OECD PRISME Project on Fire Propagation Final Seminar Overview of the PRISME project Technical Description and Main Outcomes Audouin L., Rigollet L., Pretrel H., Le Saux W. IRSN / ETIC
Summary (1) Context & Objectives (2) Description of Experimental Facilities (3) Description of Experimental Campaigns (4) Some Outstanding Results (5) Conclusion
1 Context & Objectives PRISME Project: Program overview (1) Context: the PRISME fire research program (2006-2011) was conducted in an international framework (Project Secretariat: OECD/NEA ; Operating agent: IRSN) (2) 20 partners from 12 countries: Belgium (TRACTEBEL-Suez, BEL_V), Canada (AECL), Finland (STUK, VTT), France (IRSN, EdF, DGA/IUSTI), Germany (GRS, ibmb, BfS), Japan (JNES), Korea (KINS), Spain (CSN), Sweden (Vattenfall Ringhals), UK (HSE), The Netherlands (VROM-KFD, NRG), and USA (NRC). (3) Main objectives: Investigate smoke and heat propagation mechanisms in multi-compartment fire scenarios and assess the consequences of fire on targets of interest (thermal stress on electrical cables and their potential malfunction).
2 Description of Experimental Facilities Description of DIVA facility 4 m 4 m ROOM 4 ROOM 3 ROOM 2 ROOM 1 5 m 5 m 5 m CORRIDOR 6 m 2.5 m (1) 5 compartments (4 rooms and 1 corridor). (2) Rooms 1 to 3 (120 m 3 ), room 4 (180 m 3 ) and corridor (160 m 3 ). (3) Each room is connected with a mechanical ventilation system by means of inlet and outlet duct. (4) Each room can be connected with its adjacent rooms through doorway and/or simple openings. (5) Instrumentation: up to 800 possible measurement channels on the data acquisition system. (6) Measurements: Fuel mass, temper., gas (CO, CO2, O2 and HCT) and soot concentrations, heat fluxes, pressures, flow rates in all compartments and in ventilation network, video.0
2 Description of Experimental Facilities Description of SATURNE calorimeter (1) Hood: 3 m in diameter. 3 m 4.2 m (2) For these PRISME support tests, the height between the floor and the bottom rim of the hood was about 4 m. (3) The smoke exhaust system is connected to a ventilation network. Its exhaust flow rate can be ranged from 10,000 to 25,000m 3 /h. (4) This calorimeter is designed for studying fire source up to nearly 1.5 MW. (5) Measurements: Pressures, gas flow rates, temperatures, gas concentrations (O 2, CO, CO 2 and HCT) and soot concentration, heat fluxes, video (.
3 Description of Experimental Campaigns Summary of fire tests carried out during the PRISME project
3 Description of Experimental Campaigns Description of PRISME Source campaign Free Atmosphere (= => no air viciation) PRISME Source (1) Objective: Fire behavior of HTP pool fire (HTP=Hydrogenated Tetra- Propylene, C 12 H 26 ) in open atmosphere. (2) Parameters: pool area (0.1 to 0.4 m²). Confined and Ventilated Rooms (=> vitiated envir.) Single Compartment (PRISME Source) Name D1 D2 D3 D4 D5 D5a D6 D6a S (m 2 ) 0.4 0.4 0.4 0.4 0.2 0.2 0.4 0.4 Tr (h - ) 4.7 8.4 1.5 4.7 4.6(2 ) 1.6 4.7 1.7 Inlet H H H H H H B B
3 Description of Experimental Campaigns Description of PRISME Door campaign PRISME Door (1) Objective: The spread of smoke and hot gases through open doors for two and three rooms + Heat transfer to surrogate and real cables + Effect of oxygen depletion on fire source. (2) Parameters: Pool area, the initial mass of fuel (HTP), ventilation flow rate, number of compartments and location Two and Three Compartments of air inlet. (PRISME Door) Name PRS_D3 PRS_D2 PRS_D1 PRS_D4 PRS_D5 PRS_D6 S (m 2 ) 0.4 0.4 0.4 0.4 1 1 Tr 4.7 1.5 0 8.6 4.7 4.7 Nb of room 2 2 2 2 2 3
3 Description of Experimental Campaigns Description of PRISME Leak campaign PRISME Leak (1) Objectives: + Propagation of smoke and hot gases through leakages (Leak 1 to 3) + thermal transfer on a duct crossing the fire room and flowing in the adjacent room (Leak 4) + Cable performance tests (fire thermal stress, T>430 C nearby upper cables). (2) Parameters: Fuel (HTP, 0.6m²), type of leakages (3), duct or not.
