Insight On Demand Tuning, Error, Level Control


I need to tune a level control loop output from oscillating to big. The level is used to control level in a column and simple feedback loop, normally in AUTO.

The problem is the pv tracks the sp ok, but the output (valve) swings too much which things downstream. I have been tasked with reducing how much the valve opening and close just to control sp.

I do not have full license Insight, but I do have Insight On Demand Tuning. However, when I try a test (with step response of about 2 or 3), I get error such as "process response from step change occurred in the wrong direction. Check the control action (revers or direct) for correct setting, or adjust tuning settings). I believe this mean, my output swings too much to perform the tuning.  (L101LC, Set-point vs Process Variable) (L101LC, Set-point vs Process Variable vs output) 

Is there away to tune the loop so the valve output does not oscillate so much? Is Insight On Demand Tuning giving error because my loop is oscillating too much?

Thank you, 

5 Replies

  • Yes, there is a systematic way to stop the oscillation but still make the loop respond in the time you require. It is called Lambda tuning and is taught in Emerson Educational Services Course 9030, 9032, or 9035. Emerson consultants and our local business partners (including some people available to your plant) have used Lambda tuning to fix thousands of problems like yours.

    Your DeltaV Insight Tune has some capability for Lambda tuning. However, as you asked, there may be a problem that prevents Insight from getting a good process model. It could be a problem in the control valve, or just getting the process to a stable condition before starting the test. This may not be a trivial task--that's why I suggest you gain a little knowledge before proceeding.

    A good article for you to read was published in,
  • Sounds like this might be a surge tank, and rather than control level at setpoint, it should be absorbing swings in the process to stabilize flow down stream. This is a common problem in industry where level control is applied incorrectly. If stabilizing the flow downstream is your main criteria, you should not be attempting to maintain a tight level control. There are several options available on the PID loop to create a gap control or other behavior to meet the requirements of the process as a whole. You should be looking at the retention time of the tank and what its intended purpose in the process is. But you description appears to indicate that stable flow downstream is more important than tight level control.

    Andre Dicaire

  • In reply to Andre Dicaire:

    If it is a surge tank application, I have a couple of posts you may want to see:

    Hopefully these might help.

  • In reply to Jim Cahill:

    In truth, every tank level application is a surge control problem. In the case of distillation, usually both the base level and reflux accumulator level need to be held within fairly narrow limits. Lambda tuning, as Mark Coughran noted above and as is mentioned in your links, provides a nice first order response to a load disturbance. Lambda, the closed loop time constant, should be selected as large as possible in most level control applications, but short enough that the level will remain within allowable limits in the face of the largest possible load disturbance. In the overhead, this becomes particularly important if the level is controlled by manipulating the reflux flow. This arrangement isn't that common, but becomes necessary in distillations with a high reflux/distillate ratio. The bottom product can usually be fairly loose unless the bottom flow is feeding a downstream unit or is being used to preheat some other stream. Then it should be as stable as possible. The important thing is to tune for gradual non-oscillatory flow.
  • In reply to Mark Coughran:

    Thank you so much for responding Mark.

    Apologize for the late response. I definitely signing up for the Emerson Educational Service course. Recently, we hired a tuning expert from Emerson who used EnTech toolkit to perform tuning on some complex loops. He did a great job.

    I was able to use Insight On Demand Tuning to tune the loop above i had issue with.

    I was not selecting Integrating Process (for level application) and I did not put the Loop in manual before doing the bump test. I was trying to tune the loop while it was in AUTO and oscillating. So when I give bump test I get error stating the loop is not behaving in right direction. The bump test I use was 3, and it took about 15 minutes to get tuning recommendations. Then I use Simulate tab for further suggestions. I was able to tune the loop from these suggestions and I reading about Lambda Tuning for clear understanding of the methods.

    What is the difference between Insight Adaptive Tuning Software, Insight On-Demand Tuning, and Entech? Currently, we do not have Full Insight License, but only access to On-Demand Tuning and Simulate tabs.

    Thank you