ECSS Structural Analysis — Reduced Model Requirements (space-systems/ecss/reduced-model-requirements)
Use when the task is verifying that a structural finite element reduced
model satisfies the requirements of ECSS-E-ST-32 clause 4.4. This leaf
covers three reduction methods — Craig-Bampton component mode synthesis,
Guyan static condensation, and generic superelements — and checks
interface-node completeness, internal-mode frequency coverage, mass and
frequency accuracy, and the need for residual flexibility correction.
Domain quick reference
- Clause 4.4 addresses reduced models formed by condensing a large
finite element model into a compact representation that preserves
behaviour at a defined set of boundary (interface) nodes. Three
techniques are commonly used: Craig-Bampton, Guyan reduction, and
generic superelements (which may use either technique internally).
Each technique must be verifiable against the full model from which
it was derived.
- Craig-Bampton condensation retains all boundary degrees of freedom
and augments them with a truncated set of fixed-interface normal
modes. The retained mode set must cover the target analysis frequency
range with a prescribed margin (typically 1.5 times the target
frequency); modes above the cutoff contribute residual modal mass and
stiffness that are lost unless a residual flexibility correction is
applied. The method is applicable across a wide frequency range and
is the standard choice for dynamic substructuring.
- Guyan static condensation partitions the model degrees of freedom
into master (retained) and slave (eliminated) sets. The slave
response is recovered by a static relationship to the masters.
Accuracy is adequate only when the target frequency is well below
the natural frequencies associated with the slave degrees of freedom;
when the target approaches those frequencies, dynamic content is lost
and Craig-Bampton should be substituted. Guyan models are often
used for quasi-static or interface-stiffness purposes.
- Interface-node completeness is a prerequisite for all three methods:
every boundary node must be listed with the full set of active
degrees of freedom (coded 1 through 6), and no node may appear more
than once in the interface definition.
- Mass accuracy (fractional error between reduced and full model total
mass) and, for Craig-Bampton, frequency accuracy (fractional error
on the lowest retained frequencies) must both fall within the
allowable limits before the reduced model is accepted.
Workflow
- Identify the reduction method applied to the model (Craig-Bampton,
Guyan, or superelement) and confirm it is one of the three accepted
methods. Reject any unrecognized method before continuing.
- Verify interface-node completeness: list every boundary node, confirm
each carries a non-empty set of active degree-of-freedom codes (1–6),
and confirm that no node ID appears more than once. Flag every
violation as a blocking finding.
- For Craig-Bampton models, check that the highest retained
fixed-interface mode frequency is at least 1.5 times the target
analysis frequency; if it is not, flag the finding and determine
whether residual flexibility correction is applied — its absence
when modes are truncated below the required cutoff is also a
blocking finding.
- For Guyan models, check that the target analysis frequency is
strictly below the lowest natural frequency associated with the
slave degree-of-freedom set; when this condition is not met, flag
the reduction method as inadequate for the intended frequency range.
Also verify that the master degree-of-freedom count is positive and
less than the total degree-of-freedom count.
- For all three methods, compute the fractional mass error between
the reduced model and the full model; flag an exceedance of the
1 % limit. For Craig-Bampton, also compute the fractional frequency
error on the lowest retained modes; flag an exceedance of the 2 %
limit.
- Aggregate all findings per reduced model; the model is not accepted
for delivery or system-level integration until every finding is
resolved.
Pitfalls
- Treating a Craig-Bampton model as compliant when the retained mode
set reaches exactly the target frequency rather than 1.5 times it —
the margin exists to capture the dynamic stiffness contribution of
modes just above the target, which are not negligible.
- Omitting residual flexibility correction when the mode cutoff is
below the required frequency — without this correction the reduced
model is softer than the full model for loads near the cutoff, which
can cause under-prediction of interface forces.
