Eliminating Signal Noise in High-Velocity Gas & Liquid Metering

Signal noise is one of the most persistent and least discussed problems in industrial flow measurement. It manifests as erratic readings, unstable control loops, inflated totalised volume figures, and — in severe cases — complete loss of measurement integrity. In high-velocity gas and liquid service, where fluid turbulence, mechanical vibration, and electromagnetic interference are all simultaneously present, eliminating noise requires a systematic approach that addresses the source rather than masking symptoms.

This article examines the primary sources of signal noise in high-velocity metering applications, practical engineering techniques to eliminate or reduce it, and the role of calibration in maintaining long-term measurement accuracy.

Understanding Signal Noise in Flow Meters

Signal noise in flow metering is any variation in the output signal that does not correspond to actual changes in flow. It can be periodic (at a fixed frequency) or random. Its amplitude can range from a fraction of a percent of full-scale — invisible in most control room displays but significant when integrating to totals over time — to swings large enough to drive a control valve into saturation.

The consequences scale with the application. In a steam distribution system, noise in a flow signal used for billing allocation means one plant site systematically under- or overpays for energy. In a batch process, noise in the flow signal can cause the batch controller to deliver incorrect quantities. In a custody transfer application, noise corrupts the legal measurement.

Mechanical Sources of Signal Noise

Mechanical vibration is the dominant noise source in most high-velocity installations. Vortex flowmeters are particularly sensitive because their operating principle — detecting the frequency of vortices shed by a bluff body — relies on measuring very small pressure fluctuations. Pipeline vibration at frequencies close to the shedding frequency is indistinguishable from actual flow signal at the sensor level.

The solution is mechanical isolation. Where possible, the meter should be installed in a section of piping that is independently supported and isolated from major vibration sources — compressors, pumps, and control valves. Rigid supports within two pipe diameters of the meter inlet should be avoided; they transmit vibration rather than absorb it. Flexible couplings upstream of the meter are effective where vibration sources cannot be moved.

For pressure transmitter installation in high-vibration environments, capillary-remote seal arrangements keep the sensing element away from the vibration source while maintaining measurement accuracy. This is particularly effective on pump discharge lines where both high velocity and high vibration are present simultaneously.

Hydraulic Sources of Signal Noise

Turbulent flow profiles — caused by insufficient straight-run upstream of the meter, partially open valves, tee junctions, or pipe reducers — create hydraulic noise that corrupts the flow signal. ISO 17089 measurement of fluid flow in closed conduits specifies minimum straight-run requirements for various meter types, but these minimums assume ideal installation conditions. In practice, installations with multiple upstream disturbances may require 50% more straight run than the standard minimum.

Swirl is a particularly damaging form of hydraulic disturbance. It is caused by out-of-plane bends (two elbows in different planes within 10 diameters of the meter) and persists for 30–50 pipe diameters downstream — much longer than simple asymmetric profiles. Where swirl cannot be avoided through piping layout, a flow conditioner should be installed upstream of the meter. Modern perforated plate conditioners add 0.5–1.5 bar pressure drop but eliminate swirl effectively within 10 diameters.

In gas metering applications, entrained liquids create a second form of hydraulic noise. Even small quantities of condensate or lubricating oil carried over from compressors cause slug flow that produces severe signal spikes. Adequate gas conditioning — coalescing filters, moisture separators, and in some cases condensate knockout drums — upstream of the metering station is essential.

Digital Filtering and Signal Conditioning

For engineers managing high-velocity pipelines, consult our vortex flowmeter product page for a practical framework for reducing signal noise without sacrificing measurement response time. The fundamental tension in digital filtering is between noise rejection and response speed — a heavily filtered signal is stable but slow to respond to genuine flow changes.

Modern vortex meters offer configurable low-flow cutoffs and signal damping parameters that should be set based on actual process dynamics rather than left at factory defaults. The low-flow cutoff eliminates the erratic readings that occur when flow velocity falls below the minimum reliable shedding frequency — typically 0.5–1.5 m/s depending on the meter size and fluid properties. Setting the cutoff too high masks genuine low-flow events; setting it too low allows unreliable readings to pass to the control system.

Damping time constants of 2–5 seconds are appropriate for most process control applications. Longer constants are appropriate for totalisation where cycle-to-cycle variability matters less than long-term accuracy. Shorter constants — or zero damping — should only be used where fast process dynamics genuinely require it, as shorter damping significantly increases the noise floor.

Calibration as a Noise Reduction Tool

Signal noise that appears to originate externally sometimes has an internal cause: the meter itself is performing outside its calibrated range due to wear, fouling, or damage. A vortex meter with a partially fouled bluff body sheds vortices at an irregular frequency, producing a noisy signal even in an otherwise clean installation.

Regular flowmeter calibration services confirm that the meter's K-factor — the relationship between shedding frequency and volumetric flow — remains within specification. Meters that have drifted outside their calibrated tolerance produce systematic errors that look like noise at the control room level but are in fact consistent bias. Calibration intervals of 12–24 months are appropriate for most clean fluid applications; shorter intervals are warranted in dirty service or where the measurement is used for custody transfer or billing.

Electromagnetic Interference Considerations

In industrial environments with variable-frequency drives, welding equipment, and large motor starters, electromagnetic interference (EMI) can corrupt the signal from sensors with inadequate shielding. For instrumentation and control systems in Singapore, best practice is to run sensor cables in dedicated instrument cable trays separated from power cables by at least 300mm, use individually shielded twisted pairs with shields grounded at one end only, and verify that all instrument earth connections are made to a clean instrument earth rather than the structural steel earth.

Conclusion

Eliminating signal noise in high-velocity gas and liquid metering is an engineering discipline that spans mechanical design, hydraulic analysis, signal processing, and electromagnetic compatibility. Addressing noise at its source — through proper straight-run, vibration isolation, flow conditioning, and correct digital filter settings — produces more reliable measurement than any amount of downstream signal processing. Combined with regular calibration, a systematic approach to noise elimination delivers the measurement accuracy that modern process control and energy management systems require.

Uneeb Khan is the founder of Techager and has over 6 years of experience in tech writing and troubleshooting. He loves converting complex technical topics into guides that everyone can understand.