Poll Results: “How did THIS flowmeter end up HERE?!?” and other disasters..

In a new edition of Engineering Ideals vs. Sales vs. Real Life: “How the hell did THIS flowmeter end up HERE?!?”. Was it a shiny slide? A procurement shortcut? A great barbecue? Perfect presentation? Cheapest option? Poker play results? Maybe a galactic alignment? Shaman divination? … (it does feel like it sometimes..)

We all say accuracy is king… but let’s be honest: The throne is often shared with budget, relationships, procurement, and many heirs…

In a (not so much..) surprising turn of events, our flowmeter selection poll ended with a tie! With all options practically equal in consideration, thus perfectly reflecting how we all asked ourselves at some point this same :

“How did THIS flowmeter end up HERE?!”

Turns out to be a very valid question.

Are flowmeters selected by their technical prowess? Reputation? Sales magic? Price? or is it a process involving all of these or none at all?

The fact is, most of the best work in picking the right flowmeter is done beforehand, but then, how does a ”shall be” become a “should be”, become a “could be”, into a “could have been” to a “should have never been”?

First of all, let’s address the elephant in the room: Mature wells.

🛢️ The Elephant in the Room – Mature Wells

Most flowmeters were originally well selected. They performed within expected uncertainty limits and served the field operations accurately for years. Yet, as production matures, process conditions evolve, and the key parameters that define the flowmeter’s suitability shift dramatically.

Over time, production profiles change:

• Flow rate decline alters the Reynolds number, affecting flow regime stability and signal response.
• Water breakthrough increases dielectric contrast and changes permittivity behaviour.
• Gas lift changes the local phase distribution, bubble or slugs frequency, and slip ratio.
• Salinity changes alter conductivity-based sensors and affect electrical impedance.
• PVT properties evolve as reservoir pressure drops, shifting gas-oil ratios and phase equilibria.

What was once a perfectly matched flowmeter at commissioning becomes mismatched years later. Initial working envelopes shrink relative to the new operating conditions, and drift correction becomes guesswork. What are the consequences? Increased bias, flow instability, and false confidence in the real flowmeter performance.

This is what we in PandA call: Metrological Ageism (Mature wells have still a lot to give is not an euphemism)

This situation calls for a Continuous Suitability Management (CSM) approach; an operational philosophy ensuring that flowmeters remain within their performance envelope throughout the asset lifecycle.

🧭 Key Elements of Continuous Suitability Management

1. Routine Validation and Benchmarking: Compare your MPFM (Multiphase Flowmeter or Wet Gas Meter) readings with periodic well-test separator data, tracking bias and variance. A 5% deviation trend may signal out-of-envelope operation.
2. Adaptive Configuration: Adjust PVT tables, salinity compensation, or model coefficients as production fluid properties evolve. Most modern MPFMs support this but are rarely updated post-commissioning.
3. Cross-Flowmeter Strategy: Reassign wells to alternative MPFMs or Single-phase Flowmeters with better-fitting envelopes. Mature fields often have underused flowmeters that could improve allocation accuracy if intelligently switched.
4. Historical Drift Analysis: Maintain a rolling baseline of measurement residuals (ex: Difference between MPFM and a test separator or periodic well testing equipment). Use time knowledge, which means trending and regression, to predict failure or recalibration needs. PandA has developed a solution in this area.
5. Lifecycle Budgeting: Allocate enough OPEX not only to new projects but also to sustain measurement quality on aging assets. Mature wells still contribute significantly to production, neglecting them distorts corporate reporting. These are low-hanging fruit that could still bring a significant net.

Unfortunately, ageism in flow metering is real. Mature wells often get deprioritized in favour of new shiny developments (AKA: greenfield), albeit they still hold valuable production potential. This imbalance leads to systematic under-measurement, poor reservoir diagnostics, and unquantified production losses.

“A mature well with outdated flow metering is like a symphony played on a detuned instrument. It still performs, but no longer in tune with reality.”

