Technology transfer: how digital trends in the consumer market might benefit process plant operations and asset management.
WHILE the digital economy is enjoying strong growth in the consumer market, where virtual reality (VR) gaming, music downloads, online shopping, cloud storage, internet-connected white goods and on-demand TV are all but taken for granted, the technologies that support these activities are steadily emerging as useful tools for industry. But as compelling as this technology transfer may appear to be, can we seriously envisage consumer technology having a useful role to play in a complex, safety-critical environment such as a chemical, biochemical or process plant?
Virtual reality in itself is of little use for industrial users but subsequent developments such as ‘augmented’ reality and ‘mixed’ reality are, and they are beginning to have quite an impact in areas such as plant engineering and plant maintenance. The term ‘big data’ might conjure up visions of slick 21st century marketing strategies but for plant operations it could herald an era of heightened asset reliability and availability. And then there’s the refrigerator that ‘knows’ when milk is running low and orders more over the internet. Now wouldn’t that be an efficient way to automate the ordering, manufacture and delivery of plant equipment and spare parts?
Of course, while this is not suggesting the wholesale adoption of cheap and readily available commercial or consumer technology for industrial use, the engineering underpinning these concepts has great potential outside of these environments. Like the evolution of ethernet from the office to the factory floor, these technologies can be made sufficiently robust for industrial application, and their benefits could have very positive implications for plant operations and plant economics.
Setting aside its more familiar applications in gaming and consumer attractions, virtual reality has long been an essential tool for the design engineer as it provides valuable insights into the 3D virtual design model, iteration by iteration, ahead of any investment in the physical prototype. Augmented – and its more complex complementary technology, mixed reality – on the other hand, are digital tools that provide views of real-world objects overlain with pertinent information, and therefore have greater relevance to plant operations engineers and plant maintenance technicians.
Augmented reality adds 2D- or 3D-layered content on top of real world objects or locations, allowing the user to access relevant information about that object or location. Mixed reality goes further, combining sensors, advanced optics, audio and powerful computing that enable relevant digital content to be instantaneously and automatically assigned to the location or object being viewed, as well as providing an interactive environment for the user and a communications channel between that user and a remotely located colleague.
Recognising the value of these tools to onsite commissioning and maintenance operatives, programmers at SKF’s software development centre in Gothenburg, Sweden, have been working with application engineers and product specialists to develop mixed reality technologies that make our engineering competence more widely available, and particularly to those working in remote locations who would otherwise have limited access to these resources.
Plant operators occasionally need access to the specialist knowledge of experienced application engineers. Providing the manpower for this is a growing challenge, not least with travelling between locations taking up much of an engineer’s time. Augmented reality solutions avoid wasted travel time and allow remotely located engineers to focus on any onsite problems via the internet instead.
During an augmented reality session, the environment a user sees before them is combined with a digital model that overlays technical information, instructions and real-time asset performance data. When they move their head to the left, to look at a different part of the plant or a plant asset such as a pump or valve, so the digital overlay adjusts to reveal pertinent information about that asset. What the user sees may also be shared with a remotely located engineer, enabling more informed decisions to be taken concerning corrective actions, should they be necessary.
The real value of augmented or mixed reality for the maintenance technician lies in combining the functionality of hardware such as Microsoft’s Hololens ‘smartglasses’ with machine health monitoring platforms. The recently-launched SKF Enlight Centre, an internet-enabled condition-based maintenance (CBM) and condition monitoring (CM) system, is an example of this type of platform.
Wearing a pair of Hololens smartglasses, the commissioning engineer or maintenance technician simply connects their smartphone, tablet or laptop to the internet to set up a link with a remote service expert who is able to view what the engineer or technician is seeing in real time. The expert can interact visually with this, overlaying images or texts on the user’s screen and providing additional spoken instructions that guide the user on what actions need to be taken.
Augmented or mixed reality can certainly be of use to an on-the-spot maintenance technician but in the longer term, predictive maintenance is what really matters. Predictive maintenance avoids unscheduled, potentially costly downtime but it is also data intensive. Enormous volumes of current and archived CBM or CM data need to be analysed to provide informed decisions about when to replace a bearing, for example, or change a seal on a pump. This is the area of ‘big data’ and it has as much, if not more relevance to the plant maintenance team as it does to any consumer marketing department.
