Case Study: Tracing Erratic Flow and Pressure Readings to Two-Phase Flow in a Liquid Chemical Line

Case Study: When the Instrument Is Right and the Fluid Has Changed

A chemical plant reported a problem that survived two instrument replacements. Flow and pressure readings on a transfer line jumped without warning, while pump frequency held steady, no valve moved and the 4-20 mA loops were healthy. Every instinct pointed at the instruments. The measurements were correct: the fluid inside the pipe was no longer the single-phase liquid the line was designed for.

Key finding: a fluctuating reading is not always a measurement failure. In this case the reading was an accurate report of an unstable process condition, and replacing hardware would only have hidden it.

1. Site Background

2. Where the Gas Came From

The investigation started upstream rather than at the meter. Three gas sources were confirmed on site:

• Ingested gas: a low tank level let the pump draw a vortex at the suction nozzle, and a slightly weeping valve stem packing admitted air without leaking liquid.
• Dissolved gas breakout: a temperature rise combined with a pressure drop released gas that had been dissolved in the liquid, forming free bubbles downstream.
• Flashing and cavitation: pressure fell below vapour pressure across the control valve and orifice plate, liquid vaporised locally and bubbles collapsed in the higher-pressure region, producing noise, vibration and trim erosion.

Each source needs a different remedy, so identifying the origin came before touching any instrument.

3. Flow Regime Explained the Pattern

Once gas entered the line the mixture passed through several regimes. Bubbly flow produced slight jitter. Slug flow produced the periodic surges the operators complained about, with alternating gas and liquid slugs crossing the tapping points. Stratified flow appeared in horizontal sections, where tapping position determined a fixed offset. Annular flow at higher gas velocity created high-frequency pulsation in the differential pressure signal.

The decisive observation was that the fluctuation had a rhythm. Slug passage and migrating gas pockets both produce step-like drift, and neither means a failed transmitter.

4. Why Each Meter Reacted Differently

5. Structured Troubleshooting

1. Confirm the fluctuation is real: a mechanical gauge teed into the same tapping point pulsed in step with the transmitter. With valve position locked and pump speed stable, the trend still oscillated, and a gravel-like noise was audible at the pump discharge.
2. Locate the phase change: tank level, pump suction pressure and valve upstream and downstream pressures were trended together. Gas accumulation at high points and partially blocked impulse lines were found and cleared.
3. Converge on the loop and parameters: 24 V supply, single-point shield grounding and separation of signal from power cable were verified. Damping was set to 0.5 to 1 s for liquid service and 1.5 to 2 s for gas service, low-flow cutoff was enabled and DCS secondary filtering applied, with the note that damping smooths a signal but cannot remove genuine two-phase pulsation.

6. Corrective Actions and Results

Upstream measures carried most of the benefit: the tank operating band was raised to stop suction vortexing, a deaerator and buffer vessel were added ahead of the pump, tapping points were moved clear of elbows and valves, impulse lines were re-sloped with periodic venting, the Coriolis meter was reinstalled full-bore without stress, and a shift routine of vent, drain, zero check and condensate compensation was introduced.

After commissioning, the flow indication band on the affected line narrowed from roughly plus or minus 18 percent to about plus or minus 3 percent, empty-pipe alarms stopped, totaliser deviation against tank gauging fell within 1 percent per batch, and no further replacements were required. Range settings were also corrected so that normal flow no longer sat in the bottom of an oversized span.

Conclusion

An instrument's job is to report the process faithfully, not to flatten it. Persistent flow and pressure fluctuation is often the process telling the engineer that the fluid in the pipe is no longer the fluid the line was designed for. Reading the instrument is routine; reading the fluid state behind the instrument is what actually solves the problem.