Skip to content

Materials System Overview

materials-overviewApplies to create / edit / troubleshoot~720 tokens

The desktop AI reads from the same knowledge base — this page is its English edition. When the AI cites 📖 guide · materials-overview in a conversation, it refers to this entry.

In One Sentence

Kavvo’s materials come in two layers: uniaxial (resisting stress in one direction only) and multiaxial (nD, carrying the full stress tensor). Choosing the wrong layer is the number one source of material-related errors—more common than wrong parameter values, because it raises no error.

Where Each Layer Is Used

Layer Stress state Used by Typical members
Uniaxial A single scalar σ-ε curve Fiber sections, truss members, springs, dampers, zero-length elements Beams, columns, braces, rebar, bearings
Multiaxial 3×3 stress and strain tensors Solid elements, shell elements, plane stress/strain elements Shear walls, thick slabs, soil, dams

There is only one way to decide: look at the element type, not the member’s name.

  • Beam elements (elasticBeamColumn / forceBeamColumn / dispBeamColumn) → material attaches to a fiber section → uniaxial only
  • Solid elements (stdBrick / SSPBrick / FourNodeQuad) → material assigned directly to the element → must be multiaxial
  • Shell elements (ShellMITC4 / ShellDKGQ / LayeredShell) → layered shells use uniaxial, thick shells use multiaxial

The most common mistake: modeling a shear wall with solid elements while attaching Concrete02 (uniaxial). The software reports “material type mismatch”, but if the UI lets you force the assignment, the solve results will be completely distorted.

Where Materials Sit in the Modeling Chain

Materials are the first link; everything downstream depends on them (see 📖 guide · model-build-order):

Materials → Sections (fiber sections reference materials) → Geometry → Elements → Boundaries → Loads → Analysis steps

Knock-on effects of changing a material: once a fiber section is created, changing material parameters requires rebuilding the section; if only the material label pointer is changed, the fiber assignments inside the old section persist, and you get “the material changed but the section didn’t.” Kavvo refuses to solve when it detects this inconsistency.

Where Parameters Come From: Three Sources, Different Levels of Trust

Source When to use Reliability
Code values Routine design, no test data Medium—codes give design values, not constitutive parameters
Test values A material test report exists (research, assessment) High—but mind the test loading regime vs. in-situ conditions
Software defaults Concept comparisons, quick trial runs only Low—defaults must never enter a calculation report

Code values must not be used directly as constitutive parameters. This is the easiest trap to fall into:

  • Concrete strengths from codes are design values (already divided by partial factors) or characteristic values (prism/cube), while the constitutive model needs the actual strength at peak stress
  • Rebar yield strength from codes is fy (design value); the constitutive model needs the measured yield strength, typically 10–20% higher than the design value
  • Codes do not provide peak strain, ultimate strain, or softening-branch stiffness—these must come from the constitutive model’s own formulas or from engineering experience

So the complete parameter-selection methods live in the other three guides: steel in 📖 guide · materials-steel, concrete in 📖 guide · materials-concrete, and which to pick in 📖 guide · materials-selection.

Unit System: Where Material Parameters Go Wrong Most Often

Kavvo defaults to mm-kN-s (see 📖 guide · units-and-quantities). Under this system:

Quantity Steel Concrete (C30) Common wrong entry
Elastic modulus E 200 30 200000 / 30000 (MPa-system values)
Strength f 235 / 345 / 420 20.1 (fck) 235000 (kPa)
Peak strain ε ~0.0015 ~0.002 —
Density ρ 7.85e-12 2.5e-12 7850 / 2500

Strain, Poisson’s ratio, and hardening ratio are dimensionless quantities unaffected by the unit system—the only values you can copy over with confidence.

Self-check: after filling in a material, look at the E/f ratio. Steel should be about 850–1000, concrete about 1000–1500. If you get something on the order of 1e6, the units are wrong.

Relationship to OpenSees (Compliance Note)

Kavvo’s solution kernel is OpenSees 3.7.2. Material types correspond at the script level to OpenSees’ uniaxialMaterial / nDMaterial commands, so the type names (Steel02, Concrete04, ElasticIsotropic, etc.) are generic identifiers; we keep them so you can cross-reference literature and existing scripts.

However, this handbook’s parameter meanings, selection methods, and misuse lists were all written independently for Kavvo and do not reproduce the official OpenSees documentation—the OpenSees documentation’s license is a non-standard clause held by the Regents of the University of California (commercial entities: internal use only, no external redistribution or adaptation), and Kavvo, as commercial software, cannot republish it.

For full command syntax (parameter positions, optional arguments), consult the official documentation directly: OpenSees Documentation · uniaxialMaterial · nDMaterial

This handbook tells you “what values to enter and why”; the official documentation tells you “in what order the parameters are written.” They complement each other; they don’t repeat each other.

Five High-Frequency Errors

Error 1: Using a design value as the peak strength

For C30, fcu,k = 30 MPa (cube) and fck = 20.1 MPa (prism). The constitutive model’s peak stress takes the prism value; entering 30 overestimates capacity by about 50%.

Error 2: Leaving the concrete tension branch empty

Concrete01 has no tension branch (zero tensile strength). Use it for a beam and the stiffness drops abruptly after cracking, inflating deflections. Flexural members must use a material with a tension branch (Concrete02 / Concrete04).

Error 3: No hardening branch for steel

Entering b = 0 in Steel01 gives elastic-perfectly-plastic behavior: internal force redistribution in statically indeterminate structures cannot develop, and capacity runs conservative but unrealistic. For common Q235/Q355, use b = 0.01–0.02.

Error 4: Treating confined concrete as unconfined

The column core is confined by stirrups: both strength and ultimate strain increase. Filling the entire section with unconfined concrete underestimates the column’s ductility and capacity. Fiber sections must be zoned: unconfined for the cover, confined for the core.

Error 5: Forgetting to enter density

Static analysis won’t care, but the moment you run a modal or dynamic analysis, mass is zero → infinite frequencies or a singular matrix. Kavvo enforces a density check before modal analysis but not before static analysis—that’s intentional; don’t assume a static solve passing means you’re fine.

Criteria: Three Questions Before Solving

Before solving, ask three questions; if any answer is missing, the parameters aren’t settled:

  1. Does this material’s strength correspond to a design value, a characteristic value, or a measured value? If you can’t say, it isn’t settled
  2. Where did the peak and ultimate strains come from? Computed from a formula, or copied? If copied, you must be able to name the source
  3. Are the units right? Check that the E/f ratio falls within 850–1500

The calculation report prints every parameter of every material together with its source annotation—parameters left blank show as “source not stated” and will be questioned during review. That friction is by design, not a bug.

materialsuniaxialmultiaxialfiber sectionconstitutiveunits