{"actors":["Mechanical design engineer","Manufacturing engineer","Quality engineer","Metrology technician","Assembly process owner","Material review board chair"],"affected_objective":"Decide how to allocate shaft-and-bearing fit tolerances and which otherwise permitted component pairings warrant engineering review, while preserving continuous clearance information and existing hard acceptance limits.","arm":"ORDINARY_DIVERSE_P2","authority_safety":{"authorized_first_step":"Construct and evaluate the continuous fit-margin model retrospectively for one fixed shaft-and-bearing configuration using existing measurement and press records; outputs may be used only for comparison and may not determine component acceptance, pairing, or drawing changes.","decision_authority":"The responsible design authority and material review board retain authority over drawing tolerances, component disposition, pairing rules, and assembly-process changes under existing procedures.","excluded_actions":["Accept a component that violates an approved drawing limit","Automatically pair or reject components from the model output","Change press-force limits, tooling settings, or assembly instructions","Interpolate across material, coating, supplier-process, or drawing-revision boundaries","Treat a calculated fit margin as more precise than the underlying metrology uncertainty"],"halt_rollback":"Stop the evaluation if component identities cannot be reconciled with assembly records, measurement uncertainty is unavailable or comparable to the modeled distinctions, configuration boundaries are ambiguous, or the calculation labels a known unacceptable assembly as ordinary. Discard operational recommendations and retain current drawings, inspection rules, and disposition procedures."},"baseline":"Inspect shaft outside diameter and bearing bore against separate drawing limits, record each component as conforming or nonconforming or place it in a prescribed size bin, and assess the joint through worst-case stack calculations and established assembly checks. Components within the same accepted category are treated as interchangeable for the pairing decision.","candidate_id":"discrete_continuous_model_selection__engineering_design__ORDINARY_DIVERSE_P2","causal_chain":["Shaft diameter, bearing-bore diameter, temperature correction, and measurement uncertainty vary continuously within a fixed joint configuration.","Separate pass/fail labels or size bins collapse distinct measured values before the interference fit is evaluated.","The joint-relevant quantity is a signed diametral fit margin formed from the paired measurements, so two conforming components can occupy materially different positions within the permitted fit range even when their labels match.","Conversely, nearly identical measurements on opposite sides of a component-bin boundary can receive different dispositions despite producing nearly identical joint margins.","The intervention retains measurements as continuous values, propagates their uncertainty into the paired fit margin, and delays quantization until an authorized decision explicitly requires a category.","Material, coating, supplier-process, and drawing-revision changes remain non-interpolable configuration boundaries, preventing continuous estimation from smoothing across different physical regimes.","Retrospective validation compares calculated margins and uncertainty with press force-displacement traces, seating results, final inspection outcomes, and current dispositions near both component and joint boundaries.","The comparison reveals whether continuous or categorical representation better preserves the information needed for tolerance allocation and pairing review; existing authority and rules remain controlling regardless of the result."],"cell_id":"discrete_continuous_model_selection__engineering_design","consequence":"Premature binning can conceal joint-level fit variation, create arbitrary differences between nearly identical components, and obscure which tolerance contributes to an assembly concern; unjustified smoothing across configuration changes can instead combine physically unlike joints and produce misleading fit estimates.","diversity_from_prior_proposals":"This opportunity addresses dimensional tolerance allocation and component-pairing representation for a static interference-fit joint. It selects a continuous, uncertainty-aware model in place of component-level quantization; its problem, intervention, evidence, and causal path are independent of fatigue accumulation, load-event preservation, and maintenance scheduling.","experiment_id":"eoa_inverse_innovation_exp13_second_slot_policy60_20260806","intervention":"Create a non-controlling continuous fit-margin record for one fixed shaft-and-bearing configuration. For every traceable permitted pairing, retain calibrated shaft diameter and bearing-bore measurements as continuous values, apply predeclared temperature corrections, calculate signed diametral interference, and propagate measurement uncertainty without first converting inputs into pass/fail labels or size bins. Analyze each material, coating, supplier-process, and drawing revision separately. Display the calculated margin beside the current component dispositions, applicable hard limits, and observed press force-displacement and seating outcomes. Audit cases near existing component-bin and joint-fit boundaries before considering any change to tolerance allocation or pairing policy.","mechanism_mapping":[{"counterfactual_removal":"Without continuous fit calculation, within-limit dimensional differences are collapsed into labels before their combined joint consequence can be evaluated.","mechanism_slug":"continuous_process_model","role":"Represents diameters, thermal corrections, uncertainty, and signed interference as continuous quantities within a fixed physical configuration."