A fourth industrial revolution of digital automation technology is already under way and accelerating. This is not just further automation of equipment and processes but automation of capabilities and decision making that are now only within the abilities of us as engineers, writes Professor John Barrett, Nimbus Research Centre.

Unless we understand what is happening and the technology behind it, and take control over how it is to be appropriately used in engineering, we risk a growing dominance of the tech industry in the engineering domain that will mirror growing tech industry dominance in other areas of human activity.

Our upcoming Online Seminar  

Join Professor Barrett and his colleagues from the Nimbus Research Centre by booking a place at our CPD online seminar entitled The Role of the Internet of Things (IoT) in the Energy Sector on Thursday, 25 March

The seminar will address the IoT Approach to Energy Efficient Buildings, IoT for a Decentralised Energy Grid, AI & IoT for Carbon Emissions Reduction and the Challenges and Solutions of IoT Implementation. An industry case study will also be covered.

This CPD seminar will be of interest to engineers, technicians, contractors, academics, project managers and consultants interested in IoT and how it can be applied in energy conservation technology, renewable energy technology and energy system development to ensure an economically and environmentally sustainable and secure energy system in Ireland.

How did we get from slide rules to today’s mobile digital engineering?

From an extreme reductionist viewpoint, engineering is a discipline of calculation – calculations for the design of 'engineering things' and calculations to extract meaning from measurements taken from these 'things' to ensure that the 'things' are working as designed.

Figure 1: 1950s IBM Calculators Advertisement (Wikimedia Commons)

Over time, what has changed are the ways we do our calculations and how we collect our measurements. The IBM advertisement from the 1950s in Figure 1 reflects the move from slide rules to electronic calculators with one IBM electronic calculator claimed as equivalent to the computing power of 150 engineers with slide rules.

Each global digital evolution brought a step forward in the process of digitisation of engineering. In the intervening decades, we have moved on to mainframe programmable computers, pocket calculators and desktop computers and now the mobile computing of laptops, tablets, smartphones and the massive computing and storage capabilities of the cloud that we can access via the mobile internet in the office or in the field.

How many engineers with slide rules does a tablet with mobile cloud access represent?

We see a similar trend of digitisation in how we collect measurements from our 'engineering things' – moving from in-person reading of analog gauges to remote monitoring using digital sensors attached to our engineering things.

The way we interact with our designs and measurements has also been digitised – from paper to digital touchscreens.

Engineers have fully embraced digitisation to the point that smartphones, mobile computing and the internet have become essential engineering tools we use to do our work.

At each step in the evolution of digitisation, engineers have used their skills to adapt and apply digital tools to their work and, through the parallel evolution of education and training, have ensured that engineers, both experienced and in education, acquire the engineering and digital skillsets needed to do their work in a digital world.

It is also worth reflecting, given the way the world has gone in 2020, that it is this progress in digitisation that has allowed engineers to continue to do their work and, in many cases literally keep the world running, even though locked down in their homes.

So what is next in the digitisation of engineering?

The convergence of low-cost mobile internet, cloud computing and remote monitoring has allowed the emergence and growth of the 'internet of engineering things' – the engineering subset of the internet of things (IoT): the 24-7-365 monitoring and control of anything, anywhere – from the ventricles of the human heart to a rover on Mars – from any digital device. This is commonly referred to as the 'fourth industrial revolution' or 'Industry 4.0'.

The IoT concept is not new to engineering: from industrial control to building management systems, for decades we have had the ability to measure and control the 'engineering things' we have designed, first using wired networks and later using dedicated wireless networks.

The difference in Industry 4.0 is that we can use the massive global infrastructure of the mobile internet as the communications backbone, giving us global monitoring and control literally at our fingertips from wherever we happen to be.

At a black box level, IoT technology is neither complicated to use nor expensive to buy because this is the way digital and internet technology has evolved: massive numbers of low-cost electronic devices and systems that almost anybody can use.

In IoT, sensors on our 'engineering things' are connected to electronics and communications 'edge nodes' that send sensor data to us via the internet; our control systems use the internet by return to close the loop.

In many cases, the technology is now 'plug-and-play' and as simple as connecting our phones to the nearest Wi-Fi signal. In some cases, it is a one-way data flow only – for example, we monitor the structural stability of a bridge or the water level of a river to get early warning of failure or flooding; in manufacturing systems, in autonomous vehicles, in process control, in environmental control systems and many others, it is a two-way flow of measurements and real-time control.