Future planning should therefore integrate dynamic flow metering requalification as part of field life extension. Regular reassessment, model recalibration, and smart resource allocation ensure that every barrel is properly accounted for, whether it comes from a brand-new subsea tieback or a 25-year-old legacy well.

Now that the Elephant is well accounted for (and out of the room), let’s address the processes in place which contributed to the Flowmeter Mismatch.

How does a ”shall be” become a “should be”, become a “could be”, into a “could have been”, to a “should have never been”

🔬 SHALL BE – The Technical Definition Phase

At this point, it’s tempting to reach for the latest and most accurate flowmeter on the market,  the one with the glossy brochure and impressive uncertainty numbers. But that shiny device could turn out to be a perfect mismatch. The key is not to believe the claim, but to verify whether that claim fits your actual process reality; the pressure, temperature, GVF, and WLR conditions your field really lives in, not the ones it dreams about.

To do this properly, we must translate process conditions into flowmeter working envelopes. Parameters such as gas volume fraction (GVF), water liquid ratio (WLR), viscosity, density, sand content, chemical injections, and transient flow regime frequency define the true measurement challenge. The goal is to select a flowmeter whose flowmeter performance envelope overlaps with the operating envelope for at least 90 % to 95 % of the production lifetime (reporting correctly 18 out of 20 times, or 19 out of 20 times). Anything less is wishful thinking.

Which is why every credible selection begins with a third-party qualification test, ideally a double-blind test conducted at an ISO 17025-accredited facility. In simple terms:

• The vendor doesn’t know the flow rates or composition at the test point.
• The vendor does not get any information until the data has been provided to the witness person, and this is set with no possible recovery, like in the actual field conditions.
• The flow loop operator does consider the entire set of data of the results afterward for the report.
• The witness or the end-user selects the flow rate and conditions at the last minute within a range earlier defined and representative of the field flow conditions.

Only when both sides are blind do we finally see the flowmeter’s true performance. This is where science should lead the dance, not PowerPoint slides.

Yet, this crucial first selection stage is often clouded by a dangerous mix of experience, assumptions, and optimism. Too many decisions are based on “I’ve seen this flowmeter work before” instead of “Let’s confirm it statistically for our field.”

Reality check: if nobody mentions the Binomial Law or confidence intervals, you’re still in the opinion zone, not the verification zone. In flowmeter testing, the standard deviation of error distribution (σ) relates directly to the number of data points (n), as σ/√n – too few points, and uncertainty balloons faster than the project cost curve.

A rigorous approach demands understanding that total flow uncertainty is not a mystery but a measurable function, for example:

U_total = √(U_repeat² + U_linearity² + U_bias² + U_drift²)

Each component is traceable, and each affects your bottom line. Proper definition, testing, and traceability, not brand names, define real accuracy.

Remember, we’re not here during this test to challenge the manufacturer’s claims; we’re here to verify that they comply with their own statements under realistic field conditions and with no data manipulation after receiving the references. That line between declaration and demonstration can easily be worth millions, and some companies could use a flowmeter for decades and struggle with it because of this main milestone not being met.

This phase should remain the kingdom of technical experts: engineers who speak the language of ISO 10790, ISO/TR 12748, and API MPMS 22.2, … and who understand what a proper uncertainty budget means. It’s the stage where statistics, metrology (and patience), come together.

Pro tip: Always bring an SME ( Subject Matter Expert) in multiphase flow metering to the test site. It’s like bringing a chess grandmaster to a poker table: he may not know the rules of bluffing, but he’ll spot the patterns no one else sees and decode them. (Ref: Chess master and champion vs. Enigma machine)

IE (Idem Est): this is where technical knowledge and priorities shine (if applied correctly). Bring science to the field, not just opinions. And never underestimate how many millions of barrels depend on getting this step right.

💼 SHOULD BE – When Sales Enters the Room

This is the phase where competition steps into the arena, and  more often than not, the best technical salesperson wins, not necessarily the best technology.