},{"counterfactual_removal":"Without examining the quantization rule, reviewers cannot distinguish consequences of the physical fit from information loss introduced by component bins or pass/fail categories.","mechanism_slug":"quantization_rule","role":"Makes explicit where continuous measurements currently become component categories and defers that conversion until an authorized action boundary requires it."},{"counterfactual_removal":"Without a resolution audit, apparent differences in fit margin may be artifacts of gauge precision, rounding, temperature correction, or unstable classifications near a boundary.","mechanism_slug":"transition_resolution_audit","role":"Tests whether metrology resolution supports the modeled distinctions and whether component and joint boundary cases are reproducibly classified."},{"counterfactual_removal":"Without an explicit comparison between categorical and continuous forms, the analysis could become an ordinary tolerance calculation without testing whether model granularity changes the engineering decision.","mechanism_slug":"state_machine_vs_flow_model","role":"Compares the current conforming-bin representation with a continuous signed-margin representation for the same tolerance-allocation and pairing-review task."}],"nearest_rivals":["Worst-case tolerance-stack analysis: preserves limiting combinations for design assurance but does not necessarily retain unit-level measured position within the allowed fit range or test information lost through component categories.","Root-sum-square statistical tolerancing: estimates assembly variation from assumed component distributions but may not address traceable component pairings, measurement uncertainty, or observed boundary-case outcomes.","Selective assembly by size bin: coordinates component pairing but deliberately quantizes measurements and can preserve the bin cliffs under examination.","Gauge repeatability and reproducibility study: evaluates measurement-system variation but does not decide whether the engineering quantity should remain continuous or be represented as component categories.","Tightening every component tolerance: changes manufacturing requirements without first testing whether the observed decision problem arises from physical variation or from the representation used to combine it."],"negative_tests":{"intervention_falsifier":"Reject the intervention for this decision if the predeclared continuous model cannot reproduce signed fit margins from traceable measurements, its distinctions are dominated by measurement uncertainty, or it explains boundary-case press and seating outcomes no better than the existing categories while introducing additional unresolved judgments.","problem_falsifier":"The representation problem is weakened if approved worst-case limits already guarantee the required joint behavior for every permitted pairing, current decisions use the unbinned measurements directly, and retrospective outcomes remain invariant across continuous values, existing bins, and pass/fail labels.","risks":["Measurement error, thermal correction error, roundness, surface finish, or elastic deformation may dominate nominal diametral interference.","Pair-level traceability may be incomplete or incorrectly linked to press records.","A continuous result may imply unsupported precision or encourage use outside the validated configuration.","Observed associations between fit margin and press response may be confounded by lubrication, tooling alignment, or process settings.","Deferring quantization may increase review burden without changing tolerance-allocation decisions.","Reviewers may mistakenly treat shadow calculations as authority to accept a nonconforming component."],"strongest_counterevidence":"The approved component limits may already be derived so that every conforming combination satisfies the joint requirement, with press and seating outcomes determined mainly by controlled factors not represented by diametral fit; in that case, continuous pairing calculations would not improve the relevant design decision."},"next_evidence_step":"Select one unchanged shaft-and-bearing configuration and a bounded sequence of 60 consecutive assemblies with traceable component measurements and press records. Before calculating results, specify the fit equation, temperature reference, uncertainty propagation, configuration exclusions, existing category boundaries, hard limits, missing-data rules, and comparison criteria. Reconstruct continuous fit margins in shadow mode, then compare continuous values and existing categories against press force-displacement traces, seating results, final inspection outcomes, and engineering dispositions. Report unresolvable records, category cliffs, within-bin outcome variation, uncertainty-dominated distinctions, and cases in which representation alone changes the review conclusion. Do not modify drawings, pairing rules, or component dispositions from this step.","observable_state":"For each included assembly, traceable shaft