3 Description of Experimental Campaigns Description of PRISME Integral campaign (open) Inlet 3100 m 3 /h L1 L2 L3 L0 PRS_INT_2 3100 m 3 /h Exhaust (1) Objectives: The propagation of smoke and hot gases through doorways (more than 300 in L3); The effect of the number of adjacent rooms on the propagation through doorways; The effect of sprinkler activation; The effect of fire damper closure; The behaviors of cable & cabinet fires in confined and ventilated fire scenarios. Inl et 2600m 3 /h PRS_I NT_5 (2) Parameters: Fuels (TPH, cabinet, cables), 3 or 4 rooms connected by doorways, fire dampers and a deluge system (sprinklers). 3100m 3 /h L1 L2 L3 Ex haust 500m 3 /h L0
4 Experimental Results: Pressure effect induced by fires in forced ventilated enclosures Pressure and ventilation system (1) Pressure time history: A typical behavior that consists in an overpressure peak at ignition (possible others peaks of pressure during the combustion phase) and a low-pressure peak at extinction. (2) Background: The pressure variation is the result of the energy balance within compartment, which includes the fire HRR, the energy dissipating through wall s enclosure and the net balance through the ventilation branches. (3) Ventilation network: A direct consequence of the pressure peaks is the variation of air flow in the ventilation system. Reverse flows can occur at fire ignition (inlet) and at fire extinction (outlet).
20 18 16 14 12 10 8 6 4 2 0 dm/dt (g/s) 4 Experimental Results: Effect of oxygen depletion on fuel mass loss rate Oxygen depletion and fuel MLR in fire compartment Tr=1,5 (D3) Free (S4) Free (S3) Tr=4,7 (D6) Tr=8,3 (D2) Tr=1,7 (D6a) S=0,4 m2 Tr=4,7 (D1) t (s) 0 600 1200 1800 2400 3000 Door 1.2 1.0 0.8 0.6 0.4 0.2 0.0 dm/dt * (-) Tr=8.6 Tr=0 Source Free atm. Tr=1.5 Free atm. 2R_Tr=0 2R_Tr=1.5 2R_Tr=4.7 2R_Tr=8.6 Tr=4.7 t (s) 0 300 600 900 1200 1500 1800 2100 (1) Air vitiation: The combustion products fill up quickly the fire compartment involving the oxygen depletion in it (under-ventilated condition). (2) Effect of oxygen depletion on MLR: The fire duration can be either shorter because extinction occurs quickly by lack of oxygen, either drastically longer because the decrease of MLR under steady state condition involves more time to burn all the mass of fuel available in the pan. (3) Effect of oxygen depletion on fire duration: The fire duration for a RR of 4.7 (Source, D1) is around 2.5 times longer compared with the same pool fire in free atmosphere. The same behavior is observed in Door tests.
4 Experimental Results: Effect of ventilation on the velocity profile through the doorways Quick description (1) The effects of the ventilation rate and the fire HRR on the velocity profiles are analyzed to quantify the natural vs forced flows during the steady state regime. (2) The forced flow is induced by the ventilation network as initial condition with the air flowing from the room 1 to the room 3. Integral P12 qv P2 3 L0 L2 L1 hn12 hn23 D12 Q D23 L3
4 Experimental Results: Effect of ventilation on the velocity profile through the doorways Results and discussion Effect of flow rate: The main effect of the forced ventilation is to shift significantly the value of the inflow. At the upstream doorway (D12), the rise of the ventilation flow rate increases the inflow velocity. Inversely, at the downstream doorway (D23), the flow rate increase contributes to reduce the inflow until completely disappears (forced ventilation with one-way flow). => the ventilation network could promote significantly the propagation of hot gas to preferential neighboring compartments. Doorway D 12 - HRR=100kW 1 z* Doorway D 23 - HRR=100kW 1 z* qv- 700m3/h qv- 3100m3/h qv- 700m3/h qv- 3100m3/h 0 U* -0,6-0,4-0,2 0,0 0,2 0,4 0,6 0 U* -0,6-0,4-0,2 0,0 0,2 0,4 0,6
4 Experimental Results: Cable performance testing Electrical malfunction of cables due to fire thermal stress Leak (1) IRMS system (Sandia): The Insulation Resistance Measurement System (IRMS) monitors conductor-to-conductor and conductor-to-ground insulation resistance for various conductor pairs. (2) SCDU system (Sandia): The Surrogate Circuit Diagnostic Unit (SCDU) is nominally configured to simulate a typical motor operated valve (MOV) control circuit (specific conductors were electrically energized and various modes of cable failure monitored). (3) Main outcomes: The results provide the relationship between outside gas temperature and cable electrical failures. These data were in close accordance with those obtained in laboratory tests under controlled radiative heat sources.
5 Conclusion PRISME experimental campaigns: An improvement of fire modeling and a experimental database (more than 35 large-scale fire tests) used to validate fire safety codes (as zone modeling, lumped parameter approach and CFD) Benchmarking Group: Comparisons of various fire codes (two-zone, lumped parameter, CFD) Development of validation strategy (multi-metrics) Sensitivity analysis on fire codes Flow rates through doorways Opening of PRISME Tests Results 2 tests on the IRSN Website (July 2012): PRISME Source D1 & PRISME Door D3 All experimental measurements Tests reports Photos and films Special Issue of Fire Safety Journal