Perhaps the multiphase flowmeter the operator had in mind isn’t locally available. Or maybe the local vendor team is smaller, less experienced, or simply out-marketed by a more polished competitor. Suddenly, the battlefield shifts from equations (hardcore science) to presentations (sweet confident talk).

This is when the PowerPoint parade begins: polished animations, glossy brochures, and the ever-popular claim of “± 1 % accuracy across all GVF and WLR ranges”. It’s also when the best-curated slideshow can defeat a decade of R&D, when a perfectly timed barbecue or a visit from overseas “product experts” magically turns uncertainty into enthusiasm.

The pitch becomes irresistible: “Our technology is revolutionary… near perfect!” and this precisely is where many projects start to lose the plot.

IE: Don’t fall into the PowerPoint trap. Be technical. Review the data. Demand numbers, not adjectives.

A skilled salesperson can sell confidence better than performance (said from experience), and your only defence is validation, not persuasion. If a vendor cannot produce traceable, independent performance data from accredited facilities (CEESI, DNV, TÜV-SÜD, SwRI, etc.), their claim is, by definition, unverified. Ask for data that spans realistic flow regimes: annular, slug, and bubble flow. A single homogeneous test point doesn’t mean field reliability; it just means a very calm day at the lab.

Technical Insight: Flowmeter physics is not a guessing game. Multiphase flow measurement(for instance) relies on coupled correlations between impedance, attenuation, permittivity, and phase fraction estimations. Any credible vendor report includes calibration equations, confidence bands, and sensitivity analyses – for example, dOutput/dGVFand dOutput/dWLR. These parameters reveal where the flowmeter holds and where it wobbles.

None of the flowmeter manufacturers today has the ultimate answer here. There is no unicorn MPFM immune to process variability. Every technology has its own trade-offs in rangeability, GVF sensitivity, or calibration drift. Your mission isn’t to find the “perfect” flowmeter ; it’s finding the one that’s perfectly fit for your field conditions.

That requires genuine work: reviewing uncertainty matrices, comparing vendor methodologies, and assessing data underrepresentative process conditions. Bring in an SME in multiphase flow metering if needed. Paying an expert $100k is negligible compared to the billions your field will generate over its life expectancy.

Remember: anyone who proclaims “next-generation accuracy” without third-party validation should trigger your scientific instincts. There’s no way to solve the Navier–Stokes equations for three-phase flow on a PowerPoint slide, no matter how impressive the animations are.

In other terms: If it sounds too perfect, it probably is (valid for most things in life, but mostly Flow Metering).

IE: Salespeople only shine when engineers dim their own light. Take ownership of the problem and the deployment. Request an SME to stand beside you; not to sell, but to translate between physics and promises (the hardcore reality vs. dreamland).

And if someone confidently tells you, “Don’t worry, we’ll figure it out after installation with the right setup/calibration” run! Preferably toward the nearest independent flow loop with a credible witness and a properly designed validation test plan.

🏗️ COULD BE – The Procurement enters the chat and takes the wheel: Integration Phase

Here comes the tricky part: buying the flowmeter.

Whether it’s for a shiny new development or a mature field upgrade, this is where flow metering stops being engineering and starts being procurement ; where precision gives way to politics, and “best technical choice” starts wrestling with “best commercial package.”

Buying a flowmeter in this context isn’t a relaxed shopping trip; it’s a wholesale exercise where the flowmeter is just one tiny line in a billion-dollar grocery list. Your favourite flowmeter? It’s probably tied to a not-so-favourite package. So, you end up trying to pick the best flowmeter within the best overall package, like choosing the best olive in the jar. Except someone else sealed the lid.

Now, could you buy the flowmeter separately? In theory, yes. In practice, not so much. Large EPCs and integrators usually come with their own preferred-vendor ecosystems, invisible fences built around commercial alliances and legacy relationships. When you ask about alternative options, the answer often sounds suspiciously like:

“Yes… But no.” (Said in corporate language).

