Click to learn more about author Chen Linchevski.
Predictive monitoring is the latest use case for predictive analytics in process manufacturing, which stands to bring enormous value to the process industry.
By combining machine learning (ML) analytics with the rich and extensive data gathered by IIoT sensors across the plant, predictive monitoring systems can pick up on even the smallest of anomalies and detect whether they represent inefficiencies or potential incidents, or whether they are simply part of the plant’s normal background noise.
Unlike typical control room monitoring, predictive monitoring can pick up on emerging issues before they become observable to the human eye or ear. Very often, by the time someone notices a change in operating conditions, altered sound patterns or vibrations, or a drop in product quality or quantity, the problem has been growing for some time. It’s frequently too late to apply a quick and low-cost fix.
Predictive monitoring solutions are typically more efficient than employee-based control room monitoring programs, since AI-powered solutions never get tired, never forget protocols, and never take sick days. Classic monitoring relies on workers circulating throughout the plant and noticing changes, as well as setting rules for when certain thresholds are met, so it is limited by the number of employees and how they set the rules – often through a combination of knowledge and intuition. Thus, it tends to focus on the most expensive and critical items of equipment. A predictive monitoring solution includes every valve and pipe, not just high-value or high-use parts, to paint a constantly-updated picture of processes and operations across the entire plant.
COVID-19 Accelerated Adoption of Predictive Monitoring
Although adoption has been slow, COVID-19 made a big difference to attitudes to digital tools in general and predictive monitoring in particular. The pandemic sharpened the imperative for disaster planning, reminding plant managers that it’s crucial to build in resilience and remote operations.
Plants that already had digital tools, a digital culture, and had undergone digital transformation were far better placed to weather the storm than those that still saw plant monitoring and data gathering as manual, human-led processes. Companies that weren’t prepared for remote operations are quickly upgrading their systems before the next crisis. Adoption of IIoT saw about 5 years of growth in just 8 weeks of 2020, primarily for predictive maintenance use cases.
To prevent spread of infection, most plants ran only a skeleton staff on-site. Under “normal” conditions, many nascent issues are spotted by expert employees in the course of their regular work, but with fewer people present, those early warning signs would go unnoticed. A predictive monitoring solution can spot them, though. Employees at plants with predictive monitoring solutions are able to oversee processes and parts remotely, ensuring that the entire plant continues to operate as needed.
Control room monitoring traditionally follows a rules-based system. Critical processes and pieces of equipment have defined thresholds that specify “normal” operations. When the thresholds are exceeded, it triggers a rules-based alert. However, the thresholds are set manually and may not be correct, or could reflect a past reality which is no longer relevant for the current situation. Additionally, a number of serious issues can arise without overstepping the rules-based system; this happens because humans, unlike AI, cannot possibly take every eventuality into consideration. An AI-powered solution doesn’t rely on fixed rules. Instead, it observes “normal” conditions and learns to recognize both acceptable variations and alarming anomalies.
Additionally, early alerts about potential incidents meant that skeleton maintenance teams could schedule repairs for the most convenient times, long before things reached an emergency situation. This helps reduce the number of employees at the plant at any given time, too.
The Many Benefits of Predictive Monitoring
The many benefits of predictive monitoring are thrown into sharper relief by the pandemic.
- Extend equipment life cycle
Predictive monitoring alerts teams to emerging incidents before they become serious, enabling plants to repair parts before they fail entirely and have to be replaced. As well as saving money on labor and parts, this allows plants to replace parts at longer intervals.
- Improve plant efficiency
Fixing small issues before they snowball means that all equipment is in peak operating condition at all times. This helps remove minor impediments that affect production quality and quantity, significantly boosting plant revenue. At a time when demand is in flux and supply chains are still disrupted, plant efficiency is more important than ever.
- Reduce unplanned downtime
Early alerts mean that maintenance teams can deal with issues that are small and easy to fix, instead of having to shut down production to make major repairs or replace an entire part. Lost production due to unexpected downtime costs process plants far more than all the costs of repair, labor, and replacement parts put together.
- Act strategically instead of chasing fires
Predictive monitoring gives process plant managers more visibility into the entire plant, adding awareness of what is working perfectly and what could be refined. Many process engineers spend the majority of their time reacting to unexpected incidents, events, and anomalies, which hampers their ability to plan and implement strategies towards operational excellence. Predictive monitoring places them in control of the situation so they can focus on long-term planning.
- Increase employee safety
Predictive monitoring solutions help make the plant a safer place for all employees. Equipment will be in better condition, potentially dangerous incidents are prevented before they grow, and there’s no need to request employees to enter hazardous conditions to check up on equipment, gather data, and monitor developing events.
Predictive Monitoring is Swiftly becoming a Must-have
Companies with digital solutions like predictive monitoring will be better placed to bounce back, ramp production up again as the markets reopen, and bring employees back to work safely. According to recent research, the global big data analytics market in the manufacturing industry was worth $904.65 million in 2019, and is projected to reach $4.55 billion by 2025, at a CAGR of 30.9%. This market expansion will help drive adoption of predictive monitoring solutions.
COVID-19 changed more than the number of employees present at the plant at any given time. It also disrupted supply chains and reshaped consumer demands, increasing the imperative for AI-driven data analytics like predictive monitoring that keep process plants on top of market changes and fluctuations. Predictive monitoring can ultimately help process plants to make better use of limited resources, improve uptime, enhance production quality, and drive profitability and productivity across the company.
With uncertainty continuing around business conditions and virus variants, and employees likely to want to continue to work remotely wherever possible, we expect to see predictive monitoring adoption increase significantly in 2021.