--- name: neqsim-consequence-analysis version: "1.0.0" description: "Quantitative consequence analysis for oil & gas hazards — jet fire, pool fire, vapour cloud explosion (VCE), BLEVE, Gaussian plume and heavy-gas dispersion, probit-based fatality probabilities, individual and societal risk roll-up. USE WHEN: a task requires fire-radiation contours, dispersion to LFL/IDLH/ERPG, BLEVE thermal/missile assessment, or QRA-style risk integration of multiple release outcomes. Anchors on neqsim.process.safety.fire, neqsim.process.safety.dispersion, neqsim.process.safety.qra." last_verified: "2026-04-26" requires: java_packages: - neqsim.process.safety.fire - neqsim.process.safety.dispersion - neqsim.process.safety.qra --- # NeqSim Consequence Analysis Skill Quantitative consequence modelling that converts a release scenario (mass flow, inventory, ignition probability) into thermal radiation contours, overpressure contours, dispersion footprints, and finally individual / societal fatality risk per ISO 17776, NORSOK Z-013, API 752 and the CCPS *Guidelines for Chemical Process Quantitative Risk Analysis*. ## When to Use - Jet-fire, pool-fire, VCE or BLEVE source-term modelling - Toxic / flammable dispersion to LFL, IDLH, ERPG-2/3 or 1 % fatality contours - Probit-based fatality probability for thermal radiation, overpressure or toxic dose - QRA roll-up: combine release frequency × ignition probability × fatality probability - Source-term generation as input to PHAST / FLACS / KFX Distinct from `neqsim-process-safety` (HAZOP / LOPA / SIL — frequency side) and `neqsim-relief-flare-network` (PSV sizing / flare radiation — design side). This skill is the *consequence* side of QRA. ## Standards - **API 521** §6 — flare and vent radiation, fire heat input - **API 752 / 753 / 756** — facility siting, occupied buildings - **NORSOK Z-013 / S-001** — risk acceptance, fire & explosion loads - **CCPS QRA Guidelines** — probit constants, ignition probabilities - **TNO Yellow Book** — multi-energy / Baker-Strehlow VCE - **EI 15** — hazardous area classification ## Method 1 — Jet Fire (API 521 + CCPS solid-flame) ```java import neqsim.process.safety.fire.JetFireModel; import neqsim.process.safety.dispersion.ProbitModel; // 30 kg/s gas leak, 50 MJ/kg HoC, 25 % radiative fraction JetFireModel jet = new JetFireModel(30.0, 50.0e6, 0.25); double flux50m = jet.radiationFluxAt(50.0); // W/m² double dist125 = jet.distanceForFlux(12500.0); // m where flux = 12.5 kW/m² ``` Typical thermal radiation criteria (API 521 / NORSOK S-001): | Flux (kW/m²) | Effect | Use | | ------------ | --------------------------------------- | -------------------------- | | 1.6 | No discomfort for long exposure | Public area limit | | 4.7 | Sufficient for evacuation in 30 s | Escape route | | 12.5 | Wood ignites, equipment failure (5 min) | Process equipment limit | | 37.5 | Structural failure of process equipment | Damage to steel structure | ## Method 2 — Pool Fire ```java import neqsim.process.safety.fire.PoolFireModel; // 500 kg liquid, burning rate 0.05 kg/m²·s, dike diameter 8 m, η = 0.30 PoolFireModel pool = new PoolFireModel(0.05, 8.0, 50.0e6, 0.30); double flux = pool.radiationFluxAt(25.0); // W/m² at 25 m ``` ## Method 3 — Vapour Cloud Explosion (TNO Multi-Energy) ```java import neqsim.process.safety.fire.VCEModel; // 200 kg flammable in cloud, congestion class 7 (heavy) VCEModel vce = new VCEModel(200.0, 50.0e6, 7); double overpressure = vce.overpressureAt(60.0); // Pa at 60 m double safeDist = vce.distanceForOverpressure(20684.0); // m for 3 psi ``` ## Method 4 — BLEVE ```java import neqsim.process.safety.fire.BLEVECalculator; // 50 t propane vessel BLEVECalculator bleve = new BLEVECalculator(50000.0, 50.0e6, 0.40); double fireballDiameter = bleve.fireballDiameter(); // m double fireballDuration = bleve.fireballDuration(); // s double thermalDose = bleve.thermalDoseAt(150.0); // (W/m²)^(4/3)·s ``` ## Method 5 — Gaussian Plume Dispersion (Briggs σ) ```java import neqsim.process.safety.dispersion.GaussianPlume; // 10 kg/s leak, ground-level source, 4 m/s wind, neutral stability D GaussianPlume plume = new GaussianPlume(10.0, 0.0, 