It’s a kind of flow metering mafia : polite, smiling, and quietly structured so that reliability (and future blame) shifts squarely onto you.

Once the tender documents start flying, the flowmeter becomes just another component in a sea of line items. However, it is the key device to improve your recovery factor (aka allowing you to understand the reservoir behaviour over time and squeezing it in the best way). The procurement goal is simple: minimize cost (CAPEX), maximize compliance. But flow metering isn’t a bolt-on commodity, it’s a scientific discipline that anchors your production data, fiscal allocation, and carbon reporting.

Integration is where the silent errors creep in. Once a flowmeter is “buried inside a package”, a whole range of new biases enter the scene:

• Cable interference and grounding loops;
• Suboptimal upstream/downstream straight lengths;
• Control logic mismatches;
• Signal processing shortcuts;
• Environmental compensation drifts no one notices until the numbers stop making sense;
• Fluid properties;
• Chemical Injection…

Technical Insight: Integration without uncertainty re-verification breaks the metrological traceability chain. The calibration or performance certificate that came with your flowmeter is only valid in the configuration it was tested for. Once you change geometry, orientation, vibration, or control logic, you’re not “maintaining accuracy”,  you’re in a way gambling with it.

A deviation of just ± 2 °C in temperature compensation or ± 1 bar in pressure reference can shift your density correction by over ± 0.3 %, which, at the field scale, translates into millions in allocation discrepancies.

This issue is common in subsea projects, where integrators often see the flowmeter as “just another expensive pipe section”. But those of us who’ve reconciled subsea production data know better: it’s the heart of your revenue stream, the only instrument ensuring your anti-freeze keeps flowing, your accountants keep smiling, and your partners keep arguing over something other than measurement.

IE: This is the phase where the “biggest package wins” mentality takes over. Fight against it! The integrator will be long gone when the data starts drifting, but you’ll still be there, explaining why the well test and the MPFM disagree for the next decade.

Never, ever allow a multiphase flowmeter to be installed without a clear, double-blind test showing that it can handle at least 90 % to 95 % of your production envelope. Demand it, witness it, and document it. If you need to hire an SME to back you up, do it, it’s still cheaper than the therapy you’ll need if you don’t.

And remember: “One shouldn’t let a flowmeter be bundled without independent SME validation.”

Because procurement shortcuts don’t disappear, they just mature into long-term allocation disputes.

And if all else fails? Well, at least update your CV; you’ll be explaining this decision for the next five years.

📜 COULD HAVE BEEN – The Tendering Phase

Now comes the infamous ‘black box’ – tendering. Engineers prepare the technical basis for the bid; procurement sends the RFP, and commercial competition begins. Often, the lowest bidder wins. And here, millions are lost for a few thousand dollars saved.

Example:

Field: 50,000 BOPD × $50/bbl × 365 × 20 years = $18.25 billion.

With 1% bias → $182 million error.

Independent flow loop test: $150,000, This is 8 ppm! (aka 8 10^-6).

Skipping it: It is a false economy.

ISO/IEC Guide 98-3 (GUM) shows that uncertainty grows multiplicatively if systematic errors are not isolated. A test skipped today compounds over decades of allocation.

Technical Insight: Apply CoPM (Cost of Poor Measurement) method. A small increase in CAPEX to ensure metrological assurance drastically reduces OPEX disputes, auditing costs, and reputation loss.

“Accuracy is not expensive. Inaccuracy is.”

⚠️ SHOULD NEVER HAVE BEEN :The Production Fallout

AKA: When physics sends the invoice.

This is the phase no one wants to talk about: the production fallout, and the moment when the honeymoon is over and reality shows up with a spreadsheet.

The project has been executed, the flowmeters installed, and the wells have produced for years. Everything seemed smooth, until one day, the numbers stopped making sense. Predictions drift, well tests disagree, parameters start jumping around like caffeine-fed electrons. What once looked like seamless production now looks suspiciously like a data circus.