We may have firewalled secure IoT intranets within our factories, buildings or infrastructure, we may simply use the public internet for less critical applications or we may use a hybrid with a secure IoT intranet for local operations combined with the public internet for data sharing across the globe.

It can’t be that easy?

The technology is that easy and becoming easier and cheaper all the time but this, paradoxically, is where challenges arise:

  1. Hype. At the time of writing this article, an internet search of 'internet of things' and 'engineering applications' yields 1.7 million results, many from technology companies promising the sun, moon and stars if you invest in their IoT technology for your engineering application; it is an ever-increasing challenge to identify the safest and most useful IoT applications for your individual needs and to select from the ever-increasing numbers of Industry 4.0 solution providers.
  2. We can collect so much data from so many engineering things that the data volume and flow rate can overwhelm us – it becomes increasingly difficult, even with the best data analytics, to actually extract useful information. In many applications, we have to apply machine learning techniques that themselves present challenges in how to use them appropriately. Not least is the removal of the engineer from the loop and the devolving of control to black-box machine learning and artificial intelligence (ML-AI) systems whose algorithms we may not fully understand and therefore cannot, perhaps, fully trust. ML-AI systems are becoming increasingly powerful and will undoubtedly be invaluable tools for engineering – a challenge is to judge how much of the analysis and decision process to hand over to ML-AI.
  3. Should we therefore not always keep an 'engineer in the loop' – there is, after all, still no ML-AI system (or even a near prospect of one) that can replace years of engineering experience and that ability to spot an anomaly and say “hold on, there’s something not quite right here…”. For the engineer in the loop, we have to filter the sea of IoT data down to what is truly useful and present it in an easily assimilable way. Augmented, mixed and virtual reality – xR – are moving across from the gaming world as tools that can greatly help engineers with data visualisation. Again, a challenge is the brushing away of xR hype to identify useful xR engineering applications.
  4. The technology that powers the internet has been demonstrated time and time again, and at massive scale, to be fundamentally insecure. Not so much a problem, perhaps, if a hack causes a video server to crash but a serious problem for an autonomous vehicle or a national smart electricity grid. How do we manage cybersecurity for engineering systems that were never before 'cyber-anything': vehicles, machinery, plant, utility networks and grids, traffic management…even the human body itself as a biomedical engineering system with a growing availability of remote monitoring and control medical implants.
  5. For safety and security, we can monitor the personnel in buildings, factories, construction sites and other on- and off-site workplaces but when does protection cross over into surveillance and where is the boundary between necessity and the right to privacy?
  6. Industry 4.0 promises dramatic new capabilities in automation with the ability of Industry 4.0 smart systems to outperform human operators across a myriad of tasks that were previously only within human capability. This automation will largely (but not exclusively) displace workers from lower-skilled jobs with potentially major personal and societal impacts and a skew towards larger-scale impacts in economies with a high-dependence on lower-skilled jobs.
  7. Engineers in training must acquire and use an ever-widening array of digital skills and these have to form an ever-growing proportion of engineering curricula: what is the optimum balance between fundamental engineering knowledge and digital skills? Should we hold the line that the engineering knowledge must always predominate rather than trying to create a digital-engineer hybrid? Should we instead use parallel engineering education streams to create a 'symbiosis' by giving digital experts enough engineering domain knowledge to work as partners with engineering domain experts given enough digital domain knowledge so that both can form an effective partnership? Perhaps an even greater skill requirement for both partners in the symbiosis is the ability to discriminate between tech hype and real utility.

It is not sufficient for the engineering section of a company to start monitoring and control using the IoT, the entire company must become data-driven and this brings many organisational and business challenges along with the need to 'upskill' the entire workforce. The scale of this challenge should not be underestimated.

So, what should we do?

Despite these challenges, Industry 4.0 will, like calculators, computers and the internet, change how we 'engineer' and, like those evolutions, it will give new engineering capabilities that were not previously possible or, in some cases, not even dreamed of.

It is not a case of 'it will happen' – 'it is happening' and, just as it is now unthinkable for us to do our engineering without computers and the internet, we will look back in relatively few years and say “how did we ever do our work without the IoT”.

This is why engineers, with the support of engineering-informed technology domain experts, must understand both its capabilities and limitations so that we the engineers and not the tech companies decide how it can be used optimally and safely and, in the best practice of engineering, for the greatest benefit of people and society. 

Author: Professor John Barrett, Nimbus Research Centre