Control Loop Case History 100

THE 100^TH ARTICLE IN THIS CASE HISTORY SERIES

I have been writing these Case Histories for many years, and feel I have reached quite a mile-stone by writing this hundredth in the series.  All are based on my experiences in optimising regulatory control loops (the “base control layer”) in several hundred plants in quite a few countries around the world. I have also been giving courses on the practical aspects of optimising regulatory control loops since 1990, and several thousand people have attended these courses over that time.  Some of these people were senior C&I practitioners, some with over 30 years experience in the field.  Well over 99% of these attendees came out of the courses saying they were the best courses on control they have ever attended.  Typical remarks were “mind-blowing”; “I wish I had attended this course 20 years ago”; “It’s the first time I have ever really understood feedback control”;  “I wish they taught control like this at University”

I feel it an appropriate point to try and summarise what I have found:

1. Practically all the delegates who have attended my courses, and nearly all the people that I have worked with in plants had previously very little real understanding of the practical aspects of control.  A few, mostly with long experience in the control field, have said later that they initially resented me, or coming on my courses, as they knew I would not be able to teach them anything.  Many of them, including some very senior control managers in very big plants have later apologised to me for this, and admitted they had no idea of their lack of knowledge.

2. I have also never met anyone who prior to the course really understood the features, operation, and workings of the PID control block built into his or her control system.

3. About 80% of control systems using PID controllers in PLC’s are set up incorrectly.

4. The manufacturers of the controllers are not interested in educating their clients in their PID block operation, and in fact most of their staff including trainers and writers of the technical manuals don’t really understand very much of practical control.

5. As I have stated so many times in the past, at least 85% of all control loops in plants world-wide are operating very inefficiently in automatic, and only a few work reasonably well.  This is entirely due to the lack of understanding of practical control which is not taught elsewhere, and to the complete lack of understanding of the potential huge increase in financial gains if controls work properly.

6. Very, very few plants regard optimisation as a separate discipline.  It is normally only applied if an Operator really starts complaining, or if a Manager realises that a particular control is critical to plant performance.  In general most plant personnel including senior C&I, and process management have a completely apathetic approach to improving base layer control performance.

7. Optimisation will only succeed if it becomes endorsed plant wide by all disciplines, and if really actively supported and promoted by senior management.

8. Optimisation is a team effort.  Process people, including process experts, and operating staff must work as a team with C&I people of there is any chance for it to succeed.

9. Optimisation needs to be carried out by skilled, well trained, experienced, and intelligent people.  Many plants believe completely erroneously, that it can be performed by lower level disciplines.

Why am I putting all this down?  It is not for the purpose of blowing my own trumpet, but loop optimisation is my passion.  I used to own a company supplying control systems, and made far more money from that than I can do on my own teaching and performing optimisation.  However years ago due to the help of a special friend in America who taught me the basics of this subject, I fell in love with it, and decided to devote my life to it.

However it is very often extremely frustrating with the general apathy and sometimes negative attitudes of many companies and personnel.  At times I feel like a crusader trying to push the gospel to the unbelieving.  However what makes it all worthwhile and very rewarding are the many successes achieved by both people who have embraced the teachings and apply them well in their work, and also in the practical consulting work I do which in most cases result in huge improvements in process and plant performance.

I have an example of this right now.  I have just spent a week in a process plant that has embraced optimisation.   The control manager is committed to improve plant performance, and has appointed a team of highly motivated and very enthusiastic control specialists.  It is a fairly old plant, and we examined some very interesting loops.  The following two examples are taken from  those.

The first example is of an important flow control which the operators had always been complaining about.  The “as-found” closed loop test (not shown here) was rather strange.  It appeared that the response to setpoint changes were initially extremely slow and then moved quickly at times, and then slowed down again.  It was also apparent that the valve was sticking badly on reversals.  It was quite a puzzling response.  

Fig. 1

The open loop test (see Figure 1) revealed some interesting things.   Basically when the PD (controller output) was below 20% the PV (flow signal) was at its lowest value of about 11%.  (This could indicate that one or more of the following problems could exist:

• >Valve is incorrectly calibrated

• Or the transmitter is incorrectly calibrated.

• Or there is a valve bypass open.

• Or that the valve is passing when shut.

On opening the PD by a further 10% step the flow PV also moved up by 10%.  However another 10% step in PD only resulted in the PV moving in a tiny step of about 1.5%.  Further steps in the PD resulted in very little further increase of PV.  It was completely saturated by the time the PD had reached 50%.  The test also reveals at least 5% hysteresis on the valve.  (This is not marked in the Figure but can be determined by the offset in the PV on valve reversals taking process gain into account – see Loop Signature series on hysteresis for further information). 

Some serious investigations into the transmitter and valve setups, and into the process itself, are obviously needed to find out why the flow only can be moved over a tiny span of 12% .  In the meantime the control is extremely limited.   If the flow can really only operate over such a small range then the transmitter spanning needs to be reduced.

A temporary measure that was put into place in the meantime was to put high and low limits on the controller’s output to prevent it moving into regions where it can no longer influence the control.  This prevents both mechanical and reset windups, and is very good practice.  There was however tremendous opposition to this from the operators.  They would not believe that more flow could not be obtained by opening the valve past 50%.  They had the feeling that even if the flow was limited then opening the valve further would somehow cause an increase in flow.  It is sometimes very difficult to overcome perceptions of people not trained in loop optimisation.

The second example in this article is to show  how dramatically good tuning can improve performance.  I have often stressed that tuning is the last thing one should do when optimising.  One must understand the control system, analyse it and correct problems before tuning, and most of my article have been devoted to those aspects of optimisation.  However there is of course no doubt that good tuning is also terribly important for optimum loop performance.

This example is a very important and critical temperature control loop in the plant which has been giving quite a lot of problems.  Although the process is subject to frequent load changes, the operators could not run it in automatic as the control was so poor.

 The control personnel in the plant had the valve replaced as they felt that the fault could lie there.  However automatic control was still very bad, and the loop was kept in manual.  When we examined the loop we found that the existing tuning was really way out.  

Fig. 2

Figure 2 shows the response of the process to a setpoint change with the original tuning. It took some 4 hours for the controller to move the temperature up 5%.   

Fig. 3

The tests on this loop took just over an hour to perform.  After satisfying ourselves that the valve was in fact performing well, an open loop test was performed to ascertain the dynamics of the process, and to do the tuning.  Figure 3 shows the result.  The process now takes only 10 minutes to move up the 5% to the new setpoint.  It is certainly one of the most dramatic improvements in loop performance purely due to better tuning that I have seen, and I thought that it was worth while sharing it with our readers and to finish off the hundredth Case History article.

Index to articles

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:
Tel (011) 486-0567
Fax (011) 646-2385
E-Mail:  michael.brown@mweb.co.za