Editor's Choice


Control loop: Case History 172 - Interesting controls in a copper extraction plant.

June 2020 Editor's Choice

In my 30 years devoted to optimising controls in industrial process plants in many countries, I thought that I had seen all the possible process dynamics that one would encounter. Imagine my surprise when I came across one I have never encountered before. This was in a temperature control on a copper smelter plant, where I recently presented a course on practical control to process control engineers and also performed some optimisation in a couple of their plants.

Temperature control optimisation is usually extremely onerous, mainly due to the slowness of the process, and requires a great deal of patience whilst waiting, sometimes hours, for a process to respond to a step in the controller output (PD) or a change in setpoint (SP). This loop was no different and what made it even more confounding was the fact that the PLC control block had not been set up correctly insomuch that the PD tracking had not been set, so that when the controller was placed in automatic from manual, the controller’s PD should carry on at the same value it had when in manual. Instead it dropped to zero and severely bumped the process. This is incredibly frustrating because it is absolutely necessary to make bump-free changes from automatic to manual and back again when performing optimisation.

Unfortunately, no one capable of sorting out the PLC program was present over the weekend on which the work was done, so we had to improvise.

The loop in question was top of the operators’ problem list. It was in a terrible continuous slow cycle with the PD being either fully open or completely shut and the PV slowly going above or below the setpoint. The operators had unsuccessfully tried to control the loop in manual, but the problem was that large load changes took place every 20 or 30 minutes when a converter was turned out of stack and emptied. Therefore they had had to leave the loop in automatic, but it was really badly upsetting production. Incidentally, it had been like this ever since the plant was built and commissioned eight years previously.

Strange response in the PV

The process was the temperature control of a cooling heat exchanger. Now heat exchangers normally always have self-regulating dynamics. On starting our tests, which mainly had to be performed in manual, we saw that after an hour the PV showed no sign of settling out as normal with a self-regulating process, but continued ramping either up or down depending on the magnitude of the PD. I then managed to get hold of a senior process engineer who explained that the process fluid was actually recycled back to the converter’s hood after passing through the exchanger. Therefore after the pass through the hoods, colder water was coming back to the exchanger, cooling it even further. Therefore the process is integrating.

Figure 1.

The next thing that I noticed is that every time a step change was made on the PD, the PV responded in a strange way. It firstly slowly curved into the ramp. This normally means that a large lag is present in the process, which is quite normal in temperature processes. However after a while the slope of the ramp curved back into a different slope, which is a very strange response that I had not seen before. This can be seen in Figure 1.

Second order lag

After thinking about it I realised that it was a second lag coming back into the PV as the recycled cooling water started coming back into the exchanger. Therefore the dynamics of the process are a slow integrator with a second order lag. This is really unusual dynamics. I have encountered second order lags occasionally on self-regulating processes, which is usually a sign of loop interaction, but never on integrating processes.

Fortunately for me, the Protuner tunes from a frequency response that it gets by performing a digital Laplace transform over a step change in manual. I strongly believe that no model based tuning system would be able to tune this process, as it would be almost impossible to make a sufficiently accurate model.

The difficulty here was that we couldn’t get the process into balance due to the PLC problem. Therefore we put the controller into automatic and changed the tuning to P only with a proportional gain of 8. I should mention that the original unstable tuning had P, I and D. Also the derivative filter had been set completely incorrectly – typical trial and error ‘desperation’ tuning.

Figure 2.

A P only controller is the equivalent of a ball valve on a level control, and automatically finds the balance point. In this case, as established later, the integral retention time was 1 hour and 10 minutes which is very slow, so it took a long time to try and get the process to near balance, particularly with a few small load changes that were occurring. We got it reasonably close, but unfortunately the operators had to make changes in the process that would have caused a dramatic bump, so we did a step quickly even though we were not exactly at balance. This step can be seen in Figure 2.

Conclusion

We then performed a graphic edit on this PV response to try and get the response to what it may have been if we had actually reached balance. Tuning on this step and then selecting a fairly conservative P+I tune, we first tried it out on the Protuner’s simulator with the process modelled from the frequency plots. It seemed to work well. Also trying the same model with the original tuning exhibited the same instability as originally seen in the loop. Therefore with a great deal of confidence we put this tune into the actual controller and put it into automatic which of course caused the usual huge bump, as the PD went to zero. It then took only about 20 minutes for the loop to get close to setpoint and settle out. It was slightly below setpoint due to the conservative tune, which had an extremely long integral. However within about 30 minutes it reached setpoint and stayed there.

