Well Design Study
A reusable, offset-well-aware workflow for a screening well design study and its report. It turns a pressure basis, a planned (or proposed) casing programme and offset-well evidence into a checked design: window and seats, casing strings, completion screening, drilling feasibility, time and cost range, a standards register and a validation list. Detailed design stays in the operator's well-design and casing-design software; this skill frames the concept, checks a design that already exists in those tools, and prepares the evidence and the report.
When to Use
- Exploration, appraisal or development well at concept or pre-FEED stage needs a casing programme, kick-tolerance check, time and cost range or a design report.
- A planned programme (for example a "dry case" and a "discovery case" design) must be sanity-checked against offset-well strength tests and mud weights.
- A reservoir fluid and well deliverability model needs a tubing, wellhead and barrier basis (see
neqsim-subsea-and-wells,neqsim-near-well-and-injectivity). - Do not use it for final design, rig selection, safety-critical or regulatory work, relief-well or HPHT qualification.
Inputs
- Reference: RKB elevation (air gap), water depth, TD (MD, TVD), vertical or deviated path, target and fluids.
- Pressure basis: pore (low/base/high) and fracture (low/base/high) EMW curves against TVD. Build them from the prospect estimate (reservoir pressure, depth, regional trend) and calibrate to offset wells (
calibration_points,envelope). - Planned programme: sections with hole size, casing OD, top and shoe TVD/MD, mud weight, top of cement, grades to try.
- Kick basis: kick intensity (sg), influx density (sg), required kick tolerance (m3).
- Offset evidence tables (CSV in one folder, filled from public regulator data and well reports or by hand):
offset_casing_lot.csv,offset_mud_weight.csv,offset_pressure_temp.csv,offset_time_depth.csv,offset_problems.csv,offset_documents.csv. - Completion basis (producer): shut-in wellhead pressure from the NeqSim fluid and reservoir pressure, tubing candidates, rates, packer fluid, maximum temperature.
Outputs
- Mud-weight window versus depth with planned mud weights, shoes and offset points (
window_plot). - Kick tolerance per section; bottom-up seat proposal (
propose_seats_bottom_up). - Casing-string selection with burst, collapse, tension and triaxial design factors and the governing load case (
select_string). - Wellhead pressure class, tubing velocity versus erosional and Turner limits (
tubing_table). - Torque and drag envelope and ECD versus flow (
torque_drag_envelope,circulation). - Time and cost P10/P50/P90 with the contributors (
monte_carlo). - Validation checks, standards register and a Word report (
run_well_study,report.WellReport).
Engineering Method
- Evidence first. Identify the country and operator (
neqsim-jurisdiction-regulatory-basis,activity_basis(country, ("well",))): the regime decides the barrier standard, the approval and verification steps, the BOP and well-control rules and the flaring and abandonment requirements, and it goes in the standards register asTO_VERIFYrows. Collect offset data (LOT/FIT at each shoe, mud weight versus depth, pressure tests, kicks, losses, stuck pipe, days per section). Tag each datum with a tier (public regulator data, final well report, internal system, analyst estimate). Empty tables are a result: record them as data gaps. - Pressure model. Pore curves from the prospect estimate and offsets (low/base/high); fracture curves from LOT points (LOT = estimate, FIT = lower bound) with a matrix-stress (Eaton) fallback and an overburden trend. Window = [pore high + trip margin, fracture low - margin].
- Seats and kick tolerance. Single-bubble gas kick with the shoe at the fracture pressure; formation pressure = (mud + kick intensity) hydrostatic at the section bottom. Bottom-up seat proposal, and a section-by-section check of the planned programme.
- Casing design. API TR 5C3 collapse (yield, plastic, transition, elastic), Barlow burst, body yield; load cases gas kick (gas to surface, capped at shoe fracture), tubing leak, pressure test, evacuation; triaxial von Mises with Lame stresses; temperature derating of yield (assumption). Lightest catalogue pipe that passes all design factors.
- Completion screening. Erosional velocity (API RP 14E, C factor), Turner unloading, tubing thermal force and ballooning checks, wellhead class from the maximum surface pressure.
- Drilling feasibility. Soft-string torque and drag with friction ranges, sinusoidal buckling margin, Bingham ECD and pump pressure.
