State-based control isn’t new but chemical companies now are showing increasing interest in the approach, according to vendors such as ABB, PAS, ProSys and Siemens.
This has much to do with the 2010 launch of the ISA106 committee on Procedure Automation for Continuous Process Operations by the International Society for Automation, Research Triangle Park, N.C. Prior to this, only a small number of chemical companies, notably Dow, had extended the concepts of batch operations into their continuous manufacturing environments. The new standard brings an opportunity to formalize this activity.
“Some companies had automated portions of startup procedures, equipment switching procedures, etc., but few did things on a grand scale. And part of the drawback was also the horsepower and performance factors on the process controllers to support running large-scale procedures for long periods of time,” notes Dave Huffman, OGP business development, chemicals, ABB, Wickliffe, Ohio.
ISA106 defines a process state as “a definable operating condition of process equipment as it progresses from shutdown to operating and back to shutdown. Each process state represents a unique operating regime that supports the process equipment’s objectives of processing an input into a desired output.”
“To more clearly state this in terms of the state-based control objective, it is a definable operating state in which the control system needs to take action to achieve the desired objective for that operating state,” says Dustin Beebe, president/CEO, ProSys, Baton Rouge, La.
So, a state is any well-defined phase in which a system can exist. The phase can be either steady state (such as distillation) or transitional (such as heating up), where specific process values are being controlled to a desired set point.
“Each phase must include definition of the specific inputs that allow entry into the phase, as well as those required to exit to a subsequent state. Unlike a typical sequence of a batch system, which is executed and ultimately completes, a state-based control system is always in one — and only one — of its defined states. Examples of states may include: stopped, ready to start, starting, production, stopping. fault, etc.,” explains Paul Morgan, senior consultant in process automation, Siemens USA, Spring House, Pa.
But why choose state-based control in the first place? “There are many reasons why state-based control is the way to go. Process plants have become increasingly more complex, leading to greater potential for accidents due to human error. The challenge is to optimize production rate and increase uptime without compromising process safety. State-based control presents a proven and effective strategy for achieving safe and profitable production. It utilizes historical production information… and the knowledge of the best operators to automate complex processes,” says Eddie Habibi, CEO, PAS, Houston.
Morgan points to four key benefits: better safety (through risk reduction), improved alarm management, reduced operator workload, and better maintenance planning with asset management.
“Improved safety is one of the key benefits to utilizing state-based control. The startup and shutdown operations of the continuous process are arguably the states most likely to encounter a dangerous demand to trip, yet these states are frequently manually monitored and controlled. Couple this with the high demands on the operators to achieve full operation, and mistakes are more likely to occur,” he says.
Such situations also can complicate alarm management, which he describes as a critical issue at the corporate level. With state-based control, alarms not appropriate during certain phases of the plant can be concealed from the operator using smart alarm hiding. As a result, the operator doesn’t get distracted by nuisance alarms and can retain focus. However, these alarms still are recorded in the archive system for completeness.
Alongside the lack of distractions, the automated nature of state-based control also helps reduce operator workload by making a far simpler and less stressful task of both startups and shutdowns.
Lastly, with state-based control, maintenance of the plant’s assets can be better predicted. Starting and stopping devices typically is harder on them than continuous operation. “Having visibility into these demands allows scheduled maintenance, reducing unplanned plant downtime due to failures,” he adds.
ABB’s Huffman notes that any company interested in continuous improvement in its manufacturing discipline will have an interest in procedural control. After all, he points out, the company already has written procedures that generally describe, in a lot of words, how the process should be started up or shut down, how to prepare the process for maintenance, how to make product grade changes, etc.
But, he asks, are these procedures followed to the letter every time? And should they be followed exactly as they are written? “In many facilities, the answer to both of these questions is ‘no.’ Frequently, the written form of the procedure is out of date because minor changes to the process over a period of time have not been captured in the formal procedures. Handwritten notes may be scribbled into sections of a ‘working copy’ that are not part of the actual procedure document file, or information is simply missing that some or all of the operators just seem to know but is not written down at all,” he notes.
THE IMPACT OF OPERATORS
Then there are the differences in the way knowledgeable operators actually perform the procedures, either taking appropriate or maybe inappropriate shortcuts, or doing things in what they think is a better way that just simply isn’t written down (Figure 1). Sometimes the actions taken by one person or shift team actually are better than what is written. Therefore, to get the best results when a procedural automation project is underway requires collecting the “best practices” from the best operators, so they can be agreed upon and adopted.
“So collecting the current best practices is the first benefit. Out of this effort, the process may have improved product quality, increased first prime production, fewer disturbance periods, and even safer operations,” Huffman concludes.
For Beebe, a key focus is to capture the knowledge of operating personnel before they retire or move on.
“With state-based control, we can retain this knowledge, not just in word, but in actual form. State-based control can safeguard and shepherd the process operation. It can take some of the load to maintain operational discipline off of the operator. It also enforces good operational discipline of the tasks remaining with the operator.”
