Internet of things: connecting the unconnected

eandt.theiet.org

If ever the importance of industrial automation was in doubt, the Internet of Things - otherwise known as IoT, IPv6 or Industry 4.0 - is probably the clincher. From baking to lighting, we look at the benefits of letting devices communicate.

By 2020 global IoT revenues will be 30 times those of the Internet, according to a prediction from Forrester Research, making it the next trillion-dollar communication industry, linking 25 billion smart devices and intelligent systems in operation around the world.But while the IoT is much more readily associated with consumer electronics, the Internet of Things (IoT) is relevant to automation in industry, and it is there that many of the early applications are to be found. After all, the bottom structure of IoT is composed of sensors, networks, services, and applications. And if, in industry, we had computers that knew everything there was to know about these things - using data they gathered without any help from us - we would be able to track and count everything, and greatly reduce waste, loss and cost. We would know when things needed replacing, repairing or recalling, and whether they were fresh or past their best.

Even so, the transition to IPv6 has been slow - currently, only about 5 to 10 per cent of users are able to support IPv6. Large-scale applications are still a long way off, because the business model is at an early stage and lacking in technical standards. Developing a common language is critical.

Then there’s the data problem. All of these devices will capture and generate huge amounts of data, which in the IT world is called Big Data. It refers to data sets so large and complex that they become awkward to work with using existing management tools. In contrast, with the control of process automation based on programmable logic controllers (PLCs), we have been used to collecting just a few data points, limited by simple serial networks and the negligible storage capacity of the control devices.

Product traceability

Sectors such as pharmaceuticals, food and automotive are feeling increasing regulatory and customer pressure to log process-critical data in ever-finer detail and in ways that are fully traceable. This means not only that the days of ‘pen and paper’ data collection are long gone, but that some of the slower ways of logging information electronically are giving way to much faster real-time collection and processing.

“The list of reasons why manufacturers should integrate individual line and equipment control systems into high-level enterprise management software is growing,” says Karl Walker, automation marketing manager, Omron. “More companies aim to measure Overall Equipment Effectiveness (OEE) in order to identify bottlenecks and weaknesses in the production process. But these calculations can only be arrived at once sufficient data is being collected from the entire operation and reliably stored.”

The US Food & Drug Administration’s (FDA’s) CFR21 part 11 regulations for the pharmaceuticals sector are now well-established, and foresee the collection and storage of production data without the possibility of human intervention. The EU’s Good Manufacturing Practice Directive lays down similar standards regarding reliable and secure records.

Ethernet-based networks, such as EtherCAT, allow every part of an automation system, from sensors to robots, to exchange information at speeds that were undreamt of in the past. This gives the controller instant access to every detail of production information. Once this controller is also connected to the enterprise-level - Enterprise Resource Planning (ERP), Manufacturing Resource Planning (MRP) or Manufacturing Execution Systems (MES), for example - the Industry 4.0 vision starts to be realised.

End-user IT departments will have to work more closely with engineering to allow shop-floor equipment to have direct access to enterprise-level systems without the use of middleware. Machine automation controller CPUs, such as the Omron NJ501-1_20, have this level of functionality embedded, thanks to program-free ‘wizard’ connections to relational databases, such as Microsoft SQL, Oracle, MySQL, IBM DB2 and Firebird. Pre-written function blocks then allow data from the machine or process to be mapped, inserted or updated into the database, or a query delivered to select specific data.

The latest controllers, such as Omron’s Sysmac NJ series, combine the reliability and rugged design of a traditional PLC, but use an open hardware architecture combined with software ‘engines’ within a single CPU - rather than multiple CPUs - to manage their different functions.

The wider benefits of this approach are neatly illustrated by a recent installation at reprographics specialist Ricoh. Its toner cartridge manufacturing and refilling centre at Telford in the UK is currently migrating towards using a direct link between ‘recipes’ on a product database and local NJ controllers on the production line. Until recently, its cartridge manufacturing has relied on conventional PLCs to ensure that the correct parts are used to assemble each of its wide range of products. The code on each component is scanned and checked. But in the past, this has meant that, when a new item is introduced, new code has to be manually added to each PLC - a laborious and time-consuming business.

