• Not Answered

Tip on how filters can improve tieback model performance

Simple Analog Loop in Mimic

The DeltaV FHX utility is one of the most commonly used tools within Mimic, since it is capable of automatically generating tiebacks, I/O models, and SIOtags corresponding to an existing DeltaV configuration. The generated tiebacks then simulate control loop responses by taking the control system output and scaling it to the process value based on the controller action. This is valuable for simple loop and graphics testing or introductory training to DeltaV.

In my experience, these simple loops may still have undesirable oscillations, and that is because these tieback models are effectively a simplification of a First-Order Plus Dead Time (FOPDT) model. To fix this issue, it is important to realize that the integral action on a PID controller generally corresponds with the process time constant. Those with a controls background know that this time constant represents the speed with which a system responds to change, typically measured by the time taken for a change by a factor of 0.6321. In Mimic, the ‘Filter Time’ on the Loop Tieback or Analog Input block is a first order filter where the user specifies the rise time to 63.21% of the process value meaning that it is effectively the time constant associated with the process value.

This means that the ‘Filter Time’ can be set equal to the PID ‘Reset’ to simulate process time constants (assuming the controls are appropriately tuned) and reduce unwanted oscillations during control system development. Note that this is still what we would refer to as a low fidelity model, which has its limitations when it comes to anything more complex than simple loops (cascade, sequences, batch, etc.).