Of course, he says, it can’t prevent all bad decisions — but areas of exposure and critical tasks can be performed in a reliable manner and documented in reality not just concept. The steps executed by an operator as documented in a Word document might not match exactly how the operation is performed. In contrast, the steps performed in the state-based platform are exactly how the operation is performed.
In addition, asset utilization improves as plant reliability increases. State-based control is huge for boosting operating equipment effectiveness. Equipment envelopes are maintained. Proper procedures are followed on startup and shutdown, ensuring equipment isn’t unduly stressed or damaged, which could lead to unexpected downtime. Startups and shutdowns are performed in record time. “State-based control doesn’t wait for shift change and doesn’t take a coffee break. When it is ready, it moves to the next step,” notes Beebe.
Nevertheless, implementing a successful state-based control strategy isn’t without its challenges.
“The primary pitfall that has been recognized is that it is unlikely that procedural automation can be created to address every possible abnormal situation and operators need to maintain their operating skills at a high level to be ready to take over when the automation does not have the ‘knowledge embedded’ to react properly,” says ABB’s Huffman.
A couple of approaches are used to deal with this. A training simulator with reasonable process model fidelity is one. Another is to schedule periods of operation where the automated procedures are deactivated and operators run the plant for a period themselves. Finally, there can be the realization that when things really turn bad, it might be worthwhile to simply shut down the process or take it to some safe operating condition. This can be a difficult decision to face — but when procedural automation already has increased production and quality, maybe sacrificing a little production still can be tolerated in some processes, he argues.
Another deterrent to adopting state-based control can be the cost involved. “There can be a noticeable investment required for both resource time and capital investment. But like any other good investment, identifying and understanding the potential benefits can justify the additional upfront expenditure. And companies considering adopting procedural automation need to keep in mind that this can generally be done in a series of small steps rather than trying to do all of the likely procedures for an entire unit all at once,” says Huffman.
PAS’s Habibi also emphasizes the importance of having highly skilled operators: “We have spent the past 20 years researching and developing solutions that address the challenges of capturing operator knowledge and managing abnormal situations and have come to realize the best approach is the back-to-basics approach. We believe that empowering the console operator with the right information at the right time to take action in order to mitigate abnormal situations is the best approach.”
He also highlights cost, noting that state-based control can add as much as 40% to the cost of engineering and configuring a control system. However, many chemical companies already realize that the incremental investment for state-based control pays a return in improved production uptime and increased process safety, he adds.
ProSys’s Beebe points out that most of this additional cost goes on detailed engineering work that should already be done: “If you have the right tools, turning those operational procedures into state-based control is not as cost prohibitive as you would think. In fact, in most of our implementations, operations and process engineering can directly view and troubleshoot the state and transition logic.”
Most state-based-control failures occur when the approach is implemented as a controls migration project. The anatomy of this failure is pretty simple, he says: 1. not having a state-based-control expert on the project early; 2. believing that state-based control is solely an automation task; and 3. trying to embed the state-based control directly into the control system without the proper tools.
Siemens’ Morgan highlights one recent implementation in which a chemical reaction is brought to the production state though a hazardous transient phase/state. The reaction is highly exothermic and required high-temperature trip conditions for the “production” state. However, the “startup” state needed a different high-temperature-trip condition. The alarm management strategy also required the suppression of nuisance alarms on a per-phase basis. Examples included low tank level alarms while filling during the startup state. Similar requirements existed for the “stopping” state.
“From these requirements, certain trip conditions and alarms needed to be bypassed in some states and enabled in others. This can be readily implemented with an indicator for each state,” he notes.
Each state typically is represented and indicated as the output of a latching function block such as a flip-flop or suitable programming methodology. The conditions to enter a state are combined at the “set” input of the flip-flop. When the process transitions to another state, the original state indicator is cleared through the reset input.
“The use of this technology allowed the engineer to automate the startup and shutdown of the system with confidence of uniform response. Without it, the commands to enable/disable bypasses of trips and alarms would rely on operator adherence to documented SOPs [standard operating procedures] during high stress conditions,” he adds.
For its part, ProSys has implemented state-based controls on units including fluidized catalytic crackers, crude units and ethylene crackers, to name but a few.
“In one case, plant startup time was reduced from 36 hours to 12 hours. This not only allowed the process to get an additional 24 hours of production but also allowed the plant to change product specification more quickly to meet market demand. In another case, we enabled a petroleum refinery to respond quickly to an upset condition that traditionally caused a trip and would require a shutdown and restart of the unit, which took 8–12 hours. With state-based control, these upset conditions could be responded to more quickly and process operations could be maintained with only a 15–45-minute impact to process operations,” notes Beebe.
PAS’s Habibi points out that human error still is the predominant cause of incidents: “We have to make the automation and information systems work better with humans by designing them with intended users in mind. The best investment a company can make to improve profitability is through operator enablement. This involves situation awareness technologies such as [a] high performance operator interface, robust alarm management, and [a] reliable decision support system.”
Seán Ottewell is Chemical Processing's Editor at Large. You can email him at firstname.lastname@example.org.