Control loops are the first line of defense for safety and process stability, and act as an engine for profitability. With hundreds or thousands of loops in operation in a single facility, it is complex to monitor, analyze, assess and diagnose potential issues. Over time, basic control loop performance can deteriorate significantly, leading to sub‐optimal operations and even unplanned shutdowns. Continuous monitoring of control loops quickly identifies anomalies allowing the operator to address potential issues before they affect production.
There are risks associated with these control loops. Being the first line of defense, our loops protect us against excursions in our product quality, rates, efficiency, and environmental performance. If they are not working well, they can bring our process to the undesirable point that the safety systems activate. We want our loops to work. We want our process to draw nice, straight lines on those trends. We want the control room to be a very boring place where things run optimally.
When loops work, we get that boring control room. When they don't work, we need lots of intervention by the operator. We get minor upsets that can escalate into major upsets. We get safety system activations and shutdowns, which is certainly a safe state of the process; but, we all know the most hazardous and risky stage our plant is usually the startup. That is when bad accidents happen. To avoid unplanned startups, we have basic objectives for our control loops. We want our process to run optimally, and we want the control system to have good disturbance rejection so that when unplanned things happen, it can compensate for those without a whole lot of operator intervention. All of this has to do with the performance and the configuration of the loops.
We have given our operators a lot of stuff with these modern control systems. Very large spans of control are typical, with more than a couple thousand I/O points or sensors, and they are generally responsible for two to 500 controllers. This is a high cognitive load, and the span of control of most operators is so large that the loops must be running well.
They cannot run well if all two to 500 of those controllers are run manually. It simply wouldn't work. The process can only be effective if the operator is effective. They can't handle the load if any substantial fraction of loops is not working.
We are often surprised when we investigate just how ineffective our control loops are. There are some very simple questions that can be difficult for plant engineers and managers to answer. For example, how many of your loops right now are running in manual? That's a pretty simple question, but how long would it take for you to get an accurate answer?
We spend millions and millions of dollars on very sophisticated equipment and process automation technologies. It turns out that a whole lot of that is simply wasted. Why?
The problem that arises is that when loops are running well, and process variation is small—oscillations are small, variations are small—we're able to run near our limits. Meaning the limits where our alarms occur, where it's safe to run, where it's optimal, where the best rates are, and where the most efficiencies are. However, if we don't take care of that as the process variability increases, we need to run the process further away from those desirable limits. In doing that, we have less efficiency, less productivity, potentially more variation in our quality, and a whole host of problems.
Unfortunately, that is how many plants are run today. There have been studies about control loop performance in our actual plants, and the number of problems is surprising. More than 30% of controllers are typically run in manual, which they were never designed to do and it is generally not desirable. But, operators do so because they need to accomplish their goals of running the process correctly and efficiently.
More than 30% of loops increase variability when they're in auto versus manual and that was never the intent. About 15% of loops have design problems, equipment problems, poor tuning, incorrect valve sizes, and valve performance – hysteresis, and stiction being two of the major problems that can be detected and resolved.
The funny thing is, we put up with this. Projects were funded to install these loops, and no one said when those projects were funded, “by the way, it's OK if we start up and 30% of them don't really work. That's quite all right.” We need to have a “no excuses” approach to this issue.
To hear steps on how to improve control loop performance, watch the on-demand webinar.
You will learn how to:
- Enable continuous control loop monitoring
- Pinpoint root causes of issues
- Optimize control loop performance
This blog is an excerpt from the PAS On-Demand Webinar, Operations Risk Management: Are You In Control of Your Control Loops?