Automation, Algorithms and AI are all part and parcel of our daily lives, playing more and more of a part in our daily choices and decision making.
The predictive text on your mobile phone messaging systems for example, (usually leading to an embarrassing gaff), to Google Maps asking us if we would like an alternative route home from work based on traffic information it has gathered from a myriad of other mobile phone users, these technologies are evermore present in our lives.
Smart technology in our homes now lets us know when someone is at the front door, when the temperature has dropped upstairs and the heat has come on, whilst our refrigerators, dishwashers and washing machines all have apps and technology designed to save us energy and money based on our behaviour and personal settings. We can let a refrigerator know that we plan to be on holiday for two weeks, through an app, and the power usage will decrease over that period based on us having less use and less storage during that time. We are becoming more and more comfortable with devices being smarter, more connected and responding to input from their environment.
“we are becoming more and more comfortable with devices being more connected”
As I sat down to write this article, a friend suggested I look up a new AI platform called ChatGPT1. ChatGPT is an OpenAI platform that allows you to ask questions and have a trained model ‘ChatGPT’ answer these questions, admit its mistakes, and basically generate intelligent responses and learn as it goes. You can do anything from diagnose a problem in Microsoft Excel, discuss the weather of the day, to generating some fairly credible content on the topic I am writing about! It was very tempting to just ask ChatGPT about the use of AI in control of work and have the generated content be the article published, but I’m not sure the response would have made very good copy for this magazine. Despite it being a relatively new platform, this tool could provide very credible information and content on a range of subjects which would make you think you were talking to a seasoned and lightning-fast researcher on the other end of the line.
It does make you think though. If we are ceding decision making, choices and preferences in our daily lives to AI, what about the automation that is on the way to our workplaces and that which is already present in our workplaces? If we are happy to allow tools with vastly superior processing power and resources to write our content for us, answer our questions, and provide us with our choices, where do we draw the line in our various industries and workplaces?
Industrial Internet of Things (IIoT)
The advent of the fourth industrial revolution (4IR or Industry 4.0), introduced a new dawn of industrialisation focussed mainly on automation, connectivity, machine learning and use of real time data. Many referred to this as the Industrial Internet of Things (IIoT) bringing a new dawn of connectivity, opportunity and efficiency to our plants and workplaces. Terms like ‘smart manufacturing’ were coined to describe the interfacing of physical production assets and operations with big data projects, machine learning and smart digital solutions improving and automating manual and cumbersome processes that had existed for decades.
All this change brought about not only the procurement and management of new technology but required that companies go back to the drawing board on some processes, to look at how the business operates and ultimately grows alongside this revolution and technology. A ‘smart factory’ now became the goal in terms of leveraging this new approach and new technology all with the aim to make safer, more efficient and growing businesses. The solutions employed alongside the interconnectivity of plant and personnel aimed to make a digital workforce more informed, better equipped and ultimately help grow the business in terms of efficiency and output.
Working with businesses across the globe, mostly in the Energy and Manufacturing space, I have seen first-hand the application of some of the new technologies and solutions in the field. For example, the rise of predictive analytics is giving companies the ability to accurately forecast when maintenance is going to be required, well before a problem is likely to arise. Companies are therefore shifting from asking the question “what just happened?” or “why did it happen?” to the more proactive line of questioning such as “what is going to happen?” and “what can we do to stop similar things from happening in the future?”. In the past, companies relied on experienced personnel who knew their plant intricately to predict and plan for likely problems and outages. To a large extent they still do, but married with the data and real time information they can receive to make decisions against, it goes towards ensuring assets have a longer, more efficient and productive lifespan.
“companies are shifting from asking “what just happened” to “what is going to happen?””
For most companies the technology required, the change in practices and the organisational overhaul are quite considerable and the benefit to the business must be both quantifiable and beneficial to the goals of growth and expansion going forward.
Industry 4.0 is a wide topic, and the implications and ramifications very diverse. Given my experience within the Energy and Manufacturing sector, especially when it comes to the physical work and inspections performed onsite and the safety compliance of such, in this article we are going to look specifically at Control of Work (COW) and the extent to which automation of various types can benefit, enhance and streamline safety practices or, in some cases, have an adverse effect.
Control of Work (COW)
If you aren’t familiar with COW, it is a term of increasing prominence across various industries relating to the processes, documentation, people, assets and information relating to how companies control physical work being performed onsite. Defined as the management of business-critical maintenance processes, to ensure that tasks are carried out safely, COW touches every part of the areas of performing maintenance and carrying out routines. Incorporating documentation and processes relating to Risk Assessments, Permitting, Isolation Management, Working at Height, Confined Spaces, High Voltage and many more category areas, the aim is to link these processes and areas together to provide a compliant, safe, efficient and digital way to enhance the control of work across our workplaces.