outside-diameter and bearing-bore measurements, gauge identification and resolution, measurement temperature, component identifiers, material and coating specification, supplier-process identifier, drawing revision, current inspection category, pairing record, press force-displacement trace, seating check, final inspection result, and engineering disposition are available or explicitly marked missing.","prior_art_status":"UNSEARCHED","problem":"A shaft-and-bearing interference-fit design is governed through separate component limits and categorical size dispositions even though the assembly-relevant fit is the continuous difference between paired dimensions under measurement and temperature uncertainty. Quantizing each component before evaluating the joint can make unequal pairings appear equivalent, place nearly identical pairings on opposite sides of a category boundary, and obscure how tolerance allocation relates to observed assembly behavior.","proposal_index":2,"remaining_contrastive_claim":"For a fixed interference-fit configuration, a traceable continuous signed-margin model can be tested as a distinct decision representation from component pass/fail labels, selective-assembly bins, worst-case limits, and distribution-only tolerance calculations because it retains paired measured position and uncertainty until an authorized action boundary is applied.","revision_record":{"claim_changes":["Initial version; no prior claim was revised."],"conceptual_changes":["Initial version maps component-level false discreteness to a continuous, configuration-bounded interference-fit representation."],"evidence_changes":["No evidence is claimed; a bounded retrospective comparison is specified."],"operational_changes":["Initial version restricts the model to non-controlling retrospective use and retains existing design and disposition authority."],"parent_version":null,"progress_targets_addressed":["Independent engineering-design problem and decision task","Explicit cost of false discreteness and false smoothness","Continuous model form with configuration boundaries","Measurement-resolution and uncertainty alignment","Boundary-case validation and falsifiers","Authority limits, halt conditions, and reversible evidence step"]},"schema_version":1,"structural_mapping":[{"archetype_element":"Decision Need","domain_realization":"Determine whether tolerance allocation and permitted pairing rules for a shaft-and-bearing joint preserve the required fit behavior without relying on information-destroying component categories."},{"archetype_element":"Process Change Signature","domain_realization":"Manufactured diameters, temperature corrections, and resulting interference vary continuously within a fixed joint configuration; material, coating, supplier-process, and drawing revisions define separate regimes."},{"archetype_element":"Cost of False Smoothness","domain_realization":"Interpolating across a coating, material, process, or drawing-revision change can combine joints with different physical behavior and hide a genuine configuration boundary."},{"archetype_element":"Cost of False Discreteness","domain_realization":"Component pass/fail labels and size bins can treat different fit margins as equivalent while assigning different labels to nearly identical margins around a cutoff."},{"archetype_element":"Granularity Choice","domain_realization":"Preserve calibrated component dimensions, corrections, signed interference, and uncertainty continuously within each declared configuration; apply discrete labels only at existing authorized decision boundaries."},{"archetype_element":"Step Boundary","domain_realization":"Drawing limits, approved joint-fit limits, and changes of material, coating, supplier process, or drawing revision are explicit boundaries and are not inferred from the continuous trend."},{"archetype_element":"Continuity Assumption","domain_realization":"The fit equation is applied continuously only within an unchanged joint geometry and physical configuration, using measurements whose calibration, temperature basis, and uncertainty satisfy predeclared requirements."},{"archetype_element":"Measurement Resolution","domain_realization":"Gauge resolution, repeatability information, rounding, and temperature uncertainty must support the distinctions shown in the calculated fit margin."},{"archetype_element":"Transition Validation","domain_realization":"Cases around component-category and joint-fit boundaries are checked against repeatable measurements, press force-displacement traces, seating checks, final inspection, and engineering disposition."},{"archetype_element":"Approximation Error Check","domain_realization":"Continuous and categorical representations are compared for category cliffs, within-bin variation, uncertainty-dominated distinctions, and representation-caused changes in review conclusions."},{"archetype_element":"Hybrid Boundary Rule","domain_realization":"Although dimensions remain continuous within a configuration, no value is propagated across a declared physical or drawing change; a new configuration starts a separate model population."},{"archetype_element":"Reversibility Check","domain_realization":"The first evaluation creates only a shadow comparison record, so removing it leaves drawings, accepted pairings, inspection categories, and disposition history unchanged."}],"title":"Continuous Fit-Margin Audit for Shaft–Bearing Tolerance Allocation","version":0}