PileCalc
Changelog

What's new

Fixed
  • Combined pile punching: when the critical sections of a pile group sit exactly tangent to opposite cap edges, the check could keep a contour around the whole group with only a corner arc of resisting perimeter, reporting a demand/capacity ratio several times too high (about 8.7 instead of about 1.2 for a 2×2 group in a 3.6 m cap). Those contours are now treated like every other opposite-edge strip mechanism: excluded from the punching check and reported, with the note that a separate strip or one-way shear check is required.
  • Reopening a saved analysis no longer changes inputs it was saved with. Torsion and foundation stiffness kept a dry soil profile dry instead of adding a water table, foundation stiffness keeps its saved solver increments, segments and load–settlement points, and lateral pile groups keep their saved layering and increments. Groundwater in torsion and stiffness is now an on/off option.
  • Saved analyses no longer show “Unsaved changes” as soon as they open when the saved inputs only spell a default differently.
  • Analyses saved with an earlier release open in the current tool again instead of a read-only snapshot. Their inputs are carried into the current format and re-run on the current engine, with a note saying which engine they were saved with; results can differ, and saving a version records them with the current engine. Inputs saved in US units are converted to metric exactly. Analyses saved before units were recorded ask you to confirm they are metric before opening. Inputs that are now outside a method's validated range are listed so you can fix them, and a record that can't be carried over without an engineering choice stays read-only with a button to start a new analysis.
  • Save to project: choosing “New project…” now sticks. It used to jump back to your first project, so the analysis was saved into an existing project instead of a new one.
New
  • Project soil profiles: describe the ground once per project (for example one profile per boring, with layers, strengths and the water table) and apply it to an analysis's soil layers. Each tool takes only the values the profile states; inputs the profile doesn't carry keep the tool's defaults and are listed for review, and a profile missing a strength the tool needs can't be applied. Lateral p-y tools receive γ′ = γ − γw below the water table, and you pick each layer's p-y model. When someone edits a profile, every analysis using it is flagged on the project page and in the tool, and you review the mapped layers before updating them. Available in axial capacity, capacity vs length, downdrag, drilled shafts, helical piles, torsion, lateral p-y, both pile-group tools and foundation stiffness.
  • Drilled shafts: the unit weight of water is now an editable input next to the water table.
  • Business teams can share projects: everyone on the team can open, edit and save versions of the same analyses, and every change is recorded in the project's audit log. The project's creator or a team admin chooses who can access it, and can make it private again.
  • A new look for the app. Every analysis uses the same layout: Run, Report and Save sit in one toolbar, and API code, Excel, version history and the method docs are in the ⋯ menu. Press ⌘/Ctrl + Enter to run.
  • Tools are grouped by foundation type in the sidebar, your most recently active projects sit under Projects, and a search box finds any project or analysis by name.
  • The projects page lists your own and your team's projects with their analysis types and latest activity, and picks up the analyses you worked on last. Each project lists its analyses with the latest result, save status and who edited them last, next to recent activity.
  • On a phone or tablet, switch between Inputs and Results instead of scrolling past one to reach the other.
Changed
  • PileCalc no longer has a Free plan. New accounts start with a 7-day Pro trial (card required, monthly Pro, one per account), and the app requires an active Pro or Business plan. The no-login calculators stay free.
  • The Pro trial is now 7 days instead of 14. Trials already running keep their original end date.
  • Accounts created before the change get a one-time thank-you: 50% off monthly Pro for the first 6 months, redeemable for 60 days from the in-app notice or Settings → Billing. Saved projects are kept.
Improved
  • Billing is now managed entirely in Settings → Billing: see and replace your card, view and download invoices, pay an unpaid invoice, edit the billing email, name and address on your invoices, switch from monthly to annual with the prorated amount shown before you confirm, or move an annual plan to monthly billing at renewal.
  • You can cancel Pro or Business from the same page. We ask for the main reason so we know what to fix, and a scheduled cancellation can be undone with one click (Keep my plan).
