How Do I Perform an Effective Bump Test?

Best-Practices for Bumping a Process and Tuning Your PID Controller

Bumping a process is a standard requirement when tuning a PID controller – whether you tune manually or with the help of software.  And, either way, applying industry best-practices will save you significant time and money.

The bump test is critical as it provides the basis for understanding the cause-and-effect relationship between a control loop’s Process Variable (PV) and Controller Output (CO).  It’s where a change in CO is introduced in order to see how the PV responds.  Once that relationship is established model parameters can be calculated and converted into tuning parameters for use with the loop’s PID controller.  That may sound easy enough, but there are several things to consider:

• Keep It Quiet – Disturbances garble the process data.  When analyzing the results of your bump test it will be next to impossible to distinguish what should be attributed to your bump versus the disturbance.  To whatever extent possible, keep the process steady and quiet.

• Mind Your DLO – As shared previously the Design Level of Operation (DLO) is where the control loop spends most of its time.  It’s a range that corresponds with normal operation and it should be considered when performing your test.  Testing outside the DLO will only hamper the PID’s ability to maintain control that is both effective and efficient.

• Stay Sharp – Whether you chose to perform a simple step, full bump, doublet, or even a PRBS, the change to your CO should be distinct.  Slowly ramping the CO from one value to another may make it difficult to clearly see how the change drives the PV’s reaction.  Additionally, it’s critical that the change overcomes any deadband or stiction in the process (i.e. valve or damper).

• Mind the Noise – Most every process has some degree of noise.  Be sure to bump the process outside of the noise band so that the PV’s response can be seen clearly.  A general rule of thumb is to perform a bump test that is at least 4-times the value of the associated noise.

• Let It Show – The purpose of the bump test is to understand the CO-PV relationship, so perform your test accordingly and let those dynamics show.  It is not always necessary to let the PV steady at a new value, but a clearer response will make for a more accurate model of the process and for more optimal control.

These simple guidelines can help you avoid bump tests that will produce poor results and/or require you to conduct additional tests.  It may go without saying, but always keep your safety and that of your colleagues in mind when performing a bump test.

If you need additional guidance on how to perform a test or how to tune your PIDs for optimal performance, consider attending a training workshop that focuses on understanding process dynamics and the PID controller or contact us.