And then, the fun begins!

The teams that bought and sold the flowmeters? They’ve all moved on. The veterans who knew the history are now in different departments or different companies. What’s left is a new production team, inheriting old uncertainty, staring at data they didn’t generate, from systems they didn’t configure, and may not have been trained too.

On the vendor side, there’s a new support team too; full of enthusiasm but armed mostly with inherited headaches. Both sides try to fix the issues, both want to save face, and so begins the great cycle of “parameter tweaking.” Filters are adjusted, coefficients are changed, and the firmware is “updated” without full documentation. Everyone tries to patch the problem. Instead, they just move the bias around.

Tensions rise. Emails get sharper. The flowmeter is blamed, then the flow line, then the separator, then the laws of physics themselves. Eventually, everyone agrees that “something must have changed in the process”.

Meanwhile, the data still refuses to cooperate.

A few years down the road, data mismatches surface everywhere. The separator no longer agrees with the MPFM readings, flow rates fluctuate for no apparent reason, models diverge, and production accountants start quietly panicking.

Common root causes:

• Poor installation geometry (swirl, asymmetry, or residual debris)
• Drift in salinity or gas density compensation
• Transient flow effects with slug frequency exceeding the sensor refresh rate
• Firmware updates or re-averaging logic applied post-commissioning
• Data synchronization issues between MPFM, DCS, and SCADA historians

Technical Insight: Advanced reconciliation tools: Kalman filtering, Mahalanobis outlier detection, FFT-based noise isolation, are often deployed at this stage. But let’s be clear: these are bandages, not cures. Once a systematic bias embeds itself in the allocation chain, it becomes self-reinforcing and financially irreversible.

This is typically when PandA gets called in; not to assign blame, but to quantify and contain the damage. We re-establish traceability, rebuild calibration logic, and reconstruct uncertainty budgets to bring sanity back into the dataset. But even then, one truth remains unshakable:

Again: Remediation is always more expensive than prevention.

That’s why we say, sometimes with a touch of irony: “The teams who make the purchase are rarely the teams that pay the price”. And the rest of us? We’re left reconciling both the data and the history, one noisy dataset at a time. This is the exact situation where our services have been called.

 

🧠 CASE STUDY – Client X: The Smart Way

Client X reached out early in the shall be phase.

Every once in a while, a project comes along that restores your faith in process discipline. This is the story of Client X ; one of those rare operators who called PandA early, right at the “shall be” phase. At the technical requirement stage, Client X asked us to help select the best multiphase flowmeter using what we call the four golden parameters:

Accuracy + Field fit + Vendor reliability + Cost balance.

PandA started with the fundamentals: a detailed study of field constraints, fluid properties, and target accuracy. (As we often say, not every flowmeter needs to be a Ferrari, sometimes what you need is a reliable Land Cruiser that actually finishes the journey).

Using a blend of generated data and model-based simulations, our experts recreated the field’s conditions to explore the limits of technologies. We combined methods such as uncertainty propagation, PVT sensitivity analysis, andMonte Carlo simulations to predict performance across the full production envelope. Synthetic datasets representing twenty years of operation were tested against multiple vendor curves.

To ensure fairness, we introduced quantitative metrics:

• Envelope Overlap Index (EOI): percentage of the field’s operating range covered by the vendor’s certified envelope.
• Sensitivity coefficients (∂error/∂GVF, ∂error/∂WLR): how slight changes in gas or water fraction affect output flow-rate measurement error.
• Uncertainty propagation (RSS): total combined uncertainty on phase fractions.
• Vendor maturity index: experience, calibration recency, service responsiveness, local presence, and global track record…

The outcome? A ranked shortlist of technologies, each with quantified uncertainty and confidence levels. No opinions, no glossy sales curves, just traceable metrology based on the field expectation conditions.

PandA then facilitated contact between Client X and local vendor teams, ensuring that long-term maintenance, support, and installation location were optimized for each candidate flowmeter. We also provided an impartial internal report detailing every device’s strengths, constraints, and realistic field performance expectations.