In the future, Ricoh plans to use similar instant communication with an enterprise-level database to scan codes on used toner cartridges sent to its Telford site for recycling. In this case, an NJ controller will allow data, including the number of previous refills, to be checked. This will in turn enable the system to determine whether any given cartridge should be refilled again, while also updating the database.

This degree of intelligent, local querying and decision-making in a real-time manufacturing process was never achievable in the past, and can only be realised using IoT technology.

In the oven

In the food processing industry, in comparison to many other producing industries, productivity and profit margins tend to be relatively low.

Bread-making is an example of this. Small-​scale operations - ‘franchise bakeries’ - which supply up to 200 outlets in their vicinity typically produce several products on one line. The automation potential of such bakeries is high, but is frequently under-​estimated or considered too expensive.

Historically, small- and medium-sized operations perform the different process steps of bread-making, such as kneading, proving and baking, independently. However, overall equipment effectiveness can be significantly increased simply by networking the various isolated elements with the help of effective and cost-effective standard automation technologies. Bottlenecks, weaknesses, cost drivers, waiting times or item costs can be precisely defined and energy consumptions precisely assigned. The data from a networked, integrated system can be used to optimise processes such as oven-loading schedules or determine varieties and quantities per unit of time. This information can then be used to process and define processes and procedures.

Bakeries frequently suffer from significant discrepancies between financial planning and reality. Being able to collect actual production data including changeover times and other manual processes makes it possible to produce realistic form factors and contributes to transparency.

Sometimes networked systems are ordered for a greenfield project, but generally, it is existing plants that are modernised (brownfield projects). Within structures that have grown organically over time, companies need to retrospectively link isolated systems from different manufacturers and of different ages in order to collect all relevant data for higher-level management.

With the help of standard automation components such as the MES-IT interface module from Mitsubishi Electric, the necessary transparency can be achieved quickly and cost efficiently even in such upgrade projects. The interface module detects production data and test results from every manufacturing step and transfers them in real time to higher-level systems. These then produce transparent and reliable analyses of processes and procedures using Excel, SAP or pre-produced reports. Such equipment is said to be quick to connect to existing systems with no interruption to machinery processes.

Preventive maintenance

Optimising machine running times through preventive maintenance is a major issue in the food processing industry. If a machine suddenly fails, an entire batch is usually lost. This is expensive and damaging to productivity.

Modern control platforms, touchscreen operating devices and sensors with comprehensive, integrated functionalities as well as flexible maintenance and repair concepts can provide the necessary data transparency for simple implementation of preventive maintenance concepts.

The FAG SmartCheck from Schaeffler FAG, for example, recognises and reports the first vibrations of a system long before any noise is generated. That means there is substantially longer time to react before the machine fails. The maintenance cycle can be adjusted to production at an equally early stage. Machines can be easily retrofitted with the FAG SmartCheck. In bakery systems, for instance, it could be retrofitted to the loading system, mixer or conveyor belts.

Working in close cooperation with automation companies, systems engineers and users, the University of Weihenstephan in Germany is currently developing a universal standard for the bakery industry. The Weihenstephan standards define an interface for the connection between machinery and software. It will enable the full connectivity of all systems through a universal IT interface, making it possible to compare the cost and efficiency of individual lines. It is already established in the brewery and beverage sector and is being introduced into the meat processing industry.

Under the name ‘WS Bake’ it is now being refined for the bakery industry to enable the capture of the relevant information for this industry for quality assurance, weakness analysis, efficiency evaluation and energy management.

An aquatic robot network

Staying with developments emanating from academia, the Sunrise project, supported by the European Commission under the FP7 R&D programme, is designed to help researchers gather and share information on marine and freshwater environments through enhanced underwaterconnectivity. It is at the forefront of a revolution in communications, creating an ‘Internet of underwater things’ that will mobilise robots interacting together to work in groups.

Underwater robots will not only be able to work autonomously, having received instructions, but for the first time they will be able to communicate with each other and send data back via the Internet, regardless of swiftly changing circumstances and challenges to data transmission.

