Academic Jobs - Home of Higher Ed Logo

Inflow Performance - an overview, and top 10 academic papers

10 min read0 Opinião

At a glance

A working overview of the inflow performance relationship: how the IPR curve is built, why it bends below the bubble point, and the ten papers that still get cited.

a close up of text on a piece of paper
Photo by Cyril Muhammad on Unsplash

Pull a well test sheet off any production engineer's desk and the first two numbers they read are the rate the well made and the pressure it made that rate against. Pump sizing, tubing selection, whether a workover is worth the money: all of it gets decided downstream of that pair. The relationship between them has a name. The inflow performance relationship, or IPR, is a curve or an equation that gives stabilised production rate as a function of flowing bottomhole pressure at a fixed average reservoir pressure. It is a well's production capacity stated in one line.

Engineers build the curve from a multi-rate production test, or predict it from reservoir data when a test isn't available or isn't affordable. Either way it answers a blunt question: pull the well harder, and what do you get?

A worked example on SPE's PetroWiki makes the shape concrete. An oil well tested at several rates returns a deliverability exponent of 0.743 and a maximum rate near 1,497 stock tank barrels a day. Drop the flowing bottomhole pressure to 800 psia and the same curve says 1,253 barrels a day. That gap between what the reservoir will give at high drawdown and what it gives at low drawdown is the whole argument for building the curve properly.

Why the curve bends

Above the bubble point, the reservoir is holding its gas in solution and the well is moving a single liquid phase. Rate falls in a straight line as flowing pressure rises, and the slope of that line is the productivity index, J, measured in barrels a day per psi of drawdown. For radial flow into a vertical well, J is set by permeability times pay thickness divided by oil viscosity, formation volume factor and a term containing the logarithm of the drainage-to-wellbore radius ratio plus skin.

Below the bubble point the physics changes. Gas comes out of solution in the reservoir, gas saturation builds toward the critical level, and the relative permeability to oil drops as a result. The straight line stops being straight and starts to curve back toward the rate axis. Anyone who models that curve as linear across the whole drawdown range will overestimate deliverability, usually by the widest margin at exactly the pressures where operators want to push hardest.

The best-known fix came from J. V. Vogel in 1968. His dimensionless relationship is qo/qmax = 1 − 0.2(pwf/pr) − 0.8(pwf/pr)2, and it needs a single flow test, one flowing pressure and the average reservoir pressure to generate a full curve. It came out of computer simulation across a wide spread of rock and fluid properties rather than a single field. That generality is why it took hold across the industry so quickly.

Skin sits on top of all this. A positive skin, meaning restricted entry at the wellbore, shifts the curve inward. A negative skin, the kind you get from a stimulated or fractured completion, shifts it outward. Marshall Standing turned that into a usable number in 1970 by defining flow efficiency as the ideal drawdown minus the skin pressure drop, divided by the ideal drawdown, and approximating it from the logarithm of the radius ratio and the skin factor alone.

Close-up of text on a page with lines

Photo by Brett Jordan on Unsplash

The operating point is where two curves meet

An IPR on its own doesn't tell you what a well will produce. It tells you what the reservoir will deliver into the wellbore at a given pressure. What actually reaches the separator depends on the pressure the tubing string imposes at that rate, which is a different curve entirely, governed by fluid density, friction and the wellhead pressure you're producing against.

Nodal analysis puts the two on the same axes. The inflow performance relationship slopes one way, the tubing performance relationship slopes the other, and where they cross is the operating point. If they cross twice, the shallower intersection is usually the unstable one and the well can hunt between states. Production engineers spend a lot of time moving that intersection: changing choke size, gas-lifting, upsizing tubing, or acidising to pull the inflow curve outward.

SPE's PetroWiki entry on oil well performance remains the most-used free working reference for the mechanics, with the test-data plotting steps laid out for both Vogel-type and Fetkovich-type analysis.

