Who Should Use Feed Forward Control? What applications are best for Feed Forward?

ff-blockdia-big It’s not uncommon for process control discussions to focus on use of the PID to improve a loop’s ability to track Set Point.  Oftentimes tweaks to controller tuning parameters either seek incremental reduction in Settling Time or strive to limit Overshoot.  Indeed, many PID tuning courses position optimization of a control loop’s Set Point Tracking as the ultimate goal.  Such a bias overlooks the value of other strategies like Feed Forward Control.

While Feed Forward Control is not an appropriate control strategy for improving Set Point tracking, it’s been widely shown to help improve a process’ capacity for disturbance rejection.  If a disturbance to a process can be isolated then it can often be modeled and defused using the Feed Forward architecture.  Unlike traditional feedback control Feed Forward utilizes a model to proactively trigger an appropriate corrective action.  That action anticipates the end state of a disturbance rather than reacting to Error produced by the disturbance as it accumulates.  By getting a jump on the disturbance’s impact Feed Forward Control prepares the process to mute any ill effects.

Not every application is ideal for Feed Forward Control.  Even so most processes can benefit to some degree from its use.  Following are process attributes for which Feed Forward is well suited and for which the architecture can often improve disturbance rejection:

Load Changes

Manufacturing processes that are subject to frequent load changes are excellent candidates for Feed Forward.  Think of Food & Beverage manufacturers along with Utilities – especially power companies.  The load changes experienced by these and similar process manufacturers are largely predictable, readily modeled, and they can be neutralized through the application of appropriate counter-measures.

Slow Processes

Processes that can be characterized as slow due to a large Dead-Time are often good candidates for Feed Forward Control.  Manufacturers in the Oil & Gas and Basic Materials have many such processes.  The delayed responses of their processes offer a window of opportunity for the type of preemptive correction achieved through Feed Forward.  The counter-measure can be timed for implementation just as the disturbance impacts the process.

Big Equipment

For a variety of reasons large process infrastructure tends to be slow, important, and very expensive.  Multi-zone furnaces, steam turbines, and distillation columns are good examples and span a variety of industry segments from Chemicals and Oil & Gas to Basic Materials and Pulp & Paper.  The proactive control achieved through Feed Forward protects costly equipment that is also big and bulky.

These characteristics are shared by processes which are often good candidates for a disturbance rejection strategy such as Feed Forward.  While Feed Forward does not impact Set Point Tracking, it does serve an often equally important purpose – rejecting the negative effects of process disturbances.

These resources offer related content:

Output Distribution

What Does Controller Output Reveal About FCE Health and Asset Reliability?

Remember the Far Side?  Of course you do, but do you remember the one depicting the boy who leans his whole body forward as he pushes against a door that clearly reads “Pull”?  Although the sign leading up to the building shows “Midvale School for...
Iconic-Mark_Inverted-Color

Using Mode Changes to Uncover Tuning Issues

  Mode Changes Provide Insight Into a Process' Overall Performance and They're Often Linked to Poor Controller Tuning. Operator interventions and Mode Changes are often viewed negatively. When a PID loop is switched out of automatic there can be a sense that the controller wasn’t...
shutterstock_207048508

How Can I Correct for Noise Using Internal Filters?

  Internal Filters are an Effective Solution for Addressing Noise.  Just Don't be Fooled ...They Don't Solve Everything! Noise is inevitable. To one degree or another it’s evident in the data of most every production process. Sure it can be absent in academic settings and...

Still looking for more?

Now that you’ve gotten the basics, connect with our team to learn how our people, processes and technologies can help you optimize.