Key points
Abstract
Process industries have gone through a long journey from fully manual mechanical control systems to fully automated digitalised processes. Companies spend millions of dollars on Advanced Process Controls (APC) and ultra-modern equipment but may still suffer from heavy operating losses. APC systems are only as good as their weakest link, which is often in underlying control loops.
Loop tuning and monitoring is a challenge, and it is traditionally done manually by experts and often too late when problems already exist. Loop tuning and monitoring software packages have made this easy and potential problems can be detected before they cause harm. This enables business performance improvement by millions of dollars.
We have gone a long way.
Roughly 100 years ago my great grandfather came home from work very unhappy. He had just lost the new work overalls that his wife had made for him. His hands and back were full of scratches and bruises. What happened? Just a week ago he got a promotion in the paper mill. From lawnmower to press boy.
His duty was to stand next to a paper machine press section and watch the position of press fabric. If it came too close to the paper machine frame, he had to turn adjustment wheel clockwise. If the fabric went too far from frame, he had to turn the wheel counterclockwise. Simple, but that was quite a tough job (Pic 1.).
Press fabric was in constant movement, and he was not able to leave his position for even 10 seconds. If the fabric went to the frame, it destroyed the fabric and machine had to be stopped for hours. Quite high pressure for young papermakers. Long days and a lot of responsibility can make a person tired and prone to mistakes.
One evening he lost his concentration and got his overall sleeve squeezed between press rolls. In a fraction of a second the whole overall was pulled in the machine. Luckily, he was strong enough to keep himself from being pulled into the machine with his overalls. If not, I wouldn´t be here, writing this.
In that time, almost all controls were manual, required a lot of manpower and were indeed very dangerous sometimes. Controls were also slow and unreliable, fully dependent on the operator. They were also environmentally unsustainable manual controls. Dangerous chemicals end up in the river when a tired operator fell asleep or forgot to close the valve.
Times change, challenges remain.
Now, 100 years later, we have a very different world.
Operators are sitting in air-conditioned control rooms, watching all process areas from big flat screens. Production processes have become highly efficient, fully automated systems which can be manipulated with mouse and keyboard, requiring only a fraction of the personnel compared to old times.
However, less personnel mean higher responsibility to individual operators. Understanding of bigger process areas is required and potential mistakes can be very expensive. At the same time, stockholders require maximised efficiency, and 100% output is expected to be delivered by the production lines.
This is a challenge as there are several subprocess areas expected to work together seamlessly, while each of them has their own sub-processes, including numerous control loops. Checking the performance of those manually is close to impossible without computing power.
It is not a surprise that everyone is talking about IIoT, Industry 4.0, and Digital Transformation nowadays. Advanced process controls can save millions of euros or reduce thousands of tons of CO2 emissions per year.
That´s why plants invest millions of dollars in sensors, actuators, control systems and advanced process control solutions. However, these investments need to be properly cared for, otherwise they will fail. Picture 2 shows what unfortunately is a typical life cycle of an advanced control system. Great value gained in the beginning will slowly diminish until the system is completely shut off.
How does this happen?
Digging into details
Quite often there is resistance from operators to trust new systems, so they tend to switch it back to manual mode. This can be avoided or at least minimised with proper training. A more critical reason is running in the background, invisible for the operators: Control loops and measurements.
These are the foundations for all higher-level control systems. These foundations need to be rock-solid to enable sustainable gains in business performance (picture 3). It would be guaranteed failure to implement advanced process controls without taking care of the basics first.
And it is not only advanced process controls which rely on reliable basics. Any production lines of any industry need them to be able to produce constant quality in an efficient way.
Here is a scary fact: Did you know that on average, about 70 % of control loops are not healthy. This is valid for all industries. Being not healthy means that loops are on manual mode, are not following their setpoints, or does not have enough control capacity. In other words, they are not doing their job. After surveying hundreds of production lines, we can conclude following:
- 30% of the process controllers installed operate in manual mode.
- More than 30% of the loops increase variability over manual control due to poor tuning.
- 30% percent of the loops have equipment problems.
What does this mean? How much financial damage these unhealthy loops can cause together with APC, or in another way round, how much can you gain by making your basics right?
“A pulp-and-paper mill in north America was suffering heavy operating losses despite investing in ultramodern equipment and APCs. After implementing an improvement program, including operator training, control loop improvements and analytical models, they gained $21 million, 15% output and 4% yield” (McKinsey 2020)
Why don´t we take care of our most important assets then?
Control loops might be located on 7 different floors, 1 km apart, hot, moist, and noisy places. Sometimes in places which can only be reached by cherry picker. In other words: It takes a huge effort to maintain them.
There are often not enough resources, and additionally, unclear responsibilities, poor maintenance planning and budget constraints will not make it any easier. The loops control process parameters such as flow, speed, concentration, and temperature. Let´s have a look at single feedback control loop (pic 4):
Sensor measures the actual process variable, such as flow rate, temperature, pressure, or level. It provides real-time information about the current state of the system.
Controller processes the information from the sensor and decides the necessary control action. It compares the actual value to the desired set point and calculates the difference between present value and setpoint. It gives a signal to the actuator to adjust the process.
Actuators produce a motion by converting energy and signals going intocontrol elements like valves.