COW as a standalone concept isn’t necessarily something you would think of when you think of Industry 4.0 or automation. It is fair to say that the increase in real time data available, the technology in place and the connectivity on the shop floor has meant that COW cannot go overlooked when considering Industry 4.0. As processes are being re-mapped and traditional manual tasks and inspections becoming digital and more connected, so the environment in which Risk Assessments are performed becomes more informed, Permits have reliable real time data to base content and decisions on, and LOTO sheets can incorporate current positions of valves and other information pertinent to making a safe and effective isolation. It’s not just the environment and the asset information but also the many different departments and disciplines that go to make up the team that perform work and maintain the assets as well as ensure continued production of goods and products.
“to what extent should decisions and judgement be made by automation and big data”
The information and data points around the personnel involved in the COW model are only one aspect, but it can be made up of internal staff, external contractors, and a range of factors from levels of access to levels of authorisation which impact the ability to perform the work competently and efficiently. Ensuring that data is communicated effectively to all parties involved is a huge part of ensuring the safe execution of tasks on plant.
This all leads to a discussion that is going on currently within various industry verticals within manufacturing ‘How much Automation, AI and Algorithms are good for safety processes and systems?’. Referring to safety systems here, I am talking about a digital method of incorporating the control of work processes and documentation. One consequence of Industry 4.0 has been a vast improvement and deployment in Enterprise Resource Planning and Digital Control of Work Systems across the board. Given the increase in choice and sophistication of these digital systems, to what extent should decisions and judgement be made by automation and big data and how much should remain within the domain of the expert engineer or safety person?
There is likely no ‘one size fits all’ answer to this question, as with Industry 4.0, the solution and the choice of technology must fit the manufacturing scenario and the application rather that fitting the business around the technology. In terms then of some specific ‘on the ground’ examples, let’s look at how automation and interconnectivity can either help or hinder decision making when it comes to Control of Work processes and documentation.
Assessing the work that needs to be done will involve needing to know details of the asset, the personnel involved, and the task to be performed. In most modern ERP systems the ability to automatically generate work orders based on predictive maintenance algorithms has greatly increased the amount of work that can be done to protect valuable assets. This is a huge plus for the operations and maintenance of the plant and ensures planning can be carried out ahead of time rather than having to break into production for an emergency task to correct a problem. On the downside, sometimes automated work orders can generate an unmanageable amount of PM backlog when other tasks become a priority.
One major aspect of reviewing a work task is the proper analysis of Risk and Hazards associated with performing the work, who they will harm, and what mitigations can be put in place to ensure the risk is reduce to as low as reasonably possible (ALARP). Having connected systems and a data historian ensures that trends and pertinent information can be made available to the Maintenance or Safety Coordinator when performing this assessment. Trend analysis around previous incidents of work of a similar nature and up to date asset notes and information can give the best possible view on the environment and the task. Many digital systems now allow for Risk Assessment to be performed in the field on a tablet or similar device, to give the operator all the information they need at their fingertips whilst being in situ at the place where the work is going to be performed.
“the physical intervention of human sign off and interpretation of data is invaluable”
Permit to Work
Permits rely on information about plant conditions to accurately relay to personnel items such as the limitation of the work area, what to be aware of in the work area, the various gases that might be present and a range of other items. Having interconnected systems allows for real time information to be included at the point of creation, to give the personnel and accurate picture of the plant. Whilst it is good for all this information to be collated and presented on the permit to the responsible parties, the human input of decision making and distilling the information into a real-world situation on plant cannot be downplayed. Just because one set of conditions and information were good for a job the first time around, doesn’t mean that they can be relied upon (solely) for the second time the job is performed. Plant conditions can change, and other factors come into play that might not necessarily be represented in ‘data’ collected from the plant. This is where the physical intervention of human sign off and interpretation of data is invaluable in the process.
“technology can help us improve but can’t take away the need for human intervention”
Lock Out Tag Out
When performing a LOTO, the diagrams and asset structures available at most plants through sophisticated ERP and DCS systems means that engineers are normally very informed about what needs to be done for a job. A lot of industries have gone through the process of generating and approving LOTO libraries for repetitive jobs and routine tasks, which no doubt have saved a lot of time and effort over the years. Whilst relying on automation to give us our isolation plan might sound great, there are pitfalls. Experience from an engineer might tell us when a valve has been passing, but not yet marked up in the digital system, meaning that we still need to widen the scope of isolation beyond what the digital system is telling us, for example. It is also tempting to rely on a digital confirmation on the close/securing of an isolation point solely for the work to commence, but a physical check and the application of a Tag or Lock can add the physical barrier required to ensure safety of the personnel at work.
The above are some examples where, when it comes to the control of work, we need the data and embrace the use of automation to help us make selections, but when it comes to signing off on the safety of engineers in the field, we still require human input and judgement to ensure safety. There are a lot of areas where technology can help us improve in safety on our plants, but we can’t yet take away the need for human intervention and sign off on these most precious of decisions.