A Coriolis meter measures mass flow by sensing a tiny phase difference in its vibrating tubes, and it needs to know exactly what that phase reads when there is genuinely no flow. Zero drift is when that no-flow reference moves over time, so the meter shows a small flow when the line is actually static, or carries an offset that quietly biases every custody total. It is one of the more insidious Coriolis problems because the meter looks perfectly healthy and the error is small enough to escape casual notice while still adding up. This page explains why the zero shifts, how to separate real drift from a process leak, and when a field zero is worth doing.
Coriolis Meter Zero Drift in one line: Coriolis meter zero drift is a shift in the meter's stored no-flow reference point, so under genuinely zero-flow conditions the meter reads a small nonzero flow, or applies a constant offset that biases readings and totals across the whole range. It arises from things like changing mounting stress, temperature, pipe loading, and component aging that alter the tubes' balanced no-flow state. It matters most at low flow and for custody totals, where a small standing offset accumulates into real volume error over time.
A Coriolis meter works by driving its measuring tubes into vibration and detecting the phase shift between the inlet and outlet sensors that flowing mass creates. At zero flow that phase shift should be zero, so the meter is calibrated with a zero value that represents the sensor reading under confirmed no-flow conditions. Because the measurement is a small difference between two vibrating pickups, it is sensitive to anything that changes the mechanical balance of the tubes, and the stored zero is only correct as long as the tubes' no-flow state stays the same as it was when the zero was set.
Several real-world influences push that no-flow state around. Mounting and installation stress is a common one, if the meter is under mechanical strain from the piping, and that strain changes as the pipe settles, thermally expands, or is disturbed during nearby work, the balance of the tubes shifts and so does the effective zero. Temperature is another, both process and ambient temperature change the stiffness and geometry of the tubes, and while good meters compensate for much of this, residual effects can appear as zero movement. Pipe loading, meaning forces and moments transmitted from the connected piping, and slow aging of the sensor and its components round out the picture.
The result is a zero that was correct at commissioning but no longer represents true no-flow after months or years in service. The drift is usually small, but its impact is disproportionate at low flow, where the offset is a large fraction of the actual reading, and in totalizing applications, where even a tiny standing offset integrates into a meaningful volume over a long period. A meter reading a trickle when a line is confirmed dead-headed, or a totalizer that creeps upward with no flow, is the visible signature of a zero that has drifted.
A Coriolis meter that reports a small flow when you believe the line is shut in can mean two very different things, and confusing them leads to the wrong action. One possibility is genuine zero drift, the meter's reference has moved and it is reporting flow that is not physically there. The other is that there really is flow, a passing valve, a leaking seal, or a bleed path is letting a small amount of fluid move, and the meter is doing its job correctly by detecting it. Concluding drift when there is actually a leak means ignoring a real loss; concluding a leak when it is really drift means chasing a phantom.
The way to separate them is to establish, independently, that there is truly no flow, and then see what the meter says. That means genuinely isolating the meter, closing block valves on both sides so no fluid can move through it, and letting conditions stabilize before reading. If the meter still shows a nonzero flow when it is properly blocked in and static, that reading has to be an artifact of the zero, because no fluid can physically be moving. If the apparent flow disappears when you truly isolate it, then the earlier reading was likely real flow, and the investigation shifts to which valve or path is passing.
This distinction is also why a zero-verification procedure insists on confirmed no-flow conditions before touching the zero. Setting or checking a zero while fluid is quietly passing bakes that leak into the reference, corrupting every subsequent reading. The discipline of proving no-flow first, then interpreting a standing offset as drift only once flow is genuinely impossible, is what keeps you from mistaking a leak for a calibration problem or vice versa. Where a meter reads consistently offset in a direction that a leak cannot explain, or where the offset appears identically across conditions, drift is the more likely story.
The most reliable way to catch zero drift is to observe the meter during confirmed no-flow windows and see whether it reads truly zero. Rather than relying on a technician remembering what the meter did the last time it was calibrated, a monitoring approach trends the flow reading whenever the process is known to be static, such as during shut-ins, batch gaps, or overnight idle periods, and watches for a persistent nonzero value. A standing offset that appears every time flow stops, and that grows or shifts over months, is the fingerprint of drift, and it is far easier to see in a trend than in a single spot reading.
This is a natural fit for cloud SCADA because the platform already holds the flow history and can correlate it with no-flow conditions. When a system like Merobix logs the flow reading over time and can be told when the line is confirmed static, it can alarm on a standing offset during those windows and preserve each zero-verification result as a time-series baseline. That baseline turns zero stability into something you can actually see trending: instead of one number in a calibration logbook, you have a history of what the meter read at zero flow every time the line was down, so a slow drift stands out and a sudden jump after some field work is immediately obvious.
The decision of whether to run a field zero rests on that evidence. A field zero re-establishes the reference under actual installed conditions, which corrects drift caused by real mounting stress, temperature, and aging that the factory zero could never anticipate. But a field zero done under poor conditions, with residual flow, unstable temperature, or a disturbed installation, can be worse than leaving the factory zero alone, because it locks in a bad reference. The sensible rule is to field-zero when the trend shows meaningful, verified drift and you can achieve clean no-flow conditions, and to leave the factory zero in place when the meter reads acceptably at confirmed no-flow. Logging each zero result also means that if a field zero makes things worse, you have the before-and-after record to recognize it and revert.
The two likely explanations are zero drift and a real but small flow. Zero drift means the meter's no-flow reference has shifted so it reports flow that is not physically present, while a passing valve or leaking seal could be letting a genuine trickle move. To tell them apart, isolate the meter with block valves on both sides so no fluid can move, let it stabilize, and read it: a nonzero reading under true no-flow conditions is drift, while a reading that disappears when properly isolated was real flow.
The zero drifts when the mechanical no-flow state of the vibrating tubes changes from what it was when the zero was set. Common causes include changing mounting and installation stress, thermal effects from process and ambient temperature, forces transmitted from the connected piping, and slow aging of the sensor components. Because the measurement is a small difference between two vibration pickups, even modest changes in the tubes' balance can move the effective zero.
Run a field zero when trending shows meaningful, verified drift and you can achieve genuinely stable no-flow conditions to zero against. A field zero corrects drift from real installed mounting stress, temperature, and aging that the factory zero could not anticipate. But avoid zeroing under poor conditions such as residual flow or unstable temperature, because that locks in a bad reference; if the meter reads acceptably at confirmed no-flow, it is often better to leave the existing zero alone.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.