Today
Rates known at test frequency
- Per-well rates known at test frequency and assumed constant in between.
- Water cut trends detected when a test happens to reveal them.
- Allocation performed on periodic figures without stated uncertainty.
Rates between well tests,with the uncertainty stated.
INSUS estimates per-well oil, gas and water rates from pressure and temperature sensors already on the wellhead, calibrated against the well tests the asset already runs. Production engineers decide how to use every estimate.
01
Rates estimated between tests, with confidence bounds.
02
Runs on wellhead sensors and well tests you already have.
03
Advisory only. Not custody transfer or fiscal measurement.
The problem
Decisions made dailyfrom measurements taken monthly.
Almost every production decision depends on knowing what each well is producing: lift allocation, choke settings, water management, intervention planning and allocation accounting. Well tests are high quality but infrequent; a test separator handles one well at a time. Multiphase meters give continuous readings but drift and cost specialist mobilisation. Optimisation then runs on figures that may be weeks old, and water breakthrough or declining lift performance is found at test frequency rather than as it develops.
01Production varying between tests
02Well test: accurate, single point
03A change that falls between tests
- 01Production varying between tests
- 02Well test: accurate, single point
- 03A change that falls between tests
The shift
From periodic measurementto continuous estimation.
The well test remains the reference. What changes is whether the days between tests are blank or modelled.
With INSUS
Rates estimated between tests, with bounds
- Per-well rates estimated continuously from existing instrumentation.
- Rate and water cut trends visible as they develop, with confidence bounds.
- Calibration checked against every new well test automatically.
How it works
From wellhead sensorsto calibrated rate estimates.
This fits OptiStudio as a per-well soft sensor above SCADA and the historian. Access is read-only. It has no control authority over chokes or lift. Not custody transfer or fiscal measurement.
Connect
Link wellhead, lift and well test data
Connect read-only to wellhead and choke pressures and temperatures, choke position, downhole gauges where available, artificial lift sensors, separator and export metering, well test history, PVT data and completion configuration.
Understand
Build a hybrid model per well
A per-well model relates measurable conditions to tested rates, combining first-principles flow modelling with data-driven correction learned from that well's own test history.
Predict
Estimate oil, gas and water with uncertainty
Produce continuous oil, gas and water rate estimates per well, each with an uncertainty band. The band is as much the deliverable as the estimate.
Detect
Flag developing changes between tests
Flag water breakthrough, declining productivity, liquid loading, and choke or meter drift as they develop. Production engineers decide how to act.
Verify
Treat every new well test as a check
Every new well test is a labelled observation. Model error is tracked per well, and where uncertainty has grown too wide, the system says so rather than producing confident numbers.
Getting started
Start with one platformand the well tests you already hold.
For upstream assets where per-well rates are needed more often than a test separator can deliver. Delivered through OptiStudio as a per-well soft sensor. Expert Services can help with data assessment and validation. Advisory and read-only. Not a fiscal measurement solution.
- What we need. Wellhead pressure and temperature instrumentation, well test history with operating conditions, matching historian data, and PVT and completion records. No new hardware for instrumented wells. A well with no instruments cannot be modelled.
- What you see first. A back-test of per-well models against well tests you already hold on one platform, pad or field section, reporting error per well and where the model is not good enough, before estimates enter the production workflow.
- Where the limits are. Well tests stay the reference. Estimates are never presented as measurements. Access is read-only, with no control authority. Fiscal or custody transfer use is out of scope unless partners and regulators agree the role of an estimate.
Related for upstream
Other upstreamstarting points
Virtual flow metering sits beside gas lift allocation that consumes those rates, and methane reconciliation on the same production data layer.
- 01Gas Lift Allocation AdvisorAllocate limited lift gas across wells against current response and facility constraints.
- 02OGMP 2.0 Level 5 Methane ReconciliationReconcile source-level methane inventories with site-level measurements and keep the evidence trail ready for verification.
- 03Upstream industryThe wider picture for wells and production facilities: lift, separation, flare and the methane evidence underneath export.
Questions production engineers ask us
Well tests are accurate but infrequent, so many production decisions run on rates that may be weeks old. This use case estimates per-well oil, gas and water rates between tests from existing wellhead sensors, with uncertainty stated.
Wellhead pressure and temperature instrumentation, well test history with operating conditions, matching historian data, and PVT and completion records. No new hardware for instrumented wells. A well with no instruments cannot be modelled.
No. Well tests are the reference the model is calibrated against. Multiphase meters can be complemented, including as a cross-check on meter drift.
No. Where allocation carries legal or contractual weight, the role of an estimate has to be agreed with partners and regulators.
No. It produces estimates, not control actions, and has no authority over chokes or lift. Production engineers judge where estimates are usable.
Through OptiStudio as a per-well soft sensor. Expert Services can support data assessment, integration and validation with your production team.
Accuracy depends on instrumentation and well test record quality. Models degrade as fluid properties shift, which is why recalibration tracking is first-class. Virtual flow metering is an established category, so expect comparisons with incumbents on accuracy.
Upstream · Production measurement
What were your wells doing on the days you weren't testing?
A back-test on your own well test history will show how well that question can be answered from data you already record, and for which wells it cannot.