Making sense of complexity

Industry 4.0 and associated developments in the digital world are ushering in a new era that promises to reshape process control environments by making a vast range of tasks more manageable and understandable.

Process control environments are being reshaped by Industry 4.0 — the cyberphysical systems-based Fourth Industrial Revolution, which is itself being driven by technologies such as digitisation, the Internet of Things, big data, artificial intelligence and virtual reality. These technologies hold the promise of making many increasingly complex tasks such as fault detection and diagnosis, as well as process optimisation, more understandable and manageable.

What’s more, they are opening the door to tomorrow’s networked control centres. To be successful, however, these developments will require advanced collaborative work environments and a high-performance workforce.

Just as self-driving cars are set to transform the transportation sector, sophisticated digital automation programs are dramatically changing the operation of industrial processes. Driven by inexpensive sensors and powerful artificial intelligence algorithms capable of image recognition, vibration monitoring and much more, such programs are increasingly replacing human sensing. Indeed, in many cases, they are already making it possible for a single operator to manage an entire plant section.

Nevertheless, the picture is still far from ideal. Although integrated industrial information systems gather operational data to enable collaboration¹ and make real-time data available to appropriate individuals, McKinsey² has shown that in the case of offshore platforms, only a small portion of the resulting data is actually used as a basis for operational decisions (Figure 1).

Data availability: breaking down information silos

Why is the potential of Industry 4.0 not being more fully realised? The answer often boils down to complexity. Modern process plants are highly interrelated systems. A problem in one part of a process tends to propagate across different subsystems and plant components. Advanced automation systems add dynamic interactions between these components, making it difficult to obtain a clear assessment of potential problems. Added to this inherent systems complexity is the human element, which may involve the collaboration of a multidisciplinary team and the involvement of specialised expertise, often represented by an external supplier.

Unfortunately, collaboration between personnel from different disciplines, locations and organisations is often hindered by the fact that the information needed to solve the problem at hand is hidden within information silos. Many companies also lack the organisation and work processes to support multidisciplinary collaboration, and therefore tend to execute work based on a relay race approach instead of as a collaborative effort.

Figure 1: Only a small percentage of the data generated in many industrial processes is being used as the basis of operational decisions. For a larger image, click here.

Companies are hardly blind to these challenges. In fact, it is clear to most that, because of the trend towards growing levels of digitally generated data, they need to improve the way they work in order to be competitive.

Progressive companies are responding by improving workforce effectiveness through the introduction of digital technologies. For example, many are introducing bring-your-own-device (BYOD) policies, as well as solutions that enable employees to work effectively wherever they are, whether in the control room, on the shop floor or working remotely. The idea, in short, is to provide the same level of digital support to the industrial worker as has been available to office workers for years.

Information previously hidden within control systems or proprietary tools is now increasingly made available through improved connectivity and integration across systems and network layers. This trend is, in turn, opening the door to web-based applications that support consolidation of data from different systems and tools, making these easily accessible from any location. All in all, easy access to data and a common work environment is the first step to enabling effective collaboration to support process operation. Improvements in analytics and visualisation techniques also help workers to make sense of increasing amounts of data.

Other technology trends are also supporting a new collaborative approach to work. Videoconferencing technology has matured, allowing many companies to have remote operation centres that support local control rooms with open video links between locations. High-quality videoconferencing technology is also available from mobile devices or personal workstations, enabling operators to obtain instant access to remote expertise.

In combination, the introduction of digital technology for easy access to information, independent of location, and the proliferation of videoconferencing to support remote collaboration, are blurring the boundaries between local and remote operation.

Insights into process operation

Modern automation systems can cover most aspects of normal operation but also handle many anomalous situations. Advanced control techniques such as model-predictive control (MPC) and state-based control (SBC)³ allow the automation of very complex tasks, such as the start-up of an industrial plant. Indeed, automatic control performs better than typical human operators. As a result, operators are less and less involved in the inner control loops that are directly in contact with processes. Their tasks are increasingly shifting to supervisory control⁴, where the operator manages and supervises a large number of control modules. But being less involved in direct process control also means fewer possibilities to develop a feeling for a process by training on the job — a problem that was dramatically illustrated with the loss of flight AF447⁵.

To be able to take over when automation fails, operators need higher qualifications and a profound understanding of underlying technical processes, automation systems and control modules. Simulator training is necessary to develop a feeling for processes.

Operators should also be deeply involved in the optimisation of process operations, because this keeps them involved and helps to build up the required knowledge that allows them to take over in case of automation failure.

Not only can Industry 4.0 have a profound influence on information flows and availability, it is also set to have a huge impact on industrial quality control⁶. Big data techniques make it possible to distil historical process data into algorithms that can predict the quality of production⁷. Developing problems can be detected early and countermeasures can be taken before impact becomes significant. Previously, it took an operator many years to accumulate comparable experience.

Remote expertise can also play a decisive role in failure avoidance. For example, in the case of the Deepwater Horizon oil spill⁸, the investigation report clearly states that one major factor contributing to the accident was the incorrect interpretation of available measurements. Quite likely, advice from highly qualified remote experts would have avoided this accident.