- Applying Guyan reduction at a target frequency that is a significant
fraction of the lowest slave-mode frequency — the static relationship
used to eliminate slave degrees of freedom is no longer accurate and
the reduced model will misrepresent the dynamic stiffness seen at the
master degrees of freedom.
- Accepting an interface-node list with missing degree-of-freedom codes
or duplicate node IDs — incomplete or ambiguous boundary definitions
lead to incorrect constraint mode computation and corrupt the reduced
mass and stiffness matrices.
- Checking mass accuracy alone and skipping frequency accuracy for
Craig-Bampton models — a model may conserve total mass while still
misrepresenting the modal frequencies if the mode shapes are
distorted by an incomplete boundary definition.
Behavior contract (gate 3)
The reduction-method validation, interface-node completeness check,
Craig-Bampton mode-coverage and residual-flexibility logic, Guyan
frequency-range check, and mass/frequency accuracy checks are exercised
by the gate 3 contract test: scripts/test_reduced_model_requirements.py
against scripts/reduced_model_requirements_logic.py (stdlib unittest,
offline). Run:
python3 scripts/test_reduced_model_requirements.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and
paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
1---2name: reduced-model-requirements3description: "Use when verify that a structural finite element reduced model — Craig-Bampton component mode synthesis, Guyan static condensation, or generic superelement — meets ECSS-E-ST-32 clause 4.4 requirements: confirm that all boundary interface nodes and their active degrees of freedom are completely defined, check that the retained internal mode set covers the target analysis frequency with the required margin, validate mass and frequency accuracy against the allowable fractional error limits, determine whether residual flexibility correction must be applied for truncated modal content, and assess overall reduced-model adequacy before delivery or system-level integration. Trigger: ecss, e-st-32-structures-scope, reduced-model, superelement, craig-bampton, guyan, condensation, boundary-dof, residual-flexibility, component-mode-synthesis."4license: Apache-2.05compliance: STANDARDS-REF6standards:7 - id: ecss8 reference-only: true9gated: false10domain: space-systems11pack: space-systems12compatibility: "agentskills.io SKILL.md; any SKILL.md host (Claude Code, Hermes, OpenClaw)"13metadata:14 source: https://skillmd.com/skills/ashfordeou/reduced-model-requirements15 domain: space-systems16 subdomain: ecss17 tags: [ecss, e-st-32-structures-scope, reduced-model, superelement, craig-bampton, guyan, condensation, boundary-dof, residual-flexibility, component-mode-synthesis]18 version: 0.1.019 author: Aero Agent Skills20---2122# ECSS Structural Analysis — Reduced Model Requirements (space-systems/ecss/reduced-model-requirements)2324Use when the task is verifying that a structural finite element reduced25model satisfies the requirements of ECSS-E-ST-32 clause 4.4. This leaf26covers three reduction methods — Craig-Bampton component mode synthesis,27Guyan static condensation, and generic superelements — and checks28interface-node completeness, internal-mode frequency coverage, mass and29frequency accuracy, and the need for residual flexibility correction.3031## Domain quick reference3233- Clause 4.4 addresses reduced models formed by condensing a large34 finite element model into a compact representation that preserves35 behaviour at a defined set of boundary (interface) nodes. Three36 techniques are commonly used: Craig-Bampton, Guyan reduction, and37 generic superelements (which may use either technique internally).38 Each technique must be verifiable against the full model from which39 it was derived.40- Craig-Bampton condensation retains all boundary degrees of freedom41 and augments them with a truncated set of fixed-interface normal42 modes. The retained mode set must cover the target analysis frequency43 range with a prescribed margin (typically 1.5 times the target44 frequency); modes above the cutoff contribute residual modal mass and45 stiffness that are lost unless a residual flexibility correction is46 applied. The method is applicable across a wide frequency range and47 is the standard choice for dynamic substructuring.48- Guyan static condensation partitions the model degrees of freedom49 into master (retained) and slave (eliminated) sets. The slave50 response is recovered by a