{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"invariant_mode_decomposition_design__physics","trajectory_id":"R","attempt_index":0,"archetype_slug":"invariant_mode_decomposition_design","domain_slug":"physics","decision":"CANDIDATE","problem_id":"hidden_growing_thermoacoustic_modes","causal_lever_id":"targeted_damping_of_pressure_heat_release_modes","proposal":{"problem":"A laboratory combustor can transition into damaging thermoacoustic oscillation even while average pressure, temperature, and heat-release measures appear acceptable, because instability develops along coupled pressure–heat-release directions that single-sensor thresholds do not identify.","actors_substrate":["Combustion-physics researchers and facility safety operators","Instrumented laboratory combustor","Pressure, heat-release, temperature, and actuator time series","Fuel-modulation or acoustic actuators"],"observable_state":"Repeated runs show correlated oscillations, emerging spectral peaks, or rising pressure variance before a high-amplitude limit cycle; whether these observations form reproducible growing modes is a HYPOTHESIS.","consequence":"Late detection can permit damaging pressure oscillations, while symptom-level control may suppress a visible peak without damping the coupled instability.","affected_objective":"Predict instability onset early and select bounded actuator settings that reduce growth of the responsible coupled mode without violating combustor safety limits.","structural_mapping":[{"archetype_element":"Transformation acting on coupled variables","domain_realization":"An estimated local state-transition or continuous-time linearized operator maps pressure, heat-release, temperature, and actuator perturbations near a fixed combustor operating point.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Invariant directions with scalar response","domain_realization":"Operator eigenvectors represent approximately persistent coupled perturbation shapes; eigenvalues or their real parts quantify modal growth, decay, and oscillation.","claim_kind":"INFERENCE"},{"archetype_element":"Hidden unstable direction","domain_realization":"A pressure–heat-release mode may grow before any individual measurement crosses its alarm threshold.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Intervention in modal coordinates","domain_realization":"Actuator phase and amplitude are chosen from their measured projection onto the candidate unstable mode rather than from a single pressure peak.","claim_kind":"HYPOTHESIS"},{"archetype_element":"Residual and drift governance","domain_realization":"Prediction residuals, empirical mode-shape mismatch, mode rotation, and spectral-gap loss determine when the modal controller is no longer trusted.","claim_kind":"INFERENCE"}],"component_map":[{"component":"Transformation Scope","status":"direct","domain_realization":"A local perturbation-to-next-state operator at one specified fuel flow, equivalence ratio, geometry, and boundary condition."},{"component":"State-Vector Definition","status":"adapted","domain_realization":"Synchronized pressure, heat-release proxy, temperature, and actuator-history coordinates with declared scaling and sensor locations."},{"component":"Invariant Mode Basis","status":"direct","domain_realization":"Eigenmodes of the local operator, cross-checked against empirically measured response shapes."},{"component":"Modal Gain Spectrum","status":"adapted","domain_realization":"Discrete-time eigenvalue magnitudes or continuous-time eigenvalue real parts and frequencies, with uncertainty."},{"component":"Dominant Mode Selection Rule","status":"direct","domain_realization":"Retain modes exceeding preregistered growth-risk, outcome-sensitivity, or reconstruction-contribution thresholds."},{"component":"Stable/Unstable Mode Partition","status":"direct","domain_realization":"Classify modes as decaying, marginal, growing, or oscillatory within the stated operating window."},{"component":"Modal Intervention Map","status":"direct","domain_realization":"Measured sensitivity from bounded fuel or acoustic actuator perturbations to each retained mode and the safety outcome."},{"component":"Reconstruction Residual Check","status":"direct","domain_realization":"Out-of-sample state and pressure reconstruction error, including tests for structured residuals."},{"component":"Mode Drift Monitor","status":"direct","domain_realization":"Track eigenvalue movement, mode-shape rotation, and ordering across repeated runs."},{"component":"Interpretation Scope Contract","status":"direct","domain_realization":"Modes are local descriptive-control objects, not globally valid flame mechanisms or proof of causation."},{"component":"Mode-Coupling Register","status":"direct","domain_realization":"Record near-degenerate, nonorthogonal, and perturbation-induced cross-effects among retained modes."},{"component":"Local Linearization Window","status":"direct","domain_realization":"Declared bounds on operating point, perturbation amplitude, time horizon, and pre-limit-cycle regime."},{"component":"Spectral Gap Threshold","status":"direct","domain_realization":"Minimum separation and mode-shape identifiability required before treating one mode as a stable control target."}],"mechanism_dispositions":[{"slug":"eigendecomposition_workflow","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Decompose the estimated explicit local combustor operator and report conditioning as well as eigenpairs.","counterfactual_removal":"There would be no transformation-derived modes or growth spectrum to target."},{"slug":"modal_sensitivity_sweep","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Apply bounded actuator perturbations and rank their effects on modal gain and pressure risk while logging cross-effects.","counterfactual_removal":"Mode identification would not establish which safe actuator setting can change the instability."},{"slug":"modal_stability_analysis","disposition":"selected_load_bearing","contribution_type":"CORE_CAUSAL","adaptation_or_rejection":"Interpret eigenvalues against the appropriate continuous- or discrete-time stability boundary within the local regime.","counterfactual_removal":"The analysis could not distinguish a prominent decaying mode from a dangerous growing mode."},{"slug":"mode_shape_testing","disposition":"selected_supporting","contribution_type":"TEST_DESIGN","adaptation_or_rejection":"Use low-amplitude excitation and sensor response to test analytical mode shapes and attach test-condition bounds.","counterfactual_removal":"A computed but physically unobserved mode could be mistaken for a viable target."},{"slug":"network_spectral_centrality_analysis","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"The state is a coupled physical field measurement, not a node-importance problem; centrality entries would not measure instability leverage.","counterfactual_removal":"No causal, operational, safety, or test-design step changes."