- Time and cost. Section times and NPT events as lognormal P10/P50/P90 from offsets (or planned days), rate and materials uncertainty, Monte Carlo with reported correlations.
- Fail-closed validation. Published-rating cross-check, geometry, window, kick tolerance and design-factor checks; any failed check blocks a ready label.
- Report. Answer first; basis; offsets used and not used; window and seats; casing; completion; drilling; time and cost; risks; assumptions and gaps; standards register; validation; every figure with provenance.
Python Usage Pattern
from well_design_study import (run_well_study, load_evidence, calibration_points, monte_carlo, SectionTime, NptEvent, CostBasis,
normalise_casing_rows, normalise_wellbore_rows, write_table, check_tubing)
rows = normalise_casing_rows(raw_rows, well_info, source="regulator table", date_text="2026-10-31") # fractional inches, none/LOT/FIT, TVD basis
write_table(rows, "casing_lot", "offset_wells")
evidence = load_evidence("offset_wells") # CSV tables, empty list when a file is missing
points = calibration_points(evidence, well_refs, {"rkb_m": 25.0, "water_depth_m": 300.0}) # rebased to equal depth below the seabed
results = run_well_study(spec) # window, plan, strings, checks, standards register
mc = monte_carlo([SectionTime("17.5 in", 6, 9, 16)], [NptEvent("losses", 0.3, 1, 3, 8)], CostBasis(3.2, 140.0, 110.0), common_sigma=0.3)
Study spec options that matter on a real case:
- Sections:
casing_od_in: Nonefor the final open-hole section;riserless: truefor conductor and top-hole (no kick tolerance, window in mud weight below the seabed);shoe_strength_sgfor a planned or measured FIT/LOT;production: truefor the string that carries the shut-in wellhead pressure;fixed_pipeto check an existing design;string_top_tvd_mfor a liner hanger;evacuation_basis(fractionorpore_balance) andevacuation_external(mud_or_poreorset_cement). - Kick:
kick.basisis"pore"(formation pressure = pore + intensity, recommended when mud is well above pore pressure) or"mud"(legacy: mud + intensity, very conservative with heavy mud). Report both when the basis of a planned design is unknown. - Pore curves: build from pressure points with
hydrostatic_pressure_barandemw_curve_from_pressure_points(pressure builds from mean sea level, not from the rotary table). - Tubing:
check_tubing(shut-in and evacuated cases, annulus fluid outside); casing strings usecheck_string/select_string. - Drilling:
torque_and_drag(..., tortuosity_deg_per_30m=0.3)so that a vertical well does not return zero torque. - Time: calibrate
common_sigmaso that the productive-time P10-P90 spread equals the offsets; use only offsets without recorded NPT for productive time and the event frequency for NPT; exclude sidetrack rows (days after kick-off only).
Figures: well_design_study.figures (window_plot with extras, schematic, design_factor_plot, time_depth_plot, histogram_plot(units=...), line_plot). Report: well_design_study.report.WellReport (title page, table of contents, numbered tables and figures, provenance lines, callout, landscape/portrait, appendix, coloured status labels such as [VERIFIED], [DERIVED], [ASSUMED]).
Well plan document and graphics (0.3.0)
A drilling-programme style well plan (cover with document control, summary with key data, chapters, drawings, registers) is built from three modules. All take plain data, use synthetic inputs in the tests and never invent numbers; every figure needs a provenance line and every number a status label.