Process performance optimisation

Key performance indicators (KPIs) for process operations in areas such as control loop performance, alarm management, energy efficiency and overall equipment efficiency are described in detail in Hollender et al⁹. Managing these KPIs is not a classic operator task but is becoming more and more important in terms of ensuring good production performance. Areas such as operations, maintenance and analytics need to be managed holistically to achieve the best results. Many of the associated tasks can either be performed by centralised internal service centres or outsourced to specialised external service providers.

Typical goals are increased production plant throughput, efficiency and uptime¹⁰. These can be achieved through a structured approach to revealing the sources of and responses to process variations and upsets. By reducing process variations, operational flexibility, plant regularity, safety and integrity will be increased, while off-spec production, energy costs, environmental impact, operator stress and equipment wear will be reduced.

An example is provided by Dow Chemical, which has introduced a global analytics layer that distils vast amounts of data into readily understandable information and metrics¹¹. As a result, experts from a centralised analytical technology centre can now support plants globally to determine manufacturing obstacles, improve efficiencies and develop best practices.

Tomorrow’s control rooms and operators

As mentioned above, most simple parts of traditional operator work have been taken over by automation. Modern operators now have a very different profile from their predecessors. They supervise large numbers of control modules and must be able to quickly diagnose complex situations, collaborate with various support units and coordinate field operators and maintenance personnel. They decide when it is time to bring in external expertise and manage the integration of remote experts. To achieve their full potential, they need a work environment that supports these activities.

In this connection, a challenge will be how to design collaborative environments that will replace traditional control rooms. Often, such centres will not need to be physically close to a process, but will need to be much better integrated with remote service communities in their own company, as well as associated service providers and suppliers.

The involvement of experienced control room designers from an early stage on will be essential to the design of next-generation collaborative operation centres. Such centres will require a totally new approach that takes the full potential of Industry 4.0 into account. As the traditional way of building control rooms becomes obsolete, new best practices will have to be defined.

The new centres will contain fewer operators, while the roles of operators will evolve from reactive to predictive problem-solving and analytics. To achieve this, operators will have to be highly qualified and capable of interacting with many other specialised functions, such as IT/OT support, multifunctional support, technical and remote support, asset risk management, alarm handling, safety, cybersecurity and maintenance management.

Frequent interactive communication with a very broad range of remote service specialists — something that is still rare today but will be the norm tomorrow — demands a new look at factors such as room layout, working zones, monitors, cameras, analytical tools and remote collaboration workspaces.

Although working in a 24/7 environment can be exciting, it can also lead to reduced life expectancy. With this in mind, a well-designed operator platform should be able to be adapted and even automated to meet each user’s needs. For example, the distance between eyes and monitors can be automatically adjusted to reduce eye strain, and lighting can shift throughout the day to optimise circadian rhythms. Furthermore, big data analytics makes it possible to create a data-driven ‘day-by-day’ improvement program for operators.

Next-generation operators

As control centres metamorphose into collaboration centres, a major challenge will be to attract the next generation of operators — a generation for whom the connection between personal ergonomics and health is fundamental. This increasingly important factor calls for workplace designs that not only minimise acoustic disturbances, while optimising factors such as illumination and air quality, but take into account aspects of the psychosocial working environment such as gamification, collaboration, individual space, flexibility, learning, sustainability, social presence, emotional engagement and creativity. All in all, technological development driven by operators’ needs can be expected to transform tomorrow’s working environment.

References

1. Pfeffer J, Graube M, Reipschlaeger P et al 2015, ‘Towards Collaborative Plant Control using a Distributed Information and Interaction Space’, IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA).
2. McKinsey Digital 2015, Industry 4.0: How to navigate digitization of the manufacturing sector.
3. Hollender M 2009, Collaborative Process Automation Systems, ISA.
4. Sheridan T B 1992, Telerobotics, automation, and human supervisory control, MIT Press.
5. BEA 2012, Final Report: On the accident on 1st June 2009.
6. Lorenz M, Rüßmann M, Strack R, Lueth K L, Bolle M 2015, BCG Perspectives: Man and Machine in Industry 4.0.
7. Atzmueller M, Klöpper B, Mawla H A, Jäschke B et al 2016, Big data analytics for proactive industrial decision support. atp edition, 2016, vol. 58, no. 9, pp62–74.
8. National Commission on the BP Deepwater Horizon Oil Spill and Off-shore Drilling 2011, Deep Water: The Gulf Oil Disaster and the Future of Offshore Drilling: Report to the President.
9. Hollender M, Chioua M, Schlake J, Merkert L, Petersen H 2016, KPI-based Process Operation, atp edition, 2016, vol. 58, no. 9, pp52–61.
10. Bieker H P, Slupphaug O, Johansen T A 2007, Real-Time Production Optimization of Oil and Gas Production Systems: A Technology Survey, SPE Production & Operations, vol. 22, no. 4, pp382–391, Society of Petroleum Engineers.
11. Colegrove L F, Seasholtz M B, Khare C 2016, ‘Big Data: Getting Started on the Journey’, AIChE CEP Magazine, March 2016, AIChE.

This article is based on an article previously published in ABB Review.