},{"slug":"power_iteration_probe","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Dominant-only recovery is unsafe here because near-degenerate or subdominant unstable modes matter; the bounded test assumes a tractable explicit operator.","counterfactual_removal":"The selected full-spectrum workflow remains intact."},{"slug":"principal_component_analysis","disposition":"considered_rejected","contribution_type":"NONE","adaptation_or_rejection":"Variance directions need not be dynamical growth directions; PCA may be exploratory but cannot supply the causal stability classification.","counterfactual_removal":"No required inference or safeguard is lost."},{"slug":"reduced_order_model","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Project validated retained modes into a small offline surrogate for rapid actuator comparisons only inside the declared window.","counterfactual_removal":"The causal test remains possible, but rapid bounded control sweeps become impractical."},{"slug":"residual_reconstruction_test","disposition":"selected_load_bearing","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Evaluate held-out reconstruction error and residual structure before authorizing modal action.","counterfactual_removal":"Omitted dynamics could remain invisible, removing a hard gate on use."},{"slug":"singular_value_decomposition","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Use singular directions and conditioning to check non-normal transient amplification that eigenvalue stability could miss.","counterfactual_removal":"A nominally stable but strongly non-normal operator could receive false reassurance."},{"slug":"spectral_decomposition_report","disposition":"selected_supporting","contribution_type":"SAFETY_GUARDRAIL","adaptation_or_rejection":"Record mode meaning, uncertainty, coupling, validity window, residuals, and prohibited interpretations in the test authorization.","counterfactual_removal":"Mathematical uncertainty and operating limits would not reliably travel to operators."},{"slug":"spectral_gap_monitor","disposition":"selected_supporting","contribution_type":"OPERATIONAL","adaptation_or_rejection":"Re-estimate gaps and mode-shape drift between runs; revoke the fixed target when thresholds fail.","counterfactual_removal":"A once-valid modal target could persist unnoticed after regime change."}],"causal_chain":["Bound the operating point and estimate a local coupled-state transformation from synchronized perturbation-response data.","Decompose it, test empirical mode shapes, and distinguish eigenmode growth from non-normal transient amplification.","Select a reproducible growing mode only if stability, spectral-gap, conditioning, and held-out residual gates pass.","Map bounded actuator phase and amplitude to that mode using sensitivity sweeps and coupling checks.","Apply the setting predicted to reduce its gain; increased effective damping should slow modal growth and delay or prevent high-amplitude oscillation.","Revoke the intervention when residual, gap, mode-shape, or operating-window limits fail."],"baseline":"Fixed operating schedules plus alarms on individual pressure RMS, temperature, or Fourier-peak thresholds, followed by shutdown or manually tuned control after oscillation becomes visible.","nearest_rival":"A frequency-peak tracker with gain-scheduled feedback that suppresses the largest observed pressure peak without estimating the coupled state-transition modes.","authority_safety":{"affected_parties":["Laboratory personnel","Facility and equipment owners","Researchers relying on the results"],"decision_authority":"The combustor principal investigator may approve analysis; the facility safety lead must approve perturbation amplitudes, run limits, and any actuator trial.","authorized_first_step":"On archived runs and then a low-energy instrumented bench run, preregister an operating window, fit the local operator, compare onset forecasts with the baseline, and conduct randomized bounded target-versus-sham actuator pulses below existing trip limits.","excluded_actions":["Deployment to operational turbines or untested combustors","Disabling pressure, temperature, flameout, or fuel-system interlocks","Extrapolation beyond the tested operating window","Large perturbations or autonomous escalation of actuator authority"],"halt_rollback":"Abort the run and return to the existing fixed controller or safe shutdown if any facility trip limit, residual budget, conditioning limit, spectral-gap threshold, or mode-drift threshold is breached."}},"negative_tests":{"strongest_counterevidence":"Thermoacoustic onset may be dominated by nonlinear switching or non-normal transient growth with no reproducible unstable eigenmode; eigenvectors may also be ill-conditioned and change between nominally identical runs.","analogy_break":"Invariant-mode reasoning breaks when flame topology and boundary conditions change faster than a useful local window, because the estimated operator and its eigenvectors then cease to represent persistent physical directions.","failure_condition":"The candidate fails if no operating window simultaneously yields identifiable modes, acceptable held-out residuals, adequate conditioning and spectral separation, and safe actuator leverage.","problem_falsifier":"Across repeated pre-onset runs, coupled variables show no reproducible growing direction, and the ordinary single-sensor or frequency-threshold baseline predicts onset equally well under held-out operating conditions.","intervention_falsifier":"With the modal model still passing its validity gates, preregistered target-versus-sham actuator pulses do not reduce the selected mode's estimated growth rate or delay onset beyond uncertainty.","risks":["A locally stable eigenvalue could conceal dangerous transient amplification.","Sensor placement may leave a hazardous mode unobserved.","Actuator perturbations could excite another mode or approach flameout.","Near-degenerate modes may swap identity and create false targeting.","A compact model may extrapolate confidently outside its fitted regime."]},"null_rationale":null,"classification":{"candidate_kind":"TESTABLE_CONJECTURE","prior_art_status":"UNSEARCHED","evidence_maturity":"HYPOTHESIS"},"revision_change_log":{"revision_kind":"ORIGINAL","prior_problem_id":null,"prior_causal_lever_id":null,"problem_changed":false,"causal_lever_changed":false,"conceptual_changes":[],"operational_changes":[],"repairs_addressed":[]},"confidence":0.82,"generator_notes":"Closed-book structural inference from the supplied packet. The thermoacoustic application and all domain-specific empirical relations are hypotheses or inferences pending the bounded test."}