| Module | Functions |
|---|---|
wellplan_graphics |
A3Sheet (frame, title block, revision table, notes box), well_schematic_sheet (A3 sheet: MD / TVD RKB / TVDSS scales, formation column, nested casing with cement, pressure and mud-weight track, section table coloured by status), bop_stack_sketch, barrier_schematic (primary / secondary / environmental plugs), trajectory_plan_section (plan and section, plan_extent_m), anticollision_sweep_plot, dodge (label spacing) |
wellplan_charts |
gantt_chart (P50 bars, P10-P90 whiskers, calendar axis), risk_matrix (5 x 5, before and after mitigation), decision_tree (contingency tree), s_curve_plot, time_depth_band_plot, cost_breakdown_chart, location_map (rings, scale bar, collision-free labels), offset_comparison_chart |
report.WellPlanReport |
cover with document control (prepared / reviewed / approved left blank), revision history and distribution, contents, list of figures and tables (Word SEQ fields), header and footer with document id, classification and page x of y, key_data_box, table (repeated header, numeric alignment, kept together up to 16 rows), figure (caption, provenance line), landscape(a3=True) / portrait, sheet_page (an A3 landscape page for a drawing), appendix |
Casing load-basis helpers in casing_design: collapse_basis_comparison (mud outside with no cement credit, set cement, unset slurry), max_evacuation_level_for_collapse (fluid-level limit as an operational mitigation), alternative_pipes (all catalogue walls and grades with collapse under a conservative and a realistic basis, drift against the next bit, weight and a relative cost index) and the new StringBasis fields displacement_sg (adds the cementing load case), evacuation_level_tvd_m, evacuation_external="cement_slurry" and kick_mud_sg.
Logic fix in 0.3.0. With kick.basis = "pore" version 0.2.0 replaced the fluid left in the pipe after loss of returns (the planned next-section mud) by the pore EMW, which understated the internal collapse support and failed thin-wall liners that pass with the planned mud. The pore EMW now only sets the formation pressure of the gas kick (kick_mud_sg); legacy_evacuation_fluid: true in the spec reproduces the old numbers.
Example: examples/example_wellplan.py (synthetic). Run python examples/example_wellplan.py OUTPUT_FOLDER.
Professional layout checklist
- Cover: document id, revision, date, status (DRAFT / SCREENING), classification, prepared / reviewed / approved cells left blank for named people, revision history, distribution.
- Answer-first summary of at most one page with a key-data box (well, rig, water depth, RKB, TD, target, spud, days and cost P10/P50/P90, cost class, casing one-liner, pressure range, wellhead class, top risks) and the readiness label.
- Consistent numbering, a caption with number on every figure and table, a provenance line under every figure, units on axes, at least 170 dpi, no overlapping labels (use
dodge). - Drawings on landscape pages (A3 sheet through
sheet_page), wide tables in landscape sections, no orphan headings, no blank pages, tables not overflowing. - Run the report-design QA (
report-designof the repository, or the skillneqsim-report-design-and-visual-qa), render every page, view every page image and record the QA in a log; update fields and export the PDF through Word with the instance stopped afterwards. - Fail closed: any failed extended check (pore-pressure sensitivity, wall tolerance, displacement fluid, planned strength against offsets) keeps the label BLOCKED_BY_FAILED_CHECKS.
Validation Checklist
- 9-5/8 in 47 ppf N80 reproduces the published 4 760 psi collapse and 6 870 psi burst within 2 percent (built-in check).
- Depth references stated (RKB, TVDSS, MSL) and EMW computed against RKB TVD consistently with the offset tables.
- Every planned mud weight lies in the window of its section, or the section is flagged.
- Kick tolerance meets the requirement at every riser-connected shoe.
- Every casing string has a passing catalogue pipe, with the governing load case named.
- Offset data labelled by tier; LOT and FIT not mixed; unverified numbers marked.
- Assumption register and data-gap register filled; no number without a status label.
- Report states what was not modelled and who must confirm each item.