Changed
  • ISO clay now follows ISO 19901-4:2025 as reproduced in the LPILE v2026 Technical Manual and Cathie OPILE: its bearing-factor constants are confirmed by both, and the deep factor 9 + 3α by the NGI/BP companion paper (Zhang et al. 2017). With a gap behind the pile, the ISO strength-anisotropy correction is now applied with ISO's default ratio 0.9 unless you declare your own (up to 10% lower shallow resistance). New optional OCR and plasticity-index inputs select the ISO 2025 default p-y curves.
  • The LPILE Technical Manual, the Todeschini concrete report, ACI 318-19 clause numbering, twelve AASHTO LRFD articles, the Meyerhof & Adams uplift table and the Caltrans SPT charts were checked against primary sources or two independent reproductions; no other numbers changed.
Improved
  • Release provenance audit: every formula, coefficient, default and limit in the engine was checked against its published source and recorded with a page, equation or table reference. Every analysis type now reproduces at least one published worked example (FHWA, NAVFAC, USACE, LPILE, RSPile, Slide, COM624P, PCI and others) in both SI and US units.
  • Results no longer depend on the unit system. Lateral runs could differ by up to 7.8% between SI and US when a node sat exactly on the ground surface or a layer interface; slope, torsion and pile-cap results had smaller differences. All 18 analysis types now agree to better than one part in a billion.
  • Engineering warnings — where a published source leaves a choice open, a code minimum isn't met, or a constant rests on a secondary source — now appear for every analysis in the app (on every plan), in the PDF report and in the MCP result card. Each one appears only when your run uses the affected model or value: axial piles now flag a rock tip shallower than 1.5 diameters, the report's method notes no longer repeat those caveats on every run, Grade 60 reinforcement entered as 413 685 kPa is no longer flagged as below 60 ksi, and footing caveats that depend on your inputs (applied moment, combined passive resistance, consolidation) now appear as warnings on every plan. Choosing a PCI strand grade sets the strand modulus the PCI curve requires, so the request is no longer rejected. The lateral beam-column instability check now compares the axial load with the computed lowest buckling load, so stable piles whose head deflection changes sign under opposing loads are no longer flagged.
  • Slope stabilization reports the mesh-converged pile shear at the slip surface. Prestressed piles use the PCI strand stress–strain curve for Grade 250 and 270 strands. Drilled-shaft uplift includes the bell's weight. Pile-cap punching takes βc from the more conservative of the two FHWA readings. Anchors in intermediate geomaterials use the GEC 4 rock bond-length range (3–10 m).
  • The free pile-capacity calculator runs on its NAVFAC defaults again (it rejected its own pile-weight factor of safety of 1), defaults to FS 3 on side resistance, and limits sand to the 26–40° NAVFAC table range.
  • Axial factors of safety must be at least 2.0 (NAVFAC DM-7.02). Helical sand bearing is capped by the Chance manual's Nq and tension uses the net helix area. Report and documentation formulas, citations and blog examples were corrected to match the engine and the sources.
  • These are engineering consistency changes. They are not a PE signoff.
Fixed
  • Axial end bearing uses the NAVFAC DM-7.02 Nq table as printed (whole numbers; for example 86, not 93.0, for a driven pile at φ = 38°). The earlier values were a smoothed fit.
  • Axial clay adhesion follows the NAVFAC Figure 2 adhesion table, and the sand interface friction angle δ follows the Figure 1 table (steel 20°, concrete and timber ¾φ; δ was φ). A new optional pile material picks the column; if it is left out, the lower value is used. Uplift defaults to NAVFAC's T_all = T_ult/3 + W_p.
  • Rock along a drilled pile limits qu to the concrete strength f′c and requires f′c. Rock along a driven pile is rejected, since FHWA gives no static method for piles driven into rock. The stiffness analysis now rejects rock within the pile length when it validates the request.
  • Drilled shafts: the loose-sand β is bounded after the N/15 reduction, so it no longer drops below 0.25. With a bell, the base exclusion zone uses the bell diameter. Base Su and N are averaged over 2 base diameters below the base, and bell uplift Su over 2 bell diameters above it (FHWA-IF-99-025).
  • The SPT calculator and prefill read the Caltrans Chart 1 friction-angle and Chart 2 moist unit-weight curves. They no longer use density bands or a submerged unit weight that the source does not contain.
