Converge
EN
Create account

MODULE GUIDES

Well test

Pressure transient analysis. Flow periods, Bourdet derivative, Horner and MDH, deconvolution, analytical models and regression with confidence intervals.

Last edited

Overview

The Well test module interprets pressure transient data. You paste a gauge record, check the flow periods, read the Bourdet derivative on the log-log plot, run Horner or MDH straight lines, then match an analytical model by regression with confidence intervals. Use it for build-ups and drawdowns on oil wells, and on gas wells through normalised pseudo-pressure.

The page has seven tabs: Data & periods, Log-log diagnostic, Horner / MDH, Deconvolution, Model, Regression and Report. You can exclude any point by clicking it.

Inputs

Gauge data

Input Unit Notes
Time hr Choose Time unit: s, min, hr or day. ISO date-times and hh:mm:ss are read as hours.
Pressure psia Bottomhole gauge pressure.
Rate STB/d (Mscf/d for gas) Optional. Without a rate column, the module detects flow periods from the pressure trend.
Initial pressure pi psia / bara Reference for the first drawdown. Can be fitted in regression.

Paste with any delimiter, with or without a header. The module guesses the Time column, Pressure column and Rate column from the header names; change them if it guesses wrong. A record needs at least five usable samples.

Rate history

Each step has a start time (hr) and a surface rate. 0 means shut-in. The table is built from the rate column if there is one, otherwise from the detected pressure trend reversals. You can edit it freely.

Rock and fluid

Input Unit (Field) Notes
Net thickness h ft
Porosity φ fraction
Wellbore radius rw ft
Total compressibility ct 1/psi
Oil FVF B rb/STB
Viscosity μ cp

B, μ and ct follow the project fluid at the initial pressure. Use project fluid at pi resets them. With no project fluid, typical values carry a badge and the page warns you. Tick Gas well (pseudo-pressure) to analyse on normalised pseudo-pressure with rates in Mscf/d.

Methods

  • Bourdet derivative (Bourdet et al. 1983, 1989), with Derivative smoothing L in log-time units (default 0.1, typical 0.1 to 0.5).
  • Automatic flow-regime identification on the derivative: wellbore storage, linear, bilinear, radial, dual-porosity valley, boundary and pseudo-steady state, each with time range, slope and confidence.
  • Horner (1951) and Miller-Dyes-Hutchinson (1950) semilog analysis: k from 162.6 qBμ / (m h), skin, p* and p at 1 hr.
  • Gas wells: normalised pseudo-pressure after Al-Hussainy, Ramey & Crawford (1966).
  • Deconvolution: von Schroeter, Hollaender & Gringarten (2004), with the initial pressure fitted after Levitan (2005). The smoothing is set so the fit matches the gauge noise.
  • Analytical models in Laplace space, inverted numerically:
    • Well: Vertical, storage + skin; Infinite-conductivity fracture, Finite-conductivity fracture and Uniform-flux fracture (Gringarten, Ramey & Raghavan 1974; Cinco-Ley, Samaniego & Dominguez 1978); Horizontal well (Ozkan & Raghavan 1991); Partial penetration (limited entry).
    • Reservoir: Homogeneous; Dual porosity, pseudo-steady (Warren-Root); Dual porosity, transient slabs (de Swaan 1976); Radial composite (two zones) (Satman et al. 1980).
    • Boundary: Infinite acting, Single sealing fault, Single constant-pressure boundary, Channel (two parallel faults), Closed rectangle, Closed circle, Constant-pressure circle. Linear boundaries use the method of images.
    • Wellbore storage: Constant storage or Changing storage (phase redistribution) (Fair 1981).
  • Regression: Levenberg-Marquardt on the measured pressures, with parameters in log space. Confidence intervals come from the linearised covariance (Rosa & Horne 1983).

Workflow

  1. Open Well test and use Paste gauge data, or paste on the Data & periods tab and press Load. Check the column choices and Time unit.
  2. Check the Rate history and Flow periods. Type the actual rates if the module detected the periods from pressure alone.
  3. Open Log-log diagnostic. Adjust Derivative smoothing L, read the Flow regimes identified, in time order table and the Quick estimates from regimes (k, C and skin).
  4. Open Horner / MDH. Set the Straight-line window (From Δt and To Δt) on the radial-flow stabilisation and read k, skin and p*.
  5. If the test has two or more rate periods, open Deconvolution and press Deconvolve. Use it to see boundaries no single build-up reaches.
  6. Open Model. Choose Well, Reservoir, Boundary and Wellbore storage, and set the parameters until the model overlay sits close to the data.
  7. Open Regression. Tick the Parameters to fit, set bounds, optionally tick Fit initial pressure, then press Run regression. Press Accept to keep the result.
  8. Open Report to check results, assumptions, warnings and the action log, then press Export report.

Interpreting results

  • A unit slope on the derivative is wellbore storage; a flat derivative is infinite-acting radial flow; a half slope is linear flow; a quarter slope is bilinear flow. A late rise usually means a sealing boundary; a late fall, a constant-pressure boundary or dual-porosity behaviour.
  • On the Regression tab, read RMS residual and Converged, then the Rel. std error of each parameter. Check Parameter correlation: strongly correlated parameters mean the intervals understate the real uncertainty, because they are linearised.
  • Horner, log-log and regression should agree. If they do not, check the straight-line window and the radial-flow level on the derivative.
  • On Deconvolution, RMS misfit should be close to the gauge noise shown in the footer.

Common pitfalls

Exports

  • Export report on the Report tab downloads a text file (welltest-report.txt). It holds the on-screen Horner, log-log and latest regression results, accepted or not, with 95 % intervals, plus assumptions, warnings and the action log. The report and the screen use the same analysis, so the numbers match.
  • The page saves a version before you load a new gauge record.