Timing Is Everything, Especially When It Comes to Control Loop Tuning and the Process Time Constant
Party hosts often cringe when guests arrive early or late. Seemingly subtle shifts in timing can throw off planned details that are of importance to the host and utterly lost on guests. So too the timing of a PID control loop’s response is important. Premature or delayed responses negatively impact the controller’s performance. Previously the calculation of Process Gain – the “how far” variable – was covered. The Process Time Constant is equally important to process modeling and PID controller tuning. As with most things, timing is everything.
Time Constant is the “how fast” variable. It describes the speed with which the measured Process Variable (PV) responds to changes in the Controller Output (CO). More specifically it represents the time needed for the PV to reach 63.2% of its total and final change. Like other model parameters the Process Time Constant can be calculated through manual, graphic analysis of step test data.
Practitioners who choose to tune manually should be mindful of certain aspects of the Process Time Constant, including the following:
- Which is Which?
Process Time Constant is a model parameter whereas either Reset Time or Reset Rate would be the corresponding tuning parameter. The former describes important aspects of a given process’ dynamic behavior. The later serves as the clock of the controller and contributes to the PID controller’s responsiveness to disturbances. The two are connected, but they’re different all the same.
- Stay Positive
The value for this model parameter should always be positive. To calculate the Process Time Constant begin by determining the time at which the PV completes 63.2% of the total change, then subtract the time at which the PV clearly responds to the step test. The math involved can only result in a positive value. A negative value should raise a red flag, indicating that something went wrong with the calculation.
- Time Matters
When calculating the Process Time Constant it’s essential that the corresponding units are accurately recorded – whether they’re in seconds, minutes or some other increment. The designated units set the clock by which the PID controller will respond to change. Getting this detail wrong simply means the controller will arrive early or late to each upset.
A good portion of practitioners choose to tune PID controllers manually by modeling a process’ dynamics and – more often than not – using straight-forward calculations like the one explained above. This approach is well suited for control loops that are steady and that showcase only limited noise. However, manual tuning can be less than effective when the process’ dynamics are oscillatory and noisy as the associated step test’s start and finish are typically unclear. Controller tuning software capable of modeling oscillatory and noisy process conditions offer a viable alternative. Just be aware that some software tools claiming to model non-steady-state data are actually limited to use with integrating processes. Those represent only a small portion of control loops used in industry.