The ten papers that still get cited

Any top-ten list in a technical field is a judgement call. This one weighs how heavily each paper is cited by later work and how often the method actually shows up in field practice, rather than novelty alone. Two of the ten are reviews or extensions that exist because the original correlations failed outside their home conditions, which is itself the story of the discipline.

1. Vogel, 1968 — Inflow Performance Relationships for Solution-Gas Drive Wells

Published in Journal of Petroleum Technology, volume 20, pages 83 to 92, as SPE-1476-PA. The paper that made the dimensionless IPR standard practice. Almost everything after it is either an extension, a correction or a test of Vogel's curve. The paper is available through the SPE archive.

2. Standing, 1970 — Inflow Performance Relationships for Damaged Wells Producing by Solution-Gas Drive

Two pages in JPT, volume 22, pages 1399 to 1400, SPE-3237-PA, and one of the highest-leverage short papers in production engineering. Standing introduced flow efficiency so Vogel's curve could be applied to damaged and stimulated wells without re-running a simulator. He also showed how to average several rate tests on a single well.

3. Fetkovich, 1973 — The Isochronal Testing of Oil Wells

Presented as SPE 4529 at the SPE Annual Meeting in Las Vegas, 30 September to 3 October 1973. Fetkovich tested the idea that oil wells below the bubble point behave like gas wells, writing deliverability as rate proportional to the difference of squared pressures raised to an exponent. Across roughly 40 field back-pressure tests, the exponent fell between 0.568 and 1.000, with reservoir permeabilities spanning 6 millidarcies to more than 1,000. The multipoint method it produced has consistently ranked as the most reliable of the classic approaches in comparative studies.

4. Cheng, 1990 — Inflow Performance Relationships for Solution-Gas-Drive Slanted/Horizontal Wells

SPE 20720, presented at the SPE Annual Technical Conference and Exhibition in New Orleans. Using a three-dimensional simulator built at the National Institute for Petroleum and Energy Research, Cheng showed that applying Vogel's vertical-well curve to a deviated well can be off by as much as 22 percent, and by 27 percent for a horizontal well. The practical finding still gets quoted: a slant of at least 45 degrees is needed to add 50 percent on vertical-well productivity, and 60 degrees or more can more than double it.

5. Lekia and Evans, 1990 — Damaged-Well IPR Above the Bubble Point

Earlier damaged-well correlations assumed the reservoir sat at or below bubble point pressure. Lekia and Evans extended the treatment to undersaturated reservoirs, splitting the curve at the bubble point and applying a flow-efficiency correction to each segment. If a well is producing above its bubble point and also has skin damage, this is the version to reach for.

6. Klins and Majcher, 1992 — Inflow Performance Relationships for Damaged or Improved Wells Under Solution-Gas Drive

Published in JPT as SPE-19852-PA. The authors simulated 21 theoretical solution-gas-drive reservoirs with the Weller method, covering a wide spread of relative permeability behaviour and skin, and generated roughly 19,500 data points. Their conclusion was uncomfortable for single-point methods: the shape of the curve depends strongly on bubble point pressure and on how far the reservoir has depleted, so a curve fitted today will drift.

7. Wiggins, 1994 — Generalized Inflow Performance Relationships for Three-Phase Flow

Published in SPE Reservoir Engineering, volume 9, issue 3, as SPE-25458-PA. Water was the gap in most earlier work. Wiggins built regression-based relationships for bounded, homogeneous reservoirs with a mobile water phase present, valid for boundary-dominated flow, and showed they matched other three-phase methods while being far simpler to apply. The record and abstract are on OSTI.

8. Wiggins, Russell and Jennings, 1996 — Analytical Development of Vogel-Type Inflow Performance Relationships

Presented first as SPE 23580 in 1992 and published in SPE Journal, volume 1, pages 355 to 362. For nearly three decades Vogel's equation worked without anyone being able to say precisely why the coefficients were 0.2 and 0.8. This paper derived the form from a Taylor series expansion of the multiphase flow equations and checked it against simulation. It turned a field correlation into something with a physical basis.

