ECSS Structures — Test Prediction Report (space-systems/ecss/drd-test-prediction)
Use when the task is producing or reviewing the test prediction report (TP DRD)
for a structural test campaign under ECSS-E-ST-32C Annex Q — computing expected
structural responses at each monitoring point for every test load case, setting
safety-abort thresholds from finite-element predictions, checking that predicted
peak values are within instrument measurement capability, and defining the
analysis-to-test correlation acceptance band.
Domain quick reference
- Annex Q defines the minimum content for a test prediction report (TP).
The report must address every test case in the campaign (static proof,
modal survey, random vibration, acoustic, shock, thermal cycling) and
supply a predicted response at each monitoring point before the test
begins.
- Predicted responses are derived from a linear finite-element model (e.g.,
Calculix linear solver) at a unit load factor and then scaled to the
actual test load factor. A test load factor of 1.0 represents the
qualification or acceptance load level; factors above 1.0 are used for
over-test margins, but must not exceed 2.0 under the E-ST-32C structural
qualification envelope.
- A safety-abort threshold is set at a fixed margin above the predicted
peak response (typically 20 % above peak). During the test, if a
monitoring channel reaches its abort threshold the test must be halted.
The margin factor must be greater than 1.0; values at or below 1.0 are
not physically meaningful.
- Each monitoring point must be served by an instrument whose full-scale
measurement range covers the predicted peak response. A prediction that
exceeds the instrument range is a pre-test finding: either the instrument
must be replaced with a higher-range sensor or the load level must be
reduced.
- Correlation acceptance is the quantitative criterion against which
measured test responses will later be compared in the test evaluation
report (TE DRD). Each predicted quantity carries a band expressed as a
percentage of the predicted value. A measured response that falls outside
the band is an anomaly requiring formal disposition.
Workflow
- Collect the test specification (objectives, test cases, load factors,
directions) and confirm that every load factor is in the range
(0, 2.0] and that every load direction is a recognised structural axis
or load type.
- For each test case, retrieve the baseline finite-element predictions at
unit load factor for every monitoring point and quantity (displacement,
force, stress, natural frequency). Scale each baseline value by the
test load factor to obtain the predicted peak response.
- Compute the safety-abort threshold for each monitoring point in each
test case: multiply the absolute predicted peak by the abort margin
(default 1.20). Record the abort threshold in the report; it becomes
an instrument channel limit setting before the test begins.
- Check that the predicted peak response at each monitoring point is
within the instrument's full-scale measurement range. Flag any
monitoring-point/quantity pair whose predicted value exceeds the
instrument range as a pre-test finding.
- Define the correlation acceptance band for each quantity at each
monitoring point (default ±10 % of the predicted value). Record the
lower and upper bounds; these are the acceptance limits used in the
subsequent test evaluation.
- Aggregate all findings (invalid test cases, instrument overrange,
missing instrument definitions) and confirm that no open findings
remain before the test prediction report is approved.
Pitfalls
- Applying the abort margin to the signed predicted value instead of its
absolute magnitude — a negative displacement prediction gives a
negative abort threshold that can never be reached, leaving the test
without a meaningful halt criterion.
- Using a load factor of zero — a zero factor produces zero predictions
and meaningless abort thresholds; it is not a valid test case and
must be rejected before entering the prediction calculation.
- Omitting instrument-range checks and discovering the overrange
condition only during the test, which forces a test abort and may
damage the test article or its instrumentation.
- Setting the correlation band to 0 % or 100 % — a zero-width band
makes the criterion impossible to meet in practice; a 100 % band
accepts any measurement and provides no engineering assurance.
- Confusing the test prediction with the test evaluation: the prediction
is written before the test using analysis outputs only; measured data
does not appear until the evaluation phase covered by the TE DRD.