4.0, GaussianPlume.Stability.D, GaussianPlume.Terrain.RURAL); double conc = plume.centerlineGroundConcentration(200.0); // kg/m³ double distLFL = plume.distanceToConcentration(0.044); // m to methane LFL ``` Stability classes: A (very unstable) … F (stable). RURAL vs URBAN selects Briggs σ coefficients per Pasquill-Gifford. Use `Stability.F` for worst-case dispersion analysis (calm night, low wind). ## Method 6 — Heavy-Gas Dispersion For dense releases (CO₂, propane, butane) the Gaussian model under-predicts near-field concentrations. Use `HeavyGasDispersion` (Britter-McQuaid screening): ```java import neqsim.process.safety.dispersion.HeavyGasDispersion; HeavyGasDispersion hgs = new HeavyGasDispersion( 50.0, // continuous release rate [kg/s] 1.98, // gas density at release [kg/m³] 1.20, // ambient density [kg/m³] 4.0); // wind speed [m/s] double distLFL = hgs.distanceToConcentration(0.05); // m ``` ## Method 7 — Probit Fatality Probability ```java import neqsim.process.safety.dispersion.ProbitModel; // Thermal: Y = a + b·ln(t·F^(4/3)), 60 s exposure at 12.5 kW/m² double pFatality = ProbitModel.thermalFatality() .fatalityProbability(60.0, 12500.0); // Toxic H2S: Y = -31.42 + 3.008·ln(C^1.43·t) ProbitModel h2s = ProbitModel.h2sFatality(); double pH2S = h2s.fatalityProbability(600.0, 5.0e-4); // 10 min, 500 ppm // Overpressure (lung haemorrhage): Y = -77.1 + 6.91·ln(P) ProbitModel ovp = ProbitModel.overpressureFatality(); ``` Built-in factories: `thermalFatality()`, `overpressureFatality()`, `h2sFatality()`, `cl2Fatality()`, `nh3Fatality()`, `coFatality()`. The `ToxicLibrary` class centralises probit constants for common toxic gases. ## Method 8 — QRA Roll-up ```java import neqsim.process.safety.qra.ConsequenceAnalysisEngine; ConsequenceAnalysisEngine e = new ConsequenceAnalysisEngine( "10 mm gas leak", 1.0e-4); // release frequency [/yr] e.addJetFire(0.05, jet, ProbitModel.thermalFatality(), 60.0); e.addJetFire(0.02, pool, ProbitModel.thermalFatality(), 60.0); e.addToxicCloud(0.01, plume, ProbitModel.h2sFatality(), 600.0); double IRPA = e.individualFatalityRiskPerYear(50.0); // at 50 m String text = e.report(50.0); ``` The engine sums: `IRPA(d) = Σ_outcome f_release · f_branch · P_fatality(d, outcome)` Compare against acceptance criteria: | Criterion | Limit | Source | | ---------------- | ---------------------- | ------------------- | | Worker IRPA | 1·10⁻³ /yr (intolerable) | UK HSE R2P2 | | Worker IRPA | 1·10⁻⁶ /yr (broadly acceptable) | UK HSE R2P2 | | NORSOK FAR | 10 fatalities / 10⁸ h | NORSOK S-001 | | Public 1 % fatal | 35 m typical for 12.5 kW/m² | API 752 | ## Source Term for External CFD `ConsequenceAnalysisEngine.exportSourceTerm()` writes a JSON block usable by PHAST, FLACS, KFX or DNV Safeti containing release rate, momentum, density, duration and chemistry — the standard handoff format described in `neqsim-agent-handoff`. ## Common Pitfalls - **Stability class** — using class D ("typical") instead of F for hazard contours under-predicts safe distance by 2–3×. Always run F at low wind for siting. - **Heavy gas** — applying Gaussian to CO₂ / LPG releases under-predicts near-field concentration. Switch to `HeavyGasDispersion` when density ratio > 1.2. - **Probit constants** — different sources give different (a, b, n). Always cite the source; the factories in `ProbitModel` use CCPS values. - **Ignition probability** — small leaks (< 1 kg/s) often use 1–5 %, large leaks (> 50 kg/s) up to 30 % delayed. Use scenario-specific values, not 100 %. - **Flame view-factor** — the simple 1/(4πr²) point-source form is conservative near the flame; use API 521 solid-flame for distances < 2 × flame length. ## Verification Tests `src/test/java/neqsim/process/safety/{fire,dispersion,qra}/` contain JUnit 5 tests for every model. Run: ```bash ./mvnw test -Dtest=FireModelsTest,GaussianPlumeTest,ProbitModelTest,ConsequenceAnalysisEngineTest ``` ## See Also - `neqsim-process-safety` — frequency side (HAZOP / LOPA / SIL) - `neqsim-relief-flare-network` — PSV sizing and flare radiation - `neqsim-depressurization-mdmt` — emergency depressurization source terms - `neqsim-agent-handoff` — source-term JSON schema