- By Hiroshi Yokoi
- Digital Transformation
- Process manufacturers lose too much production and profitability to operational problems and excessive maintenance spending.
- Digital transformation technologies can help, but they can bog down when they require new work practices.
- Outside help is often needed to implement transformational asset management.
Digitalization helps petrochemical companies create a plant environment where everything just works. Why don’t more companies practice it?
“A well-managed factory is boring. Nothing exciting happens in it because the crises have been anticipated and have been converted into routine,” says Peter Drucker. And if we expand the scope of that observation we can say, “A well-managed petrochemical plant is peaceful.”
While we would like it to be so, “peaceful” is not always the nature of process plants and units, where a variety of disruptive events are still part of the routine. These include:
- process upsets
- equipment failures
- feedstock variability
- energy waste
- corrosion damage
- safety incidents
- substandard out-of-spec product
- unscheduled interruptions.
The list could go on. The common denominator of all these is that they cause reduced profitability from lost revenue and increased operating costs. How bad is the problem?
“The impact of unplanned downtime in the process industries has been vastly underestimated. ARC estimates that unplanned downtime is costing the process industries about $1 trillion per year in lost production and revenues,” per the ARC Advisory Group Process Industry Downtime and Key Performance Metrics, 2017.
That $1 trillion figure is hard to grasp, since it is aggregated across the entire process industry sector. To put it in more specific industrial terms, any given plant typically loses 3 percent to 7 percent of its annual production due to these types of problems. Adding several percentage points to most plants’ productivity has a major positive impact on profitability.
How is this situation possible? Haven’t process automation companies been providing digital transformation (DX) technologies for years that are able to prevent these kinds of things? The answer is yes, yet a group of long-standing problems never seems to go away in many plants. Some companies and departments have challenges and a perceived excessive level of risk when adopting available tools and techniques. This leaves the efforts uncoordinated and any tools that are adopted remain unintegrated, resulting in minimal improvement. Managers simply shrug, accept the status quo, and live with the poor overall performance. Some of those companies have undoubtedly been shaken out, casualties of pandemic-related market stresses.
On the other hand, many companies have made serious efforts to adopt new DX tools and techniques, only to find a few stubborn problems refuse to go away. It is clear that there are fewer problems now than would have been the case years ago, but more remain than are tolerable. What is the cause of this situation, and what is the answer?
Understanding situational awareness
One of the root causes of many problems under consideration is situational awareness. Do operators, reliability teams, and maintenance departments have enough information available to them to understand the true situation? Is this information timely and accurate? The answer may be yes, or at least could be yes, because the supporting data is available; however many of those individuals may not fully realize what they are seeing. They are unaware of what the data is telling them, so they do not fully grasp the larger situation.
It is a common struggle. “While human error remains a primary reason for unplanned downtime, problems in the process or problems with the equipment controlling the process are more likely to blame. If that information is not effectively communicated to the operator in a timely and contextual fashion, your chances of an incident will go up significantly,” per the ARC Advisory Group Process Industry Downtime and Key Performance Metrics, 2017.
We can extend that to include the equipment supporting the process. The reliability and maintenance teams do their best, but often they unnecessarily touch something before there is a need, while something else runs to unscheduled failure. With the right approach, plants can stop wasting money and introducing risks by working on equipment that requires no maintenance.
The diagnostic data is probably available, but locked up in individual device database silos, where the right people might not see or understand it in time.
Progressing toward a peaceful plant
Let’s look at the picture in more positive terms. Every plant or unit has some period when it performs just like it is supposed to. Feedstocks are ideal, good operators are at the board, all the equipment is running perfectly, product is in spec, and output is at maximum. The people, the process, and the technology are all in peaceful harmony. There might be very minor disruptions here and there, but everything is manageable, and everyone is happy. How does this become an everyday experience?
It starts with data. Whether companies realize it or not, the equipment in their plants is already generating an enormous amount of diagnostic data capable of indicating the condition of equipment and warning of developing problems. However, most companies do not extract most data and get it to the people who could use it in a way that tells them something useful.
The data needs one overarching interface that can bring it all together as clear, action-oriented information, so the system can make the necessary links between different areas and sets. Where the performance and condition of one asset affects another and the larger whole, these links must be identified. The data, interconnections, and information must be available to the plant’s human operators and technicians in all areas, operations, maintenance, reliability, and management, via relevant dashboards able to support timely decisions, scheduling, and actions.