Common Mistakes
| Symptom | Cause | Fix |
|---|---|---|
| Window looks too wide | Offset LOT at a different datum or in a different mud system | Convert all EMW to the same RKB TVD reference; use FIT as a lower bound only |
| Kick tolerance negative | Shoe too shallow for the section pore pressure | Move the seat down, reduce kick intensity only with evidence, or add a liner |
| Casing selection changes with the temperature | Yield derating is an assumption | Use vendor high-temperature derating or HPHT qualification data |
| Burst governed by gas to surface | Shoe fracture cap not applied to the last string | Set cap_gas_at_frac only where the shoe is at a leak-off limit; the production string sees reservoir pressure |
| Time estimate too narrow | Offset NPT not included, or sections treated as independent | Add NPT events from offset problem frequency and a common performance factor (common_sigma) calibrated to the offset spread |
| Kick tolerance far below a planned design | Legacy mud-based kick convention with mud well above pore pressure, or a different hole section assumed per table row | Use kick.basis = "pore", report both conventions, state the reading of the planned design's rows and do not tune to agree |
| Riserless window or kick tolerance fails at the conductor | Window judged as EMW at RKB depth; kick tolerance asked where no shut-in exists | Flag the section riserless; the window is converted to mud weight below the seabed under a seawater head |
| Hydrostatic pore EMW looks too high near the seabed | Normal gradient applied from the rotary table | Build pore curves with hydrostatic_pressure_bar (pressure from mean sea level) |
| Liner collapse fails by a wide margin | Fixed 50 percent evacuation with the mud column outside, or (0.2.0) the pore EMW used as the fluid left in the pipe | Use evacuation_basis = "pore_balance" with the planned next-section mud inside (0.3.0), compare collapse_basis_comparison, check the next bit against the drift with alternative_pipes, and confirm in the company casing tool |
| Thicker wall passes collapse but the next bit no longer fits | Drift (ID less 1/8 in) below the bit | Keep the wall or change the hole programme; grade does not help in the elastic regime (D/t above about 35) |
| Zero torque in a vertical well | Perfectly straight survey | Add a tortuosity allowance (0.2 - 0.5 deg per 30 m) |
| Days per 1000 m of 5 | Sidetrack row (days after kick-off only) mixed with full wells | Exclude sidetracks from time statistics |
| Strength trend scattered | Offsets at different RKB and water depth, FIT mixed with LOT, wells with a different overpressure regime | Rebase to equal depth below the seabed, keep FIT as lower bound, exclude other regimes (state it) |
Limitations
- Screening level: vertical or near-vertical string loads, no wear, no cyclic or thermal-transient well-bore analysis, no wellhead growth or annular pressure build-up, no relief-well design, no cementing hydraulics, no geomechanical stability model.
- The pipe catalogue holds nominal API sizes; confirm against the vendor list. Design factors are operator-practice defaults, not code values. API TR 5C3 collapse is checked on three published ratings only; thin-wall large-diameter pipe (D/t above about 35, elastic regime) is sensitive to ovality and wall tolerance that the formula omits.
- Kick tolerance is a single gas bubble without choke-line friction or gas expansion; conventions differ between tools (pore or mud basis, kick size, pipe size) and must be stated.
- Pore and fracture curves are inputs; the skill does not interpret seismic or logs. Mud weights are never pore pressures. Eaton gradients can lie well below measured leak-off values at depth.
- Time and cost parameters are assumptions unless an offset or rig contract supplies them.
Related NeqSim Functionality
neqsim-integrated-well-planninguses this skill as its drilling and casing design module (tubular formula ratings, pressure window, seats, kick tolerance, torque and drag, time and cost); the integrated skill adds the reservoir-well-facility chain, operating envelopes, barrier schematics, alternatives, hand-off schemas and the combined report.neqsim.process.mechanicaldesign.subsea.WellDesignCalculator,WellMechanicalDesign,WellIntegrityScreeningandWellCostEstimatorfor casing and cost roll-ups inside process models; this skill adds offset calibration, kick tolerance, torque and drag, time/cost ranges and the report.- Skills
neqsim-erd-reach-screening,neqsim-well-trajectory-anticollisionfor reach and collision checks;neqsim-near-well-and-injectivityandneqsim-subsea-and-wellsfor inflow and production wells;neqsim-flow-assurancefor hydrate and CO2 corrosion of the well fluid. - NeqSim gas density at wellhead and shoe conditions gives the influx density and the shut-in wellhead pressure that load the casing.
- Companion community agent:
well-design-study-agent.
References
- API TR 5C3 / ISO 10400, performance properties of casing and tubing. API Spec 5CT / ISO 11960.
- NORSOK D-010, well integrity in drilling and well operations (design basis and barriers; verify the revision).
- Johancsik, Friesen, Dawson (1984), SPE 11380, torque and drag. Dawson and Paslay (1984), SPE 11167, buckling.
- Bourgoyne et al., Applied Drilling Engineering (SPE), hydraulics and well control (kick tolerance).
- Turner, Hubbard, Dukler (1969), SPE 2198, liquid loading. API RP 14E, erosional velocity.