Armed with this report, Client X shared the findings with procurement, who now fully aware of the technical implications, felt compelled to follow the recommendations or at least discard options that didn’t meet the defined criteria. The result was a cost-versus-quality equilibrium, supported by objective data rather than corporate inertia.

Also, isn’t it fun to come with a detailed report that sends the ball of accountability their way? Yep, perhaps the best part of it!

Procurement, for once, followed the matrix , not the politics. The chosen flowmeter delivered performance within ± 1.5 % total uncertainty, fully traceable, auditable, and supported. In other words:

Flowmeter selection became a science, not a gamble.

Now, how often does that happen? Not much. Since PandA’s creation, we’ve had two true Client X cases.

Did both get exactly what they wanted? One of them did.

But did both end up prepared and armed for every eventuality? Absolutely!

The rest of the time, we’re called at the “should-never-have-been” stage, after the installation, when everyone suddenly realizes the laws of physics didn’t read the sales pitch.

At PandA, we often say our biggest competitor isn’t another consultancy, it’s the client who believes they can handle complex flow metering projects perfectly on their own. Years ago, one of our senior experts put it bluntly (and, admittedly, not everyone appreciated the honesty): “Multiphase flow metering is still an art, a combination of nuclear, electrical, optical, and thermodynamic technologies fused with PVT analysis, signal processing, modelling, verification, and validation. Only those who’ve spent years, if not decades, in this domain can truly appreciate its beauty and its traps”.

Most of our assignments arrive when politics has already tangled with physics, when no one wants to point fingers (or all fingers are pointed everywhere), but everyone wants answers. That’s when PandA steps in as the impartial technical entity: we assess the damage, restore metrological order, and design a better path forward for the next phase.

Client X avoided all that. Client X asked early, listened carefully, and treated flow metering as a science, not a checkbox.

Moral of the story: Be like Client X. Call before the PowerPoint slides become the standard operating procedure.

🧩 CONCLUSION: Physics Over Politics

IE: Clarity and focus are everything in each phase

From “shall be” to “should never have been”, comes with its own traps, shortcuts, and shiny distractions. The key is to remain technically present at every step. When engineers disappear, politics fills the vacuum.

A fully independent, impartial report is not just a bureaucratic checkbox or a nice-to-have spreadsheet. It is the factual matrix that aligns everyone on the same page: what each flowmeter can truly do, what it can’t, and what happens when physics eventually wins the argument. It’s also a mental preparation tool, helping teams confront the real consequences of their technical choices before they become contractual regrets.

An independent review brings balance to the equation. It gives both vendors and operators a fair, transparent ground, where claims meet data, and expectations meet reality. It replaces handshakes and assumptions with traceable facts and test results.

For vendors, it means fair competition: solid technology gets recognized for what it can actually deliver, not for how loudly it’s marketed. For operators, it means confidence that their flow metering choices are scientifically defensible and commercially sound.

At PandA, we like to remind our clients that a clear flow metering report is more than documentation, it’s insurance against future chaos. It doesn’t eliminate uncertainty, but it defines it, controls it, and, most importantly, keeps it honest.

Because at the end of the day, flowmeter selection isn’t about picking a brand; it’s about managing risk, responsibility, and reality.

Accuracy is not bought. Accuracy is engineered.

And the best engineering, as always, starts with clarity, focus, and a little healthy scepticism.

Remember, if you cannot measure correctly, you cannot truly understand your production, and without understanding, every decision becomes a calculated gamble, and after all: ngambling is proscribed for good reason…

Note: If you would like to discuss this article, or consult PandA for your next metering gamble (or measurement device and its uncertainty), do not hesitate to get in touch with us by email: enquiries.panda@pandassociates.consulting, on LinkedIn, or at the contact below!

© 2025 P and Associates Sdn. Bhd. 202001010646 (1366966-D) All Rights Reserved