Why doesn’t this happen in all plants? Well, it is easier said than done. The technology side of things is the first step, collecting at least some of the diagnostic data from smart devices, and many companies have gone this far. This alone makes some positive impact, but usually not enough. The next step is where the effort bogs down: changing people and procedures. DX is not only about technology; it also requires organizational transformation to integrate people, processes, and technology.
Here is a case in point per the ARC Advisory Group Improve Asset Uptime with Industrial IoT and Analytics, 2015: “Preventive maintenance assumes the probability of equipment failure increases with use, and schedules maintenance based on calendar time, run time, or cycle count. However, data on failure patterns from four different studies show that (on average) only 18 percent of assets have an age-related failure pattern; 82 percent exhibit a random pattern. These data show that preventive maintenance provides a benefit for just 18 percent of assets” (figure 3).
This illustrates a common but incorrect assumption that equipment wears out in a predictable way over time. Most maintenance efforts reflect this belief, so many companies waste resources by fixing things that do not need to be fixed while allowing others to run to failure. ARC’s observation points out a critical disconnect between what data shows and underlying assumptions. Identifying when an asset is developing problems must be determined from quality diagnostic data, captured and presented to operators so they can see what is happening in a timely manner.
Even a small plant or process unit has diagnostic data available from at least 1,000 devices, but typically this is isolated in multiple databases, with little or no cross communication. All this data must come together into a single knowledge base so analysis tools can operate effectively and present findings to operators.
To reinforce the point, how many situations have there been where something major went wrong in a plant or unit that resulted in an incident? When investigators try to determine what happened, what do they do? They go back to process and diagnostic data looking for the root cause, but they look at it after the fact. The same data was available before the incident happened and could have warned what was coming, but was anybody looking at it then?
Closing the loop
The overall picture should be clear: data is available, and it can help reduce risk, prevent incidents, and guide maintenance. However, efforts to realize such practical improvements often run out of momentum because the company or plant cannot integrate the technology with its people and procedures. Old work practices remain in place, posing a major obstacle to condition-based proactive maintenance. The implementation loop never fully closes.
To make matters worse, implementation requires cooperation between the information technology (IT) and operational technology (OT) sides of the company, which cannot always be taken for granted. Also, creating the framework for the overarching data engine can expand the surface for potential cyberattacks, which causes some companies to hesitate.
Internal efforts vs. outside help
Carrying out a full implementation as discussed calls for a variety of resources and competencies, which many companies may not be in a position to assign on such a scale. Some aspects of this type of DX can be implemented incrementally, but this kind of undertaking calls for a more concentrated effort. This point alone is a likely reason why so many companies never complete the effort.
Most companies need help to follow through to a full completion because of the scale of the problem. This can be handled via a managed service designed for exactly this kind of project, with the provider delivering maintenance and development services to build optimized operations over the entire plant life cycle. This includes a plantwide maintenance and asset management platform that shows data on asset performance, reliability, and security concerns in a single window.
This has to happen at every layer of the operation (figure 4), including:
- L0 – Field instrumentation, actuators, and rotating equipment
- L1 – Basic process control and safety instrumented systems
- L2 – Supervisory control
- L3 – Production management systems
- L4 – Business planning systems
These overlap and touch on both IT and OT systems, so appropriate cybersecurity measures must be included at strategic points. All the elements can be integrated under an overarching managed service suite (figure 5) that works with the plant’s existing IT and OT networks.
This approach brings a long list of services beyond the reach of most companies’ internal capabilities:
- connections to all sources of plant data, even currently unused sources
- the application of sophisticated analysis tools to turn data into valuable information using effective dashboards
- maintenance and development services designed for optimized operations using proactive maintenance techniques
- an integrated digital services platform providing managed services that bring together process, people, and technology.
Fully integrated dashboards identify which processes are not compliant or trending in the correct direction, and simultaneously show which maintenance actions should be performed. The ultimate objective is to eliminate 100 percent of the risk, but there will invariably be trade-offs calling for a balance between maximizing production and effective maintenance. The system’s analytical tools help optimize these choices.
The benefits of a comprehensive program to digitalize asset and maintenance management are easy to visualize, but for most companies, implementing something on this scale using internal resources alone simply is not possible. For those who want to build toward these types of operations, managed services can fill the gap.
All images courtesy of Yokogawa.
We want to hear from you! Please send us your comments and questions about this topic to InTechmagazine@isa.org.