{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","cell_id":"circuit_breaker__chemistry_materials","arm":"BREADTH_PROBE_ONE_SHOT","candidate_id":"circuit_breaker__chemistry_materials__P1","proposal_index":1,"version":0,"title":"Half-Open Feed Breaker for Semibatch Exothermic Reactors","problem":"In a semibatch exothermic reaction, continued reagent feed can generate heat faster than the reactor can remove it. Rising temperature accelerates reaction kinetics, which further increases heat generation and can turn a local cooling-capacity shortfall into a thermal and pressure excursion.","actors":["process chemist","control engineer","batch operator","process-safety authority","semibatch reactor control system"],"observable_state":"The controller observes reactor temperature, temperature-rise rate, jacket inlet and outlet temperatures, coolant-valve saturation, reactor pressure, agitator status, and cumulative reagent feed. A compound overload state is present when temperature or rise rate crosses its trip boundary while cooling is near saturation or another heat-removal impairment is detected. Separate, lower reset thresholds provide hysteresis.","consequence":"If feed remains coupled to the heat-limited reactor, unreacted reagent can accumulate or react at an accelerating rate, threatening batch containment, relief-system demand, equipment limits, and product integrity.","affected_objective":"Maintain reactor containment and controllability while preserving material accounting and allowing only bounded recovery attempts after the heat-removal margin returns.","intervention":"Place a stateful breaker at the reagent-feed boundary. When the compound overload signal persists for a validated interval, the controller closes a fail-closed isolation valve and inhibits the feed pump, recording the rejected or retained feed rather than silently losing it. The breaker remains open until temperature, rise rate, pressure, agitation, and cooling margin all satisfy lower reset thresholds for a dwell period. It may then enter an operator-enabled half-open state that admits one pre-specified probe pulse. The pulse's measured thermal response determines whether the breaker returns to open or permits progressively bounded pulses; unrestricted feed requires the approved recovery criteria to remain satisfied.","structural_mapping":[{"archetype_element":"Coupled flow","domain_realization":"A metered reactive reagent flows into a vessel whose reaction rate and heat release depend on the vessel's evolving temperature and composition."},{"archetype_element":"Capacity constraint","domain_realization":"The limiting component is effective heat removal through the jacket or coil, conditional on coolant availability, heat-transfer performance, and mixing."},{"archetype_element":"Cascade risk","domain_realization":"A heat-removal deficit raises temperature; higher temperature can accelerate reaction; faster reaction releases heat more rapidly and can further reduce the remaining control margin."},{"archetype_element":"Controllable boundary","domain_realization":"A fail-closed feed-isolation valve plus pump inhibit severs incoming reagent from the reactor without altering already admitted material."},{"archetype_element":"Overload sensor and threshold","domain_realization":"A validated compound trigger combines temperature or temperature-rise rate with evidence such as coolant saturation, pressure rise, or loss of agitation, rather than relying on a single noisy measurement."},{"archetype_element":"Feedback","domain_realization":"Post-trip temperature, rise rate, pressure, cooling margin, and response to a probe pulse govern state transitions instead of elapsed time alone."},{"archetype_element":"Sampling under degraded service","domain_realization":"The half-open state substitutes bounded reagent pulses for continuous full-rate feed, exposing the reactor to only the authorized probe quantity."},{"archetype_element":"Hysteresis","domain_realization":"Trip thresholds and reset thresholds are separated and reset requires a dwell period, limiting valve chatter near the thermal boundary."},{"archetype_element":"Preserved invariants","domain_realization":"No automatic quench, venting, relief suppression, feed-accounting bypass, or unbounded restart occurs; admitted and withheld reagent remain traceable."}],"mechanism_mapping":[{"mechanism_slug":"boundary","role":"The fail-closed valve and pump inhibit interrupt the physical reagent flow that couples additional reaction heat into a saturated vessel.","counterfactual_removal":"Without an enforceable boundary, overload detection can alarm but cannot stop continued addition, so the destabilizing coupling remains active."},{"mechanism_slug":"feedback","role":"Measured recovery margin and probe response determine whether the breaker stays open, reopens, or advances through bounded feed states.","counterfactual_removal":"Without feedback, restart becomes a timer-based or manual guess that can reapply feed before heat-removal capacity has recovered."},{"mechanism_slug":"half_open_state","role":"A distinct transitional state prevents a direct jump from zero feed to the normal recipe rate.","counterfactual_removal":"Without the transitional state, the first recovery action recreates the original load and provides no bounded test of reactor response."},{"mechanism_slug":"limited_probe","role":"A pre-specified feed pulse tests the reactor's incremental thermal response while limiting the newly introduced reactive inventory.","counterfactual_removal":"Without a limited probe, the controller lacks direct evidence that restored cooling margin remains adequate when reagent addition resumes."},{"mechanism_slug":"clean_rejection","role":"Feed that is not admitted stays isolated and is explicitly recorded, while the batch state reports a predictable breaker-open condition.","counterfactual_removal":"Without clean rejection and accounting, pump commands, held-up material, or recipe totals can diverge from actual reactor inventory and compromise recovery decisions."