Raw data off a flow computer is not yet a number you can book. Sensors fail, communications drop, inputs drift, and gaps appear, so before a measured volume becomes a production or custody figure it passes through a formal quality gate called VEE: validation, estimation, and editing. This is the measurement industry's structured process for checking data against rules, filling missing intervals defensibly, and making controlled corrections while preserving exactly what was changed. This page defines each step of VEE, what a good audit record of the process contains, and how it sits between raw EFM data and a reportable volume.
Flow Measurement Validation, Estimation & Editing (VEE) in one line: VEE stands for validation, estimation, and editing, the workflow measurement analysts run on raw flow data before volumes are booked. Validation checks the data against rules such as range and rate-of-change limits, estimation fills intervals that are missing or failed validation using defined methods, and editing applies controlled corrections, with the original and changed values both preserved so the process leaves an auditable before-and-after record.
Validation is the first gate, where incoming measurement data is tested against rules that flag values unlikely to be real. Range checks catch readings outside physically possible or expected bounds, such as a differential pressure that has pinned or a temperature that has gone off-scale. Rate-of-change checks flag jumps too large or too fast to be genuine process behavior, which often signal a sensor glitch or a communication artifact rather than a real event. Differential and cross-checks compare related inputs to catch a reading that is internally inconsistent with the others.
The purpose of validation is not to fix anything yet but to separate the data that can be trusted from the data that cannot. Intervals that pass are provisionally good; intervals that fail a rule, or that never arrived because of a communication gap, are marked for attention. This triage is what makes the rest of the process manageable, because it focuses the analyst on the specific intervals that need estimation or editing instead of forcing a review of everything.
Good validation rules are tuned to the measurement point and the process, since a rule too loose lets bad data through while a rule too tight buries the analyst in false flags on normal variation. The rules encode what normal looks like for that meter, so validation is really the codified judgment of what the data should and should not do. Everything downstream depends on validation correctly identifying which intervals are suspect.
Estimation addresses the intervals where real data is missing or has failed validation and cannot simply be used. Because a booked volume usually cannot have holes, the analyst fills those intervals with an estimated value derived by a defined method, such as carrying forward the last good reading, interpolating between known good points, or using a representative value from comparable conditions. The key is that the method is deliberate and documented rather than an arbitrary guess, so the estimate can be explained and, if better data later appears, replaced.
Editing is the controlled correction of data that is present but known to be wrong, for a reason the analyst can justify, such as a mis-scaled input or a known sensor bias during a period. Unlike casual overwriting, editing in a VEE process is bounded: the analyst records the reason, keeps the original value, and stores the new value alongside it. The discipline is what separates a legitimate correction from tampering, because the change is transparent and reversible rather than hidden.
The non-negotiable feature of both estimation and editing is the preserved before-and-after record. Every estimated or edited interval retains its original value, its new value, the method or reason, and typically who made the change and when. This audit trail is what lets a corrected volume remain defensible: a reviewer can see not just the final number but the full history of how it got there, and can undo or revisit any change. Estimation and editing without that record would be indistinguishable from altering the data.
VEE sits in a specific place in the measurement pipeline: after raw electronic flow measurement data is collected but before volumes are reported into production accounting, allocation, or custody settlement. Raw EFM data is the input, and a validated, gap-filled, corrected data set with a complete edit history is the output. Framing VEE as this quality gate is what keeps unreliable readings from silently flowing into figures that money and reporting depend on, while still allowing legitimate repair of the inevitable gaps and errors that real field measurement produces.
A cloud SCADA and EFM platform is what makes VEE practical at scale, because the process needs both the raw data and a place to record the edits. A platform such as Merobix collects the interval data and event logs from flow computers across many sites, giving analysts one workspace to run validation rules, apply estimates and edits, and store the resulting audit trail. Doing VEE on centrally collected data means an analyst can process a whole gathering system's meters from the office rather than chasing records device by device, and the before-and-after history is retained alongside the volumes.
The distinction worth holding onto is that VEE is a formal, auditable measurement process, not just generic data cleaning. General data validation might discard or smooth bad points; VEE requires that every estimate and edit be made by a defined method, justified, and preserved with its original value, precisely because the output feeds custody and production figures that may be challenged. That rigor, backed by a system that retains the full edit history, is what turns messy raw field data into a volume an operator can defend.
VEE stands for validation, estimation, and editing. Validation checks raw measurement data against rules such as range and rate-of-change limits, estimation fills intervals that are missing or failed validation using a defined method, and editing applies justified corrections to data known to be wrong. Each estimated or edited value is preserved alongside the original so the whole process is auditable.
Ordinary data validation may simply flag, discard, or smooth suspect points. VEE is a formal measurement-industry process that additionally requires missing intervals to be estimated by defined methods and known-bad data to be edited with a documented reason, and it insists that the original value be kept alongside every change. That preserved before-and-after record is what makes the resulting volumes defensible for custody and production reporting.
Because the corrected volumes feed custody settlement, allocation, and production reporting, any estimate or edit must be transparent and reversible rather than hidden. Keeping the original value, the new value, the method or reason, and who changed it and when lets a reviewer see exactly how the final number was produced and revisit it if better data appears. Without that record, editing would be indistinguishable from altering the data.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.