The success of a process plant operation will depend on any number of factors, but the reliability and efficiency of its production assets – the pumps, valves, fans and other motor-driven machinery – will feature prominently among them. Monitoring the condition of these assets is vital to ensuring their continuing reliability and availability, but this requires constant measurement and has the potential to generate vast amounts of data. A CM system comprising multiple, networked sensors delivering instantaneous measurements of temperature, vibration level, speed or acceleration provides valuable insights into the condition of a machine at any point in time. When these measured parameters are analysed over a period of time, they may also identify trends that indicate early signs of wear or lubrication issues that need to be addressed.
The digital manufacturing model proposed by Industry 4.0 places a considerable burden of responsibility on a company and its ability to collect and analyse the vast quantity of data it generates – not just from its commercial operations but from its engineering and maintenance related activities. The question is: how to manage these data effectively and achieve the benefits of big data analysis to operations, such as improved machine reliability and availability?
An issue that manufacturers face these days is a skills shortage. Traditional maintenance departments have, in many cases, been disbanded and this function is outsourced – often to equipment and component supplier companies that possess the necessary expertise and servicing infrastructure. For example, we have been monitoring customer equipment remotely for 15 years. We currently has around 1m of our bearings connected to the Cloud, with data relating to such parameters as operating temperatures and changes in vibration levels or vibration signatures being gathered and interpreted on a continual basis. Careful manipulation of these data not only provides an early prediction of impending bearing failure, but may also provide a better understanding of a bearing’s dynamics, suggesting an alternative approach to its design or in-service lubrication requirements.
Software platforms that are capable of gathering a CM system’s big data efficiently, while aiding its interpretation, are becoming increasingly important in this area. Along with the Enlight platform , we have the Insight, which is designed to monitor bearing performance in real time. Enabled by self-powered, wireless sensors integrated into the bearing or its housing, Insight measures load, speed, vibration and temperature and sends these measurements to the Cloud. Users can connect directly to remote diagnostic services for condition monitoring support and expert analysis, or the data may be used to generate reports providing early warnings of imminent bearing failure.
The application of modern CM and CBM technologies has risen sharply in recent years and, in turn, this has seen an exponential rise in the volumes of related data requiring meaningful interpretation. Software platforms are achieving this now and, what’s more, are placing big data manipulation directly into the hands of maintenance operatives.
This might be one of the finance director’s ultimate goals, but is it at all feasible in the case of a complex process plant operation? To this point, we have seen how augmented reality and big data can be exploited to provide extraordinary insights into the current condition and serviceability of a plant’s assets. Would it be possible to go a step further and use this information to guide the ordering – even the manufacturing – of replacement equipment and spare parts, automatically and well in advance of part failure? Let’s narrow this down to bearings, which are found in all manner of plant assets.
Automatic detection of a failing bearing may represent a big step forward in terms of efficiency but the process of ordering a replacement – including sending the purchase order through to manufacturing, estimating the lead time, and delivering the part – still involves major human intervention.
Our manufacturing operation in Gothenburg is already gearing up for a future in which a faulty part effectively places an order for its own replacement. Our platform can provide a bearing condition diagnosis and report it via the Cloud; it’s not hard to imagine that it might also send an automated message all the way back through the supply chain when a replacement is deemed necessary. Meanwhile, back at the supplier’s factory, digital manufacturing has significantly reduced machine re-setting times, so a specific replacement part can be scheduled for addition to the production line with minimal disruption.
Combining these two factors – accurate prediction of a failing part, with ‘manufacturing to order’ – ensures that some ‘projected demand’ for parts is replaced by ‘actual demand’. This extends the ‘just in time’ manufacturing concept down as far as the individual component and could one day bring stock levels close to zero. It’s hard to imagine a world without stock, but this vision is within sight.
The digital manufacturing model as exemplified by Industry 4.0 is upon us and we would be remiss in ignoring its implications for the future of manufacturing. The technologies we now take for granted in our domestic lives have roles to play and we are seeing the steady transfer of these technologies into the industrial environment. This poses many engineering challenges but the rewards will be too good to miss out on.