“The gaps in our knowledge of the underwater world are extensive. We know so little despite the fact that marine ecosystems are central to the health of our planet and vital to our economies,” explains project leader Dr Chiara Petrioli of the Sapienza University of Rome. “Identifying threats to oil and gas pipelines, monitoring the environment, protecting archaeological sites and finding out more about the geology of our planet - the ways teams of aquatic robots could help us learn more is endless.”

Just like marine mammals, the robots communicate with each other using acoustic signalling, but whereas a dolphin will adapt the way it signals according to what is around it, robots need to be programmed to do so, presenting researchers with the challenge of developing machines capable of responding to a rapidly shifting set of variables. “Salinity, temperature, interference in the form of waves or passing shipping, all these will change the range of effective communication,” says Petrioli. “This unpredictable environment is one of the key ways the Internet of things underwater differs from our land-based use of WiFi and the Internet.”

The need to respond reliably to the shifting environment means multiple robots are needed, so if one cannot communicate temporarily, another will take over the signalling. Schools of robots will carry a greater number of sensors and cover a larger area, cooperating and communicating together. Those operating them will send messages through modems transmitting acoustic waves. The waves are modulated to send information - but bandwidth is limited, meaning transmission rates are low. Additionally, sound waves only travel at 1500m/s, five orders of magnitude slower than radio communication in the air.

Results are starting to come in: work done in the port of Porto has helped to find a lost container. Now that the project has working prototypes, the next stage is to bring in new partners from different areas of interest and set up centres off the coast of the USA, in Dutch lakes, and in the Black Sea in Turkey.

Boom in smart lighting?

According to principalscientist Janne Aikio at Finland’s VTT Technical Research Centre, the lighting systems of the future could be multi-purpose devices, not dissimilar to smartphones. Smart lighting could also save as much as 80 per cent of energy compared to traditional lighting, and could be used to survey surroundings and transmit information.

“Forecasts suggest that smart lighting will become one of the key trends in the context of the Internet of Things,” says Aikio. “Demand for smart lighting is expected to boom over the next 10 years: as much as €7.7bn in 2020. The comparable figure in 2011 was €1.8bn.”

The first-generation smart lighting systems currently available are mostly designed for commercial use, and can integrate with building automation systems. In 10 years, such lighting systems could become everyday consumer goods. More and more wireless lighting systems that can be controlled via devices such as mobile telephones are becoming available.

“Smart lighting systems are becoming increasingly popular in both new builds and renovation projects. The next major step will be to integrate better sensors and new functions into lighting systems, which will allow the occupants of a room to adjust lighting with increasing accuracy and flexibility according to their movements and activities,” explains Aikio.

In the future, smart lighting technology will enable the direction, power and colour of lighting to be adjusted automatically according to whether a room is being used for watching television or eating dinner and according to where people are in the room. Lights positioned near windows will change colour according to outdoor temperature. Wall-mounted light switches will detect when a person enters the room. New smart features for light fittings will be available to download from the Internet. In office buildings, smart lighting technology will help shift-workers to adapt to changes in their circadian rhythms.

Back to earth

Inevitably, as we approach the fourth Industrial Revolution - the Internet of Things - it’s natural to ask whether we control technology, or does technology now control us? What happens when technology catches us up in the thinking business?
“Components of the Internet are increasingly carrying out sophisticated cognitive operations that we associate with the brain, such as finding patterns in large amounts of data. And they are increasingly making decisions autonomously, without human intervention. So just what it would take for the Internet to acquire a ‘mind of its own’?” said Murray Shanahan, professor of cognitive robotics at Imperial College, London, speaking at a recent event at Robinson College, Cambridge. “The Internet of Things and the Boundaries of Humanity.”

Jonathan Cave, senior tutor, Department of Economics, University of Warwick, believes the new world of the Internet is one of speed and complexity that may hide the consequences of our choices, and even change the way we think. “Understanding may not survive the journey into and out of the database, and critical thought may be replaced by the mechanical and collective logic of search. When we transfer our powers to algorithms the results may not be predictable, auditable or manageable. How can we trust machines? And how can we stop ourselves, when immersed in the constant flow of urgent and highly temporary decisions that the machines have generated, from turning into machines ourselves?”

Perhaps all the time it is humans asking these questions, we remain broadly in control of these philosophical issues.

Or do we? *