  • A lateral run whose axial load exceeds a buckling load of the pile–soil system is now reported as not converged, with a warning. Prestressing strands carry the declared effective prestress at zero load. The t-z toe spring no longer resists uplift.
Changed
  • Source review: each open method question was checked against the published source. Where the source is explicit the engine now follows it; where the source leaves a choice open, the conservative choice is kept and the result carries an engineering warning in the app, the report and the API response.
  • Global slope stability no longer rejects cohesive trial circles whose crest goes into tension. By default it adds a dry tension crack at the depth giving the minimum factor of safety (USACE EM 1110-2-1902 App. C-10.c), so the critical factor of safety no longer depends on the slice count; crestTension: "permit" matches Slide's tension-allowed default.
  • Drilled shafts follow FHWA-IF-99-025: side resistance is dropped from the bell to 2 bell diameters above it in uplift, the bell breakout factor uses depth into the bearing stratum, large bases are reduced when settlement is not computed, sand side resistance is limited to 200 kPa, rock-socket side resistance uses min(qu, f′c) (rock sockets now need concreteStrength), and the top 1.5 m exclusion applies to cohesive soil only.
  • NAVFAC Nq is used only inside its tabulated 26–40° range; a bearing layer outside it is rejected instead of silently using the end value.
  • Footings use the GEC 6 / AASHTO table groundwater factor (conservative against Eq. 5-9), the GEC 6 undrained shape factor, and one-dimensional consolidation alone for clay layers with Cc/Cr (it was also adding elastic settlement). Footing rotation is flagged as a conservative upper estimate.
  • Anchors accept GEC 4 test loads from 1.33 DL, with a warning recommending 1.5 DL in creep-susceptible soil; plate uplift accepts the same water unit-weight range as the rest of the API.
  • Lateral: up-slope sand and silt c-φ now use the RSPile up-slope wedge factors, ISO clay λ uses the mudline strength intercept, silt c-φ under Georgiadis matches LPILE Example 8, layers at or below weak rock are no longer Georgiadis-corrected, and the app picks the ISO-clay soft or stiff backbone from strength.
  • Moment–curvature reports the pre-cracking peak and warns when a section is below the minimum flexural reinforcement intent; Todeschini EI and cracking moment use the declared tension modulus.
  • Torsion applies the CDOT 1.5D surface exclusion to cohesive soil only (matching CDOT's published sand example) and integrates exactly at layer interfaces.
  • Pile caps size flexure and one-way shear from the heavier outer pile row (FHWA NHI-04-041 Step 8.11) and warn when the pile pattern is off the cap centre.
  • These are engineering consistency changes. They are not a PE signoff.
Fixed
  • Liquefied sand (Rollins 2005a) in US units now reads the API's feet and kips. Before, a US request with a pile under 11.8 ft was rejected, and larger piles would have been read as inches and pounds.
  • Drilled-shaft bell uplift now takes Su from the soil the bell sits in. A base placed exactly on a layer interface used the layer below, which could overstate uplift several times over.
  • Georgiadis layering now builds each lower layer's equivalent depth from the middle layers' actual properties; with strength gradients the lower-layer ultimate was up to about 8% high.
  • LRFD load cases now keep the pile-head fixity you chose (fixed or rotational spring) instead of running every case as a free head.
  • Moment–curvature first-yield moment is interpolated on steel strain (it read 0.6–1.1% high), now includes casing and steel-core yield in composite sections, and a plain section at zero thrust now reports its cracking moment instead of 0.
  • Torsion takes the transfer rotation from the same layer as the ultimate resistance at a layer interface.
  • Vertical pile-group settlement no longer fails to converge for square and H-piles.
  • Slope stabilization applies imposed soil movement by depth below ground when a pile stands above grade (core callers only; the API already required a pile head at grade).
  • Requests the engine cannot evaluate are now rejected at validation with a field-level message: capacity-vs-length safety factors of 1 or less, battered piles or rock within the sweep; group-cap SI material limits at the exact engine bounds; Georgiadis layering with elastic, user, liquefied or Piedmont layers; LRFD or sweep moments on restrained heads; and incomplete user p-y and stiff-clay k inputs.
  • Footing results, reports and docs now label bearing FS and design load on the gross basis the engine computes, and a footing whose correct factor of safety is 0 returns that result instead of an error.
