How to Set Up Interface Level Measurement

Back to Blog

Choose a Technology for Your Fluids

The first decision is which measuring principle can even see your interface, and it depends on how cleanly the two liquids separate. A buoyancy displacer works where the density difference is real and stable, sensing the change in apparent weight as the interface rises past the displacer. A capacitance probe distinguishes the two liquids by their very different dielectric constants, water being highly conductive and oil nearly an insulator, which makes it strong for oil-on-water but sensitive to coating. A magnetostrictive dual-float system floats one float on the total surface and one at the interface density.

Match the technology to the emulsion. A thick, slow rag layer defeats a sharp-boundary sensor and favours a profiling approach, while a clean split suits a simple displacer. If you are still deciding, the overview of interface level measurement compares the principles, and the sensing physics behind the capacitance option is covered in the capacitance level sensor. The wrong technology cannot be tuned into working, so this choice matters more than any later configuration.

Configure Both Liquid Densities

Almost every interface device needs the specific gravities of both the upper and lower liquid, because it locates the boundary from the density contrast, not from an absolute head. Enter the oil specific gravity as the light phase and the water or brine specific gravity as the heavy phase. A displacer computes the interface from the buoyancy difference the two densities create, so a wrong density puts the interface in the wrong place even when the hardware is healthy. Get both figures from a lab analysis of the actual produced fluids, not a textbook default.

Be honest about how stable those densities are. Produced water carrying dissolved solids and oil of varying gravity will shift the contrast over time, and a large swing is a real limit on interface accuracy that no calibration removes. Where the process runs an interface level controller to dump water, note that the controller acts on this measured boundary, so a density error propagates straight into how much water or oil leaves the vessel. Document both densities and their expected range on the setup sheet.

Zero, Span, and Set the Control Point

With the densities entered, set the measurement range to the physical span over which the interface can travel between the low and high try-cocks or nozzles, mapping 4 mA to the lowest credible interface and 20 mA to the highest. Then decide where in the vessel the control point should sit, remembering that on a separator the interface setpoint balances water carry-over into the oil against oil carry-under into the water. This is a process decision as much as an instrument one.

If a rag layer or emulsion band exists, define explicitly where in that band the measurement is intended to report, because different technologies latch onto different parts of a diffuse boundary. A capacitance probe may switch at the dielectric transition while a displacer responds to the integrated buoyancy across the band. Knowing which behaviour you have prevents a surprise when the control valve cycles on what looks like a moving interface but is really the sensor tracking a shifting rag layer.

Verify Against a Manual Interface Sample

Prove the setup by finding the interface independently. Use the vessel's try-cocks, cracking each in turn from the bottom up until you go from producing water to producing oil, which brackets the true interface between two known elevations. Alternatively pull an interface sample or use a water-finding paste on a dip tape where the vessel allows it. Compare that physical boundary against the interface the transmitter reports on the SCADA screen, and if they agree within tolerance the setup is sound.

Verify at more than one interface height if the process lets you move it safely, because a single point can hide a density error that only shows elsewhere in the range. Log the try-cock result and the transmitter reading together. Once the tag feeds a monitoring history, the periodic manual checks become a running verification: a slow divergence between the sampled interface and the reported one flags a density drift or a fouling probe before it upsets separation and sends water to the sales line.

Avoid the Common Mistakes

The classic error is configuring only one density, or entering the total-level density instead of the two-phase contrast, which leaves the device unable to place the interface correctly. Another is choosing a sharp-boundary sensor for a fluid that carries a thick emulsion, then blaming the calibration when the reading wanders through the rag layer. A coated capacitance probe reads a false interface until it is cleaned. And a displacer that has sprung a leak floods and reads a fixed level regardless of the real boundary.

Interface control is unforgiving because both errors of the setpoint have a cost: too high and oil escapes with the water, too low and water carries into the oil. Trending the measured interface alongside downstream quality indicators, such as water cut on the oil leg, turns a monitoring platform into an early-warning system: when the reported interface and the actual separation quality diverge, the sensor or its densities need rechecking against a fresh manual sample.

Frequently Asked Questions