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 similar lab environments where simulations often generate sanitized data. However, in the real world of industrial manufacturing noise is a factor that cannot be avoided. Failing to account for or manage noise can be a recipe for – well – failure.

Noise is defined as random variation in a signal and is generally categorized as either Signal Noise or Process Noise. Signal Noise is high frequency and tends to result from electrical interference or improperly specified instrumentation. As examples, consider a pump that kicks on then off sending an electrical upset throughout a plant’s circuitry, or a PID controller that captures imprecise sample time stamps and thereby creates irregularities in the information processed. Process Noise on the other hand is low frequency interference emanating from the process itself. Whether the result of bubbles, splashes, or other process dynamics this form of noise impedes accurate measurement and subsequently hampers effective control.

Filtering is a common approach for mitigating the negative effects of noise. Most modern PID controllers such as the Honeywell UDC2500 and Siemens S7-300/400 series controllers include internal filters just for that purpose. An internal filter mathematically manipulates the process’ signal so that the true, underlying dynamics can be analyzed. Wonderful as that capability may sound the pros and cons of internal filters merit a bit more attention. Consider the following:

  • Do’s and Don’ts of Derivative

As shared in a previous post the Derivative Term responds directly to variability in the Process Variable (PV) signal. As such a noisy PV signal results in an equally agitated reaction by the Final Control Element (FCE), accelerating wear and tear unnecessarily and potentially resulting in unplanned downtime. Internal filters effectively correct for Process Noise in the PV signal, damping the signal’s variability and curbing excessive wear on the process’ mechanical resources.

  • Dangers of Adding Delay

 

Internal Filters and Impact on Observed Dead-Time

While the use of an Internal Filter smooths the PV signal it also contributes to an increase of the process’ observed Dead-Time and Time Constant.

Correcting for Process Noise does come at a cost – the mathematical manipulations that remove volatility in the PV signal also mute the controller’s responsiveness. Specifically the newly filtered signal increases the observed Dead-Time and Time Constant of the process – albeit marginally – which can further limit control loop performance. If Dead-Time is considered the “Killer of Control”, then increasing its value should be done with eyes wide open.

  • Low Frequency Focus

While internal filters can correct for low frequency Process Noise they leave the negative influence of high frequency Signal Noise unattended. Recall that Signal Noise is most commonly the result of electrical interference – signals that are not produced directly by the process. That form of signal variability is best corrected through normal maintenance procedures.

Internal filters are an effective way of addressing Process Noise so that accurate measurement of a process’ dynamics can be made and effective control can be achieved. That said, be aware that there are other available filtering options – topics that will be covered in future posts.

These resources offer related content:

Aug2018_blog-01

Rockwell PID Execution Time: the Difference Between Continuous and Periodic Tasks

Have you ever seen an Integral Term on a PID controller that seems an order of magnitude larger or smaller than what should be used?  Often, the issue stems from the method in which logic is executed on the Process Automation Controller (PAC). More specifically,...

How Do I Isolate the Root-Cause of Control Loop Performance Issues?

  Power Spectrum: A Forensic Tool for Advanced Loop Interaction and Root-Cause Analysis Finding controller performance issues in a typical production facility is not hard, but isolating and correcting their root-cause can present real challenges. Ask around stories abound of production staff who tune a...
shutterstock_590471012

Is the Harris Index Right for Your Loop Assessment Needs?

The Harris Index is an Often Cited Measure of Control Loop Performance.  Even So, Its Complexity Could Mean that Other KPIs Are a Better Alternative. Predicting the weather is difficult – just ask a meteorologist. Even with the help of advanced analytics few seem to...

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.