static relationship to the masters.51 Accuracy is adequate only when the target frequency is well below52 the natural frequencies associated with the slave degrees of freedom;53 when the target approaches those frequencies, dynamic content is lost54 and Craig-Bampton should be substituted. Guyan models are often55 used for quasi-static or interface-stiffness purposes.56- Interface-node completeness is a prerequisite for all three methods:57 every boundary node must be listed with the full set of active58 degrees of freedom (coded 1 through 6), and no node may appear more59 than once in the interface definition.60- Mass accuracy (fractional error between reduced and full model total61 mass) and, for Craig-Bampton, frequency accuracy (fractional error62 on the lowest retained frequencies) must both fall within the63 allowable limits before the reduced model is accepted.6465## Workflow66671. Identify the reduction method applied to the model (Craig-Bampton,68 Guyan, or superelement) and confirm it is one of the three accepted69 methods. Reject any unrecognized method before continuing.702. Verify interface-node completeness: list every boundary node, confirm71 each carries a non-empty set of active degree-of-freedom codes (1–6),72 and confirm that no node ID appears more than once. Flag every73 violation as a blocking finding.743. For Craig-Bampton models, check that the highest retained75 fixed-interface mode frequency is at least 1.5 times the target76 analysis frequency; if it is not, flag the finding and determine77 whether residual flexibility correction is applied — its absence78 when modes are truncated below the required cutoff is also a79 blocking finding.804. For Guyan models, check that the target analysis frequency is81 strictly below the lowest natural frequency associated with the82 slave degree-of-freedom set; when this condition is not met, flag83 the reduction method as inadequate for the intended frequency range.84 Also verify that the master degree-of-freedom count is positive and85 less than the total degree-of-freedom count.865. For all three methods, compute the fractional mass error between87 the reduced model and the full model; flag an exceedance of the88 1 % limit. For Craig-Bampton, also compute the fractional frequency89 error on the lowest retained modes; flag an exceedance of the 2 %90 limit.916. Aggregate all findings per reduced model; the model is not accepted92 for delivery or system-level integration until every finding is93 resolved.9495## Pitfalls9697- Treating a Craig-Bampton model as compliant when the retained mode98 set reaches exactly the target frequency rather than 1.5 times it —99 the margin exists to capture the dynamic stiffness contribution of100 modes just above the target, which are not negligible.101- Omitting residual flexibility correction when the mode cutoff is102 below the required frequency — without this correction the reduced103 model is softer than the full model for loads near the cutoff, which104 can cause under-prediction of interface forces.105- Applying Guyan reduction at a target frequency that is a significant106 fraction of the lowest slave-mode frequency — the static relationship107 used to eliminate slave degrees of freedom is no longer accurate and108 the reduced model will misrepresent the dynamic stiffness seen at the109 master degrees of freedom.110- Accepting an interface-node list with missing degree-of-freedom codes111 or duplicate node IDs — incomplete or ambiguous boundary definitions112 lead to incorrect constraint mode computation and corrupt the reduced113 mass and stiffness matrices.114- Checking mass accuracy alone and skipping frequency accuracy for115 Craig-Bampton models — a model may conserve total mass while still116 misrepresenting the modal frequencies if the mode shapes are117 distorted by an incomplete boundary definition.118119## Behavior contract (gate 3)120121The reduction-method validation, interface-node completeness check,122Craig-Bampton mode-coverage and residual-flexibility logic, Guyan123frequency-range check, and mass/frequency accuracy checks are exercised124by the gate 3 contract test: scripts/test_reduced_model_requirements.py125against scripts/reduced_model_requirements_logic.py (stdlib unittest,126offline). Run:127128python3 scripts/test_reduced_model_requirements.py129130## Compliance131132- ECSS standards are freely downloadable (ESA); cite the source and133 paraphrase per standards-map.yaml.134- compliance: STANDARDS-REF, gated: false.