Behavior contract (gate 3)
The test-case validation, linear FEM scaling, abort-threshold computation,
instrument-range check, and correlation-band logic are exercised by the
gate 3 contract test: scripts/test_drd_test_prediction.py against
scripts/drd_test_prediction_logic.py (stdlib unittest, offline). Run:
python3 scripts/test_drd_test_prediction.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and
paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
1---2name: drd-test-prediction3description: "Use when prepare a structural test prediction report per ECSS-E-ST-32C Annex Q: determine predicted response quantities (displacement, load, stress, natural frequency) at each monitoring point for every test load case, derive safety-abort thresholds from scaled finite-element analysis outputs, verify that predicted peak responses fall within instrument measurement ranges, and establish the analysis-to-test correlation acceptance band for each monitored quantity. Trigger: ecss, e-st-32-structures-scope, test-prediction, drd, structural-test, monitoring-point, abort-threshold, test-correlation, calculix-linear, annex-q."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/drd-test-prediction15 domain: space-systems16 subdomain: ecss17 tags: [ecss, e-st-32-structures-scope, test-prediction, drd, structural-test, monitoring-point, abort-threshold, test-correlation, calculix-linear, annex-q]18 version: 0.1.019 author: Aero Agent Skills20---2122# ECSS Structures — Test Prediction Report (space-systems/ecss/drd-test-prediction)2324Use when the task is producing or reviewing the test prediction report (TP DRD)25for a structural test campaign under ECSS-E-ST-32C Annex Q — computing expected26structural responses at each monitoring point for every test load case, setting27safety-abort thresholds from finite-element predictions, checking that predicted28peak values are within instrument measurement capability, and defining the29analysis-to-test correlation acceptance band.3031## Domain quick reference3233- Annex Q defines the minimum content for a test prediction report (TP).34 The report must address every test case in the campaign (static proof,35 modal survey, random vibration, acoustic, shock, thermal cycling) and36 supply a predicted response at each monitoring point before the test37 begins.38- Predicted responses are derived from a linear finite-element model (e.g.,39 Calculix linear solver) at a unit load factor and then scaled to the40 actual test load factor. A test load factor of 1.0 represents the41 qualification or acceptance load level; factors above 1.0 are used for42 over-test margins, but must not exceed 2.0 under the E-ST-32C structural43 qualification envelope.44- A safety-abort threshold is set at a fixed margin above the predicted45 peak response (typically 20 % above peak). During the test, if a46 monitoring channel reaches its abort threshold the test must be halted.47 The margin factor must be greater than 1.0; values at or below 1.0 are48 not physically meaningful.49- Each monitoring point must be served by an instrument whose full-scale50 measurement range covers the predicted peak response. A prediction that51 exceeds the instrument range is a pre-test finding: either the instrument52 must be replaced with a higher-range sensor or the load level must be53 reduced.54- Correlation acceptance is the quantitative criterion against which55 measured test responses will later be compared in the test evaluation56 report (TE DRD). Each predicted quantity carries a band expressed as a57 percentage of the predicted value. A measured response that falls outside58 the band is an anomaly requiring formal disposition.5960## Workflow61621. Collect the test specification (objectives, test cases, load factors,63 directions) and confirm that every load factor is in the range64 (0, 2.0] and that every load direction is a recognised structural axis65 or load type.662. For each test case, retrieve the baseline finite-element predictions at67 unit load factor for every monitoring point and quantity (displacement,68 force, stress, natural frequency). Scale each baseline value by the69 test load factor to obtain the predicted peak response.703. Compute the safety-abort threshold for each monitoring point in each71 test case: multiply the absolute predicted peak by the abort margin72 (default 1.20). Record the abort threshold in the report; it becomes73 an instrument channel limit setting before the test begins.744. Check that the predicted peak response at each monitoring point is75 within the instrument's full-scale measurement range. Flag any76 monitoring-point/quantity pair whose predicted value exceeds the77 instrument range as a pre-test finding.785. Define the correlation acceptance band for each quantity at each79 monitoring point (default ±10 % of the predicted value). Record the80 lower and upper bounds; these are the acceptance limits used in the81 subsequent test evaluation.826. Aggregate all findings (invalid test cases, instrument overrange,83 missing instrument definitions) and confirm that no open findings84 remain before the test prediction report is approved.8586## Pitfalls8788- Applying the abort margin to the signed predicted value instead of its89 absolute magnitude — a negative displacement prediction gives a90 negative abort threshold that can never be reached, leaving the test91 without a meaningful halt criterion.92- Using a load factor of zero — a zero factor produces zero predictions93 and meaningless abort thresholds; it is not a valid test case and94 must be rejected before entering the prediction calculation.95- Omitting instrument-range checks and discovering the overrange96 condition only during the test, which forces a test abort and may97 damage the test article or its instrumentation.98- Setting the correlation band to 0 % or 100 % — a zero-width band99 makes the criterion impossible to meet in practice; a 100 % band100 accepts any measurement and provides no engineering assurance.101- Confusing the test prediction with the test evaluation: the prediction102 is written before the test using analysis outputs only; measured data103 does not appear until the evaluation phase covered by the TE DRD.104105## Behavior contract (gate 3)106107The test-case validation, linear FEM scaling, abort-threshold computation,108instrument-range check, and correlation-band logic are exercised by the109gate 3 contract test: scripts/test_drd_test_prediction.py against110scripts/drd_test_prediction_logic.py (stdlib unittest, offline). Run:111python3 scripts/test_drd_test_prediction.py112113## Compliance114115- ECSS standards are freely downloadable (ESA); cite the source and116 paraphrase per standards-map.yaml.117- compliance: STANDARDS-REF, gated: false.