  • API reference response units now read for both unit systems, and the docs correct the lateral default increments (100), the weak-rock RQD direction, the above- vs below-water sand k values, and the dry-profile water table.
  • These are engineering consistency fixes. They are not a PE signoff.
Fixed
  • Cyclic RCK stiff clay now uses the FHWA GEC 9 Table A-1 representative cyclic modulus default (200 pci) instead of keeping the static 500 pci default; static stays 500 pci. Cycle count N is used only for Welch–Reese under cyclic loading; fixed-branch models stay invariant to N.
  • This is an engineering consistency fix. It is not a PE signoff.
Improved
  • Hitting the Free 1-project cap now offers a 14-day Pro trial at /app/settings/billing?offer=trial (New project, Save to project, and the projects dashboard). Pro and Business are not nagged.
Improved
  • After a successful Free analysis, the in-app summary now offers a 14-day Pro trial at /app/settings/billing?offer=trial instead of a dead-end “Full results are a Pro feature” overlay. Pro and Business are not nagged.
Fixed
  • Vesić tip stratum now uses the bearing layer below a tip-on-interface, instead of the layer above when the tip sits on a layer boundary.
  • Downdrag no longer forwards NAVFAC zeroZones/negativeZones into t-z skin.
  • Group-vertical docs now match the released driven-sand rule: group capacity is n · Q_single with η = 1, not min(individual, block). The equivalent-block check remains a diagnostic.
  • Slope Q-z uses the same tip-on-interface stratum rule as axial Vesić.
  • Cyclic Reese 1975 / Reese–Cox–Koop stiff clay below water builds the cyclic curve on yp = 4.1 · As · y50, the static A coefficient, as printed by Meyer & Reese (1979, CTR 244-1, p. 32); the cyclic Ac scales peak and residual p only.
  • These are engineering and documentation consistency fixes. They are not a PE signoff.
Changed
  • Pro is $29.99/month ($24.99/month billed annually).
  • Free plan is now 3 in-app analyses per month and 1 saved project, with summary results only.
New
  • One-time Pro trial, 14 days, for eligible new subscribers.
  • Paid MCP and ChatGPT app integration for the same 16 analysis types as the API.
Improved
  • Scientific validation expansion, including public stiffness and buckling benchmarks.
New
  • Lateral pile workflow articles: Excel as a front end to the p-y API, and automation compared with LPILE batch runs and RSPile Python.
Improved
  • Published scientific validation program across the site and registry.
  • Public A/B-coefficient copy now matches the shipped COM624P v2.0 sand A/B and stiff-clay A tables.
Improved
  • UI coverage for every analysis feature, including sectioned and tapered piles, additional head and tip conditions, LRFD combinations, and per-input unit dropdowns.
New
  • Seven more analysis types: capacity vs length, torsion, downdrag, foundation stiffness, pile buckling, helical piles, and pile-cap analysis — 16 tools in total.
  • Show API code and Use in Excel (WEBSERVICE) on every tool, plus a schema-generated API reference.
  • Free SPT correlations calculator (FHWA N₆₀ / (N₁)₆₀ corrections and Caltrans cohesionless ranges).
New
  • Free public calculators (no login): lateral deflection, p-y curves, pile capacity, pile settlement, and group efficiency.
  • Billing moves to Stripe. Business is $99.99/month ($999/year) for teams.
New
  • Browser-generated PDF reports for every analysis type.
New
  • Projects, saved analyses, version history, and audit logs.
New
  • Public REST API on api.pilecalc.com — key-authenticated and plan-gated.
Improved
  • Backend is Clerk + Convex (Convex is the API and database).
  • Per-tool benchmark suite and a public validation page.
New
  • Laterally loaded pile analysis with live deflection, moment, shear and soil-reaction profiles.
  • Axial capacity & settlement with the t-z / q-w load–settlement curve.
  • Info buttons on every input — definition, why it matters, and typical values.
Improved
  • Georgiadis equivalent-depth treatment for layered soil, matching LPILE and RSPile.
  • Validation report published — benchmarks against LPILE, RSPile and closed-form solutions.
New
  • Public launch of the PileCalc engine and API.