PFC Burger Viscoelastic Contact Model

Use this skill for the viscoelastic contact-law stage of pfc5-asphalt-workflow. Burger may represent homogenized asphalt mastic behavior at selected contacts, but it does not by itself create an explicit binder phase or guarantee mixture-scale behavior.

When to use

• assign or audit Burger contacts in a PFC5 asphalt specimen;
• interpret Maxwell and Kelvin spring/dashpot roles;
• fit temperature- and rate-dependent mixture/mastic response;
• build a two-contact or small-RVE verification probe;
• review a Burger command block for non-PFC5 contamination.

Do not use this skill to invent production constants from a generic table.

Required inputs

• PFC2D or PFC3D 5.0 and the unit system;
• object/contact pairs that Burger is intended to represent;
• all Burger properties with units, or raw calibration curves with units;
• test temperature, loading rate/frequency and conditioning history;
• specimen/contact geometry and resolution used during inverse fitting;
• target creep/recovery, dynamic modulus/phase or other macro curves;
• material level and test provenance: binder, mastic or mixture; for JTG 3410-2025 inputs, record the exact DSR/BBR/MSCR/LAS or mixture-performance method rather than writing only “rheology data”;
• contact replacement policy for existing versus future contacts;
• bur_mode tensile-mode choice and the required gap/opening behavior; for any parallel-bond/Burger mixture, also record fraction, selection rule, random seed and spatial-distribution audit;
• any accelerated-time mapping, the preserved constitutive invariants and an unaccelerated pilot used to quantify its error;
• timestep/damping/stop criteria and acceptance tolerances.

Constitutive interpretation

Burger combines a Maxwell branch (spring plus dashpot in series) and a Kelvin branch (spring plus dashpot in parallel). Normal and shear directions have separate parameters. Stiffness-like terms control instantaneous/delayed compliance; dashpot terms control time scale and permanent or delayed deformation. Friction and interface choices remain separate physical assumptions.

Property names and dimensions are syntax-family-specific. Read references/burger-theory.md for theory and references/pfc-command.md for version-labelled syntax notes.

Workflow

1. Freeze units, temperature, rate/frequency, geometry and contact-pair meaning.
2. Verify the PFC5.0 Burger property names with a minimal runtime probe.
3. Run a minimal PFC5 two-body contact probe: contact creation, load/hold/unload, recovery, save/restore and property listing.
4. Install the model using PFC5 cmat semantics, distinguishing future defaults from current-contact changes.
5. Re-equilibrate the specimen and preserve a pre-load save.
6. Freeze Burger bur_mode before calibration. PFC5 documents mode 0 as allowing tensile normal force and mode 1 as not allowing it, while the contact is active only for non-positive gap. If Burger is mixed with a damageable bonded model, freeze the populations and selection rule; record which model carries transient tension, separated-gap cohesion, friction and damage.
7. Classify every curve as binder, mastic or mixture before fitting. JTG T 0627/T 0628, T 0647 and T 0648 binder data are priors or trend constraints, not direct mastic-contact targets. Only a declared test on the same mastic may target a mastic RVE; contact constants still require inverse fitting through the fixed PFC discretization.
8. Treat beam/contact closed-form mappings as dimension-checked initializers, then inverse-fit the fixed PFC discretization; do not report the initializer as calibrated.
9. Validate creep and recovery at the calibration condition, then confirm at another duration/rate/temperature or independent specimen.
10. For time compression, derive the dashpot transform from the declared t -> t' convention, compare with an unaccelerated pilot, and keep the case labelled as a surrogate unless moving-load/path equivalence also passes.
11. Check timestep sensitivity and rerun the target Marshall/rutting pilot.

Working rules

• PFC5 is the only supported target in this package. Unsupported-major-version command blocks are rejected, not retained as migration references.
• Do not use numerical “60 °C reference values” as production defaults; their units, geometry scaling and provenance are not established for the user's case.
• Do not copy a published random Burger-contact fraction into another specimen. Contact topology is a calibrated stochastic variable, not only a material constant.
• Do not describe Burger as categorically “tensionless.” In PFC5, bur_mode=0 can retain tensile normal force and bur_mode=1 suppresses it; however, Burger becomes inactive at positive gap and has no bond-strength/damage law. If another model supplies separated-gap cohesion or damage, expose and validate that division of responsibility under loading, unloading, opening and reversal.
• Burger has no property inheritance, contact methods, or energy partitions in the PFC5 help. Assign all required properties explicitly and mark Burger-specific energy output unavailable rather than inventing a history.
• Never reuse one parameter set at another temperature without an explicit, experimentally supported shift/fitting method.
• Do not mix binder-level JTG 3410-2025 outputs with mixture-level T 0738/T 0745/T 0719 curves as if they had the same geometry or stress measure. Record the scale transition and validate it through the RVE/specimen.
• Do not label a binder DSR/BBR curve as mastic evidence merely because the Burger contact represents homogenized mastic. If no same-material mastic test exists, record the mastic calibration layer as assumption-bound and test its sensitivity.
• Do not delete/recreate all contacts merely as a generic recipe. Choose and verify a version-correct current/future contact transition that preserves the intended state.
• Treat a linear fallback as a diagnostic branch, never as evidence that Burger is calibrated.
• A transform such as t' = t/lambda, c_m'=c_m/lambda, c_k'=c_k/lambda preserves selected linear-Burger creep terms only when stiffness and the load path satisfy the derivation. It does not establish equivalence of wheel motion, rest/recovery, frictional slip, damage, inertia or contact rearrangement.
• Mark commands runtime_verified only after a PFC5.0 runtime accepts them and the probe response passes.

Output contract

• PFC2D/PFC3D 5.0, units, temperature/rate and contact-pair meaning;
• exact version-labelled assignment command and current/future contact policy;
• parameter table with units, bounds, provenance and fitted values;
• bur_mode, opening/gap evidence, contact-model population/seed audit and tension/friction/damage responsibility map;
• two-body/RVE probe histories and timestep sensitivity;
• accelerated-versus-unaccelerated comparison with time-map convention and error metrics;
• calibration/confirmation curves and error metrics;
• pre/post contact-installation save map;
• unresolved runtime or identifiability risks.

Local contents

• references/burger-theory.md — Maxwell-Kelvin mechanics.
• references/pfc-command.md — version-labelled command notes and legacy examples.
• references/calibration.md — guarded macro-response fitting route.
• ../pfc5-asphalt-workflow/references/intersection-rutting-research-evidence.md — research evidence and limits for mixed contacts and accelerated intersection rutting.
• ../pfc5-asphalt-workflow/references/standards-method-map.md — reviewed test-method roles and scale boundaries.
• scripts/burger_contact_probe.p3dat — executable PFC3D 5.0 two-ball relaxation probe; parameters are not calibrated material values.
• agents/openai.yaml — Agent metadata.
• dependencies.json — package-level link to the standards method map.