Figure 3.

We ran the control overnight and saw the next day that it had performed extremely well with the load changes. The process manager was really happy to say the least, and said he couldn’t have believed it was possible to have tuned the loop so well.

Figure 3 is a screenshot taken from the actual control system’s trends package which shows the performance before and after the optimisation exercise. (The green line is the PD (controller’s output) – all in all, another very successful and pleasurable day at the ‘coal face’.

About Michael Brown


Michael Brown

Michael Brown is a specialist in control loop optimisation with many years of experience in process control instrumentation. His main activities are consulting, and teaching practical control loop analysis and optimisation. He gives training courses which can be held in clients’ plants, where students can have the added benefit of practising on live loops. His work takes him to plants all over South Africa and also to other countries. He can be contacted at Michael Brown Control Engineering.


Credit(s)



Share this article:
Share via emailShare via LinkedInPrint this page

Further reading:

The thermal combustion balancing act
Editor's Choice
From carbon taxes to export tariffs, and cost containment to security of supply and sustainability, companies are under increasing pressure to switch to greener fuel sources. Associated Energy Services warns that this pivotal change has some potentially serious knock-on effects.

Read more...
What’s driving the IE3 motor revolution?
WEG Africa Editor's Choice
The International Efficiency 3 (IE3) motor standard will soon become South Africa’s legal minimum standard, mandating that local suppliers offer more efficient electric motors. What is driving this change, and how does it affect the many industries that rely on these modern electric workhorses?

Read more...
Unlocking the smart factory
ElectroMechanica Editor's Choice Motion Control & Drives
At ElectroMechanica, we recognise that transitioning to smart automation isn’t just about adopting new technology; it’s about solving real challenges. Labour shortages, rising costs and downtime due to outdated machinery make digital transformation essential for long-term competitiveness.

Read more...
Case History 197: Bad reboiler temperature control.
Michael Brown Control Engineering Editor's Choice Flow Measurement & Control
It is very important that reboiler temperature controls operate well in petrochemical refineries, or the product quality can really suffer. I was asked to check such a control in a refinery where they were having problems with one of these controls.

Read more...
The future of industrial automation: fieldbus and industrial networking
LAPP Southern Africa Editor's Choice
As a global leader in integrated solutions in the field of cable and connection technology, LAPP recognises that fieldbus and industrial networking technologies are pivotal in shaping the future of manufacturing and production processes.

Read more...
AI-driven innovations with CCTV and cyber security
RJ Connect Editor's Choice Fieldbus & Industrial Networking
The fast progress of artificial intelligence (AI) and video analytics is redefining the rail surveillance landscape. Advancements have bolstered proactive event detection, predictive maintenance and enhanced situational awareness.

Read more...
Loop signature 27: SWAG tuning of simple integrating processes.
Editor's Choice
The chief control engineer of one of the largest petrochemical refineries in South Africa once sent me an email after a course at his plant. He wrote that he had found the section on SWAG tuning of simple integrating processes one of the most informative of the whole course.

Read more...
Harnessing industrial AI agents for reliable automation
Editor's Choice IT in Manufacturing
The excitement around generative AI (GenAI) has been undeniable, promising wide-ranging changes across industries. However, for those of us in the world of industrial control and automation, the realities of implementing these powerful technologies are a little more nuanced.

Read more...
Futureproof your industrial network security with OT-centric cyber security
RJ Connect Editor's Choice
To achieve digital transformation, industrial operators must first address the daunting task of merging their information technology (IT) and operational technology (OT) infrastructure. In this article, we focus on the importance of strong OT network security and provide some tips on how to strengthen cybersecurity for industrial operations.

Read more...
The symbiotic relationship between OEMs and SIs
Schneider Electric South Africa Editor's Choice System Integration & Control Systems Design
While businesses tend to turn directly to original equipment manufacturers OEMs or vendors when embarking on IT projects, the role of the SI as a key facilitator and partner cannot be overstated.

Read more...