9. Aragón, Moya and García-Gutiérrez, 2008 — Inflow Performance Relationships in Geothermal and Petroleum Reservoir Engineering

Published in Geothermics, volume 37, pages 635 to 650. A rare cross-discipline review that lines up the petroleum correlations against geothermal inflow curves for pure water, H2O–CO2 mixtures and saline H2O–CO2–NaCl systems. It concludes that IPRs can pin down a well's maximum flow rate at any point in its producing life, in either industry. The review is published by Elsevier.

10. Tabatabaie and co-authors, 2020 — Generating IPRs for Unconventional Reservoirs from Reservoir Models

The paper that says, plainly, the old correlations don't transfer. In multi-fractured horizontal wells in ultra-low-permeability rock, boundary-dominated radial flow rarely arrives and transient linear flow dominates instead. The authors propose a triple-combo IPR built at three flow durations, such as 30, 90 and 180 days, replacing the single curve tied to average reservoir pressure. They also document a result that unsettles intuition: under some conditions, lowering the flowing pressure reduces the rate, because drawdown, compressibility and relative permeability pull against each other.

Close-up of a book page with printed text

Photo by Brett Jordan on Unsplash

What the industry is doing about the gap

The obvious response to correlations that stop working is to stop relying on correlations. Recent work couples the IPR with a dynamic material balance and trains a model on the output. One 2024 study in Energies combined data from three offshore gas fields, built a vertical flow performance table in a commercial simulator, and trained a multilayer perceptron with regression on top of it. Predictions of bottomhole pressure came within an average error of 1.5 percent against the simulator. The paper is open access in Energies.

That is a useful result, and it is not a reason to throw out Vogel. A model trained on simulator output inherits the assumptions of the simulator, and it will not tell you when it has stepped outside the range it was trained on. Correlations are transparent about their assumptions even when those assumptions are wrong for your well.

The failure mode to watch for is a familiar one. Someone opens a nodal analysis package, selects the default inflow method, and generates a forecast for a fractured horizontal well in a shale play. The curve looks plausible. It is built on a relationship developed for vertical wells in conventional solution-gas-drive reservoirs. The error does not announce itself.

The practical takeaway

Match the correlation to the flow regime you actually have, not the one the software defaults to. Vertical well, conventional reservoir, pressure below bubble point: Vogel, with Standing's flow efficiency if there's skin. Deviated or horizontal: Cheng's form, or a model. Three-phase with mobile water: Wiggins. Unconventional, transient: a reservoir model, not a curve.

Then get one real test point and overlay it. A correlation that misses a measured rate by 30 percent is telling you something about your reservoir that no amount of curve-fitting will. Build the curve, test it, and let the field correct the paper.

  • Inflow Performance Relationship
  • IPR Curve
  • Petroleum Engineering
  • Reservoir Engineering
  • Well Deliverability
Retrato do Jarrod Kanizay
Sobre o autor

Jarrod KanizayVeja o autor

Academic Jobs In House Author

Visionary leader transforming academic recruitment with 20+ years in higher education.

Discussão

De sorte em:

Seja o primeiro a comentar este artigo!

Você

Você será solicitado a entrar antes que seu comentário seja postado.

Browse por Faculdade

Browse por assunto

Frequently Asked Questions

📄What is an inflow performance relationship (IPR)?

An inflow performance relationship is a curve or equation that expresses a well's stabilised production rate as a function of flowing bottomhole pressure, at a fixed average reservoir pressure. It describes what the reservoir will deliver into the wellbore, not what the surface facilities will take. The curve is normally plotted with rate on one axis and flowing pressure on the other, and the point where it meets the tubing performance curve sets the well's actual operating point.

📄Why is the IPR curve a straight line for some wells and curved for others?

Above the bubble point the reservoir holds its gas in solution and the well produces a single liquid phase, so rate falls linearly as flowing pressure rises. The slope of that line is the productivity index. Once flowing pressure drops below the bubble point, free gas appears in the reservoir, gas saturation rises toward the critical level, and the relative permeability to oil falls. The straight line then bends toward the rate axis.

