factory, for example. That means you can give the exact network resources and capabilities that an application needs.

“If you have to reconfigure your production line and you have a lot of cables to manage it’s a huge problem,” Held adds. “With the low latency enabled by 5G you can connect robots without cables. And cables can break so it would be a major benefit to get rid of them. There are a lot of projects underway that are looking specifically at these
use cases and are trialling 5G technologies for these environments.”

“In terms of factory automation we have made some progress; we’re in discussion with some factory automation activities,” says Adrian Scrase, ETSI’s CTO.

“It’s early days, so it would be wrong to give the impression that we’re fully engaged with [the sector]. That probably shouldn’t be a surprise as to some extent the timelines for different verticals are slightly different. In the mission-critical sector we had an urgent need to do something because we had certain countries making very early moves and demanding early results. 

“In the case of factory automation the timeline is not so urgent, there’s not a compelling need that says this needs to be done yesterday, so people are planning for the longer term and they want to get things absolutely right, so we have this different expectation for a timeline.”

Returning to the use of machine learning, it’s not difficult to imagine a scenario in which a manufacturer’s plants across the world autonomously adjust their output based on market trends and sales data. While this could be incredibly powerful in isolation, once everyone is using such a system might it expose manufacturers to some of the volatility and instability seen in the financial markets after the introduction of high-frequency trading and automated trading systems?… 


Leave a comment


  • Captcha Code
    Click the image to see another captcha.
  • Submit