Instrument noise, and eliminating instrument noise, is important to consider in
process control instrumentation. Noise represents variations in process variable measurement that is not reflective of actual changes occurring in the process variable. Typically, electrical devices such as high voltage wiring, electric motors, relays, contactors, and radio transmitters are the primary sources of instrument noise.
No matter the cause for the process noise, the measurement signal in the process is being distorted and is not reflecting the true state of the process at a certain time. Accuracy and precision of process measurements are negatively affected by noise, and can also contribute to errors in control system. Controller output can reflect the noise affecting a process variable.
Grounding allows for the reduction of noise stemming from electrical systems. Shielded cabling and separating signal cabling from other wiring, as well as replacing and repairing sensors, allows for noise reduction. Low-pass filters are a way to compensate for noise, and much of the instrumentation used in process systems incorporates noise dampening features automatically. Determining the best kind of filter to use depends heavily on cut-off frequency, alpha value, or time constant.
The ideal low-pass filter would eliminate all frequencies above the cutoff frequency while allowing every frequency below the cut-off frequency to be unaffected. However, this ideal filter is only achievable mathematically, while real applications must approximate the ideal filter. They calculate a finite impulse response, and also must delay the signal for a bit of time. To achieve better filter accuracy, a longer delay is needed so that the filter computation “sees” a bit further into the future. The calibration of these filters heavily relies on the desired accuracy level of the process, while also taking specific steps in calibration to best fit a particular process.
Noise is important to mitigate because the noise observed while measuring the process variable can produce “chatter” in the final control element of a process. The resulting “chatter” increases the wear of mechanical control elements, such as valves, and will generate additional cost for the process as a whole. The filtered signal lagging behind the dynamic response of the unfiltered signal is a result of the filtered signal’s increased dead time, meaning that signal filters add a delay in sensing the true process state. The solution is to find a mid-point between signal smoothing and information delay, which allows for elimination of noise while not negatively affecting the speed by which information is delivered.
For question about any process control application, or challenge, visit
https://meadobrien.com or call
(800) 892-2769