}],"causal_chain":["Reactive reagent feed increases the vessel's potential reaction and heat-generation load.","Cooling degradation, mixing impairment, or an overly aggressive feed rate reduces the margin between heat generation and heat removal.","Temperature or temperature-rise rate crosses a validated boundary while a corroborating capacity signal indicates coupling to a constrained heat-removal path.","The breaker closes the feed boundary, preventing additional reagent from adding to the developing overload.","The controller continues observing the already charged reaction while hysteresis and a dwell interval prevent premature reopening.","After recovery criteria are met, an operator-enabled half-open state admits one bounded pulse.","The pulse's incremental temperature, pressure, and cooling response either returns the breaker to open or supports another bounded recovery step.","Normal feed becomes eligible only after the approved recovery sequence remains within all limits."],"baseline":"The baseline is recipe-rate feed governed by ordinary continuous temperature control, with operator alarms and a separate emergency shutdown. It lacks a stateful compound trip tied to automatic feed isolation and lacks a feedback-governed half-open probe state.","nearest_rivals":["Continuous PID feed throttling, which modulates rate around a setpoint but may continue adding reagent during a compound overload state.","A one-way emergency shutdown or quench, which terminates or chemically suppresses the batch but does not provide cautious feedback-governed reentry.","A high-temperature feed interlock with timer-based reset, which interrupts flow but lacks corroborating capacity signals, hysteretic recovery criteria, and a limited probe.","Pressure relief, which protects containment at a downstream limit but does not interrupt the reagent flow creating additional reactive load."],"remaining_contrastive_claim":"The candidate's distinguishing proposition is the combination of compound overload detection, enforced feed isolation, and an operator-enabled half-open pulse whose measured thermal response controls reentry. Removing the stateful probe-and-feedback transition reduces it to an interlock; allowing continuous modulation during the overload reduces it to throttling.","authority_safety":{"decision_authority":"The site's designated process-safety authority approves the sensed variables, trip and reset boundaries, maximum probe quantity, dwell times, eligible chemistries, and recovery envelope. The batch operator may keep the breaker open or abort the batch but may not widen those limits. The controller may execute only preauthorized state transitions.","authorized_first_step":"Use archived reaction-calorimetry traces and a hardware-in-the-loop simulated pump and valve to replay normal, cooling-loss, agitation-loss, sensor-fault, and delayed-reaction cases. No reactive material is charged and no production control logic is deployed.","excluded_actions":["testing the initial logic on a live production batch","automatically reopening feed without an approved recovery mode and operator enable","changing relief, quench, vent, or emergency-shutdown settings","using the breaker to justify exceeding an established safe operating envelope","bypassing independent high-high temperature or pressure protection","treating a single failed sensor as proof of thermal overload without the approved fault-handling rule"],"halt_rollback":"The offline test halts on any simulated violation of temperature, pressure, inventory, or state-transition limits, any valve-command ambiguity, or any unaccounted feed. Rollback is removal of the proposed breaker logic from the simulator configuration and restoration of the unchanged baseline model; no plant configuration is modified."},"negative_tests":{"strongest_counterevidence":"The strongest counterevidence would be calorimetry showing that, after the trigger becomes observable, heat release is dominated by material already accumulated in the vessel, so feed isolation and bounded reentry do not meaningfully control the excursion and may delay the required abort or quench decision.","problem_falsifier":"The problem framing is falsified for a target reaction if excursions arise primarily from sensor error, phase separation, blocked venting, incorrect chemistry, or an internally propagating decomposition whose trajectory is not materially coupled to ongoing reagent feed or recoverable heat-removal capacity.","intervention_falsifier":"The intervention is falsified if trace replay and validated dynamic models show that its triggers miss credible feed-coupled excursions, produce unsafe reopening, chatter despite hysteresis, admit an excessive probe inventory, obscure an abort condition, or perform no better on the defined safety-state metrics than the existing independent interlock.","risks":["A delayed exotherm can make the probe response appear safe before its heat release becomes observable.","Temperature measurements may not represent local hot spots in a poorly mixed vessel.","Closing the feed valve may leave reactive material between the valve and injection point.","Repeated probe pulses can accumulate reagent even when each pulse appears individually acceptable.","Shared sensor or controller failures can defeat both detection and actuation.","Nuisance trips can strand batches in unstable or off-spec states and encourage unauthorized bypasses.","Automatic recovery logic can distract from conditions that require irreversible shutdown, quench, or evacuation.","The selected probe may be unrepresentative of the normal feed rate or later batch composition." ]},"next_evidence_step":"For one specified reaction recipe, conduct a bounded offline replay using an approved calorimetric model and historical traces: compare the baseline and proposed logic across a fixed matrix of normal operation, cooling-capacity loss, agitation loss, delayed heat release, single-sensor fault, and valve delay. Record trigger timing, maximum simulated temperature and pressure, admitted inventory after first warning, state oscillations, false trips, and whether every half-open pulse remains within the preauthorized envelope. The result only determines whether the logic merits independent process-hazard review; it does not authorize reactive testing or deployment.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Not assessed against other proposals because runtime isolation forbids consulting them; this candidate is defined solely as a semibatch-reactor feed interruption and feedback-governed recovery system.","revision_record":{"parent_version":null,"progress_targets_addressed":["Initial one-shot candidate grounded in the supplied circuit-breaker archetype and chemistry/materials domain card."],"conceptual_changes":["None; this is version 0."],"operational_changes":["None; this is version 0."],"evidence_changes":["No external or prior-art evidence consulted."],"claim_changes":["No novelty, prevalence, demand, or effect-size claim is made."]}}