Process response: Is the actuator controlling a valve of liquid or gas, what type of valve, what system volume does it apply to, just few to mention. Example outside industry: If you press a gas pedal of truck or rally car, the response is totally different.
Failure of any of the boxes in the diagram above will lead to failure of the whole control loop.
Drilling deeper into details
Typical challenges of control loop elements are described below.
Stiction (picture 5) is extra force needed to move a valve from dead-stop and is a sum of the static friction in the moving valve parts. It is affected by packing gland torque, process fluid viscosity and valve plug properties.
Hysteresis means that you have a different valve position with the same signal depending on valve being opened or closed. Backlash is a part of hysteresis caused by lost motion on valve or positioner mechanical parts. Most commonly these come from loose or worn mechanical parts.
Incorrect valve or pipe sizing is purely an engineering fault where too small valve or pipe cannot pass through the required flow or too big valve makes the loop overly sensitive to small valve changes. This happens often when processes are modernised and the production rate changes. As a rule of thumb, valve should run 60% – 80% open at maximum required flow.
Poor control loop tuning
Controllers need to be tuned according to process requirements, conditions, and equipment. Too fast or aggressive control action does not help the process in any way, it only causes equipment wearing and potentially oscillation of the signal. Too slow control or completely manual mode passes process variation through.
Poor physical design of control loop
In addition to the abovementioned valve sizing, general design of the loop is particularly important. If the sensing element is placed in the wrong place and the control device installed improperly, it makes controlling exceedingly difficult if not impossible. In the illustration below the sampler and sensor would show different concentration if the setup were incorrect. That makes proper calibration impossible.
Faulty measurements
If you cannot measure it, you cannot control it. Online process measurements are often indirect, where the actual sensing element measures a different property than the final signal shown. That´s why measured values need to be converted to final signal by sensor calibration which needs to be repeated periodically. For proper calibration, correct process information and samples are crucial. As shown in picture 7, an incorrect sampling point would lead to incorrect calibration.
Having up to 1000 control loops, spread over 1 square km area, in 7 buildings and knowing all the potential issues described above, it is not a surprise that only 30% of control loops are healthy. This goes together with the fact that in some cases only 10% of installed APC systems are working as expected. However, it does not have to be like that.
How to get there? How to increase the percentage of healthy control loops and how to maintain this improved level? How to sustain the best performance of APC systems?
Fixing it and start being profitable
Dedicated software solves the abovementioned problems by taking care of two crucial tasks:
Loop tuning
The first step on improvement journey is to identify all the important control loops in the system and tune them. Loop tuning software takes in data from the historian and uses model predictive algorithms to simulate control loop behaviour. Picture 8 shows the user interface of sophisticated loop tuning software.
The user can easily see if a loop model is satisfactory or if there is additional information needed. The user can add some more parameters like properties of the media controlled in the loop. For example, gases have completely different behaviour compared to incompressible liquids.
Additional bump tests may be used to identify disturbances from the outside system, but in most cases PID constants calculation can be done in seconds by pressing a button. Trial and error are not needed in most cases.
Using this kind of tool will enable the tuning process to be standard across organisations, preventing the use of different techniques or approaches based on individual engineers potentially leading to hidden problems. Picture 8 shows an example of a model building procedure.
Loop condition monitoring
Tuning the loops is only halfway to success. Nobody can check hundreds of loops manually in daily business. Tuned loop may work like a charm for about one or two months, then too many variables change in the process, the valves are not lubricated properly, the air pressure in the pneumatic system changes, production increases or decreases… in short, several variables affect the PID constants used.
Once these parameters are deteriorating loop performance, the operators complain that the control valve is not working (opening fast enough, taking too long to respond, valve is getting stock, valve is oscillating, and some other problems), they put the valves in manual and this is a really bad situation for the plants, having the operators control the valves manually is inefficient and dangerous.
Loop performance monitoring software does scheduled checks of each loop and shows the results visually in the report. Site performance charts are organised by process area where single loops are shown as boxes.
The bigger the box, the more important the loop. User-defined colours indicate the condition of the loop (picture 9). Details of the loop condition and trend of related tags can be opened by clicking the box. With cloud integration, it is possible to monitor an entire enterprise globally.
If there is an important loop which needs immediate action, emails and SMS messages can be set to be sent to key personnel. This prevents unplanned shutdowns efficiently.
Being sustainable
Deteriorating benefits were mentioned earlier in this article. By taking care of your valuable control loops, you can turn this phenomenon around by not just sustaining the benefits, but continuously improving your process. This is possible only when you can trust your numbers and processes. Which graph describes your process best?
Conclusions
In today’s industrial processes, automatic controls are required to keep the system stable. But different process areas often interfere with each other, and it is challenging for operators to have a holistic understanding of all involved process areas and interactions.
Advanced process controls can significantly improve business performance by assisting operators to make quick adjustments based on relevant process conditions. This reduces process variation and helps to produce a more stable quality with an optimised cost structure.
However, advanced process controls still rely on the underlying control loops, so it is of utmost importance to keep these loops in good shape. Ensuring good loop performance continuously is difficult and requires a lot of resources.
Luckily, there are good software tools available nowadays for tuning control loops and monitoring their performance automatically. This helps to keep advanced controls running and ensures that the savings potential of millions of Euros continues to be delivered.