📄What is the Vogel equation and why is it still used?

Vogel's relationship, published in the Journal of Petroleum Technology in 1968, is qo/qmax = 1 − 0.2(pwf/pr) − 0.8(pwf/pr)2. It generates a full inflow curve from one flow test, one flowing pressure and the average reservoir pressure, with no need for permeability, pay thickness or relative permeability data. It was built from simulation across a wide range of rock and fluid properties, which is why it transfers reasonably well between conventional solution-gas-drive reservoirs.

📄What does skin factor do to an IPR curve?

Skin represents an extra pressure loss or gain at the wellbore face. Positive skin, from mud damage or a restricted perforation tunnel, pulls the curve inward so the well makes less at every flowing pressure. Negative skin, from a fracture or a successful stimulation, pushes the curve outward. Standing's flow efficiency ratio lets you correct a Vogel-type curve for skin without running a new simulation, using only the drainage-to-wellbore radius ratio and the skin value.

📄How does Fetkovich's method differ from Vogel's?

Fetkovich treated oil wells below the bubble point the way gas wells had long been treated, writing deliverability as rate proportional to the difference of squared pressures raised to a deliverability exponent. That requires multiple test points rather than one, plotted on log-log axes, with the exponent taken as the inverse of the slope. The payoff is accuracy: in comparative studies the multipoint Fetkovich approach has generally outperformed single-point methods across the pressure drawdown range.

📄Can I use Vogel's curve for a horizontal or deviated well?

Not without a correction. Cheng's 1990 work using a three-dimensional slanted and horizontal well simulator found that IPR data generated with the Vogel curve can differ by as much as 22 percent for slanted wells and 27 percent for horizontal wells. The same study found that a slant angle of at least 45 degrees is needed to increase productivity by 50 percent over a vertical well, and 60 degrees or more can more than double it.

📄What is the operating point in nodal analysis?

The operating point is where the inflow performance relationship crosses the tubing performance relationship. The IPR says what the reservoir will deliver at a given bottomhole pressure; the tubing curve says what pressure the wellbore and surface设施 impose at that rate. Where they intersect is what the well actually produces. When the two curves cross more than once, the lower-rate intersection is often unstable and the well can oscillate between flowing states.

📄Why do conventional IPR correlations fail in unconventional reservoirs?

Unconventional wells are typically multi-fractured horizontals in rock of extremely low permeability, where boundary-dominated flow is rarely reached and transient linear flow dominates instead. The classic correlations assume boundary-dominated radial flow and a curve tied to average reservoir pressure. Work presented in 2020 proposed instead a triple-combo IPR built at three flow durations, such as 30, 90 and 180 days, and reported cases where lowering the flowing pressure actually reduces the rate because drawdown, compressibility and relative permeability pull against each other.

📄Does the IPR change as a reservoir depletes?

Yes, and that is one of the more inconvenient findings in the literature. Klins and Majcher simulated 21 solution-gas-drive reservoirs and generated roughly 19,500 data points, concluding that the shape of the curve depends strongly on bubble point pressure and on depletion. Fetkovich addressed the same problem by assuming the deliverability exponent stays constant while the maximum rate declines with reservoir pressure, which works well enough in many fields.

📄Is machine learning replacing IPR correlations?

It is supplementing them, mostly in forecasting rather than in day-to-day well analysis. A 2024 study in Energies combined a dynamic material balance with an inflow performance relationship for three offshore gas fields, built a vertical flow performance table in a commercial simulator, and trained a multilayer perceptron with regression on top of it. Predicted bottomhole pressures came within an average error of 1.5 percent of the simulator. The catch is that such a model inherits the assumptions of whatever generated its training data.

📄What is the single most important check before trusting an IPR curve?

Overlay one measured test point on the curve. A correlation that misses a real rate by a wide margin is telling you something about the reservoir that no amount of curve-fitting will fix. Match the method to the flow regime and the fluid first, apply skin and flow efficiency second, then validate against field data before the curve gets used to size a pump or justify a workover.