Emerging Risks from Smoldering Fires

INTRODUCTION The project aims at improving safety for people and property in relation to smoldering. Smoldering is a combustion type with no flames. Smoke is produced, and the process evolves slowly. Smoldering may be started from weak and imperceptible heat sources - and is difficult to detect. A smoldering fire involves risks as such. The smoke is more toxic than smoke from most flaming fires. Furthermore, smoldering fires may lead to flaming fires and explosions. Therefore, they represent hidden dangers in many situations: for dwellings, for stored biomass, for powders and grains in silos, for stored industrial waste, for goods under transportation. Smoldering may start without any external heat source. Such self-ignition may occur in stored organic materials like wood pellets. Chemical or biological processes release heat, and the temperature may increase sufficiently for smoldering to start. A crucial parameter is the oxygen concentration. Smoldering may survive at low oxygen concentrations. This is part of the reason why smoldering fires often are very difficult to extinguish, like in waste deposits. The transition to a flaming fire may be explosive: A smoldering fire in a closed space will lead to a very low oxygen concentration. Should a door be opened, fresh air will enter the space and processes intensify. As a result, combustible gasses produced by the smoldering may ignite in an explosive way. RESULTS Many experiments have been carried out with wood pellets as material. Wood pellets has been slowly heated - and it was investigated at which temperature smoldering starts, as one varies sample size, heating, air access, and type of wood pellets. In these experiments with wood pellets, several interesting phenomena have been observed after smoldering has started: Temperature measurements indicate combustions processes that vary in intensity in an apparently random way. Still, the most intense combustion periods seems to be governed by (proportional to) the initial sample size. This applies both for the time till these intense combustion periods and their extent. Furthermore, in the project a striking self-synchronization has been observed in smoldering wood pellets under certain conditions: The entire sample oscillates between intense smoldering and apparent extinguishment - in contrast to the more disordered smoldering combustion one usually has. The smoldering and reaction behavior of wood pellets and grinded wood pellets were investigated in two different cylindrical combustion chambers (lab scale and medium scale). The packed beds were ignited at the top under controlled conditions (temperature, air flow). The temperature of the reaction front and the composition of the gaseous products were measured at different points. At low flow rates an unexpected reverse smoldering reaction occurred in the medium-scale apparatus. With cotton as a material, necessary conditions for a spontaneous transition from a smoldering fire to a flaming fire has been investigated in the project. The transition occurs more easily, the larger the sample is. One (single) side wall at the sample is crucial. The transition occurs relatively late during smoldering experiments, as a partly burnt-out sample allows increased oxygen access. During smoldering, cotton generates smoke consisting of water vapor and non-oxidized substances called pyrolysis oil. Thin films of condensed pyrolysis oil on inert porous media can undergo a surface smoldering process with amazingly high front speeds, up to 1 mm/s, and can also generate sparks. Experiments are performed to study the behavior and the governing parameters of the fronts. It is also investigated whether pyrolysis oil which has condensed on un-burned fuel (cotton) or any porous boundaries close to the fuel might facilitate the transition to flaming. The project also includes work on smoldering in various waste materials, in particular in piles of wood chip waste. We varied the wood chips size and moisture content. In piles of larger wood chips we observed higher temperatures and more intense temperature variations than in piles of fine wood chips. Furthermore, whereas flaming fires were not observed for piles of fine wood chips, both larger wood chips and moist wood chip piles went to flaming. A new cellular automaton developed uses a multi-layer structure and is able to model the dynamics of different smoldering fronts. It has been adapted to account for smoldering fronts that develop finger structures. Preliminary experiments conducted on industry-specific materials for three companies where smoldering is a challenge, short reports completed. With one of the companies, we have obtained additional funding through a VRI project. Collaborators: Western Norway University of Applied Sciences, Universities of Magdeburg and Lund, Rise Fire Research, Imperial College, Weizmann Institute, and industry.

VERKNADER: Ulmebrannar er lite forska på samanlikna med flammebrannar. I prosjektet har Høgskulen på Vestlandet og samarbeidande institusjonar i Tyskland, Storbritannia, Sverige, Israel og Noreg gjennomført eit allsidig forskingsprogram på ulmeprosessar. Prosjektet har vekselverka tett med industriverksemder med ulming som potensielt problem. Dette har gitt kompetanseoppbyggjing og eit sterkt og varig fokus på ulming for alle deltakarane i prosjektet. Gode og varige samarbeidsrelasjonar internasjonalt er bygde opp. EFFEKTAR: Nasjonalt har to viktige forskingsmiljø innan brannsikkerhet vore med. Begge vidarefører forsking på ulmebrann etter dette prosjektet. Dette inneber at ulmebrannar vil ha større fokus nasjonalt - på lang sikt - enn det ville hatt utan dette prosjektet, og dette vil spela over på konkret sikkerhetsarbeid. Derfor vil prosjektet ha som langsiktig effekt betre sikkerhetsarbeid nasjonalt i høve til farane ulmebrannar representerer.

We propose a joint project between Stord/Haugesund University College, University of Magdeburg, Germany and Lund University, Sweden - on emerging risks associated with smoldering fires. Smoldering is a flameless form of combustion. It is initiated at re latively low temperatures, operates at intermediate temperatures, and smoke is produced. Smoldering is difficult to detect - and is frequently a precursor of flaming fire (at high temperature) and explosions. Therefore, it represents a hidden danger in many situations: for dwellings, for stored biomass, for powders and grains in silos, for stored industrial waste, for goods under transportation. The smoldering process is complex: It is governed by a self-tuned balance between air entering and smoke leaving the reaction zone, heat generated and heat lost (transported) to the surroundings - and depends in addition on geometry, material chemistry, etc. Furthermore, the distribution among various chemical reactions during the oxidation (smoldering) pro cess is coupled to the temperature and the gas and heat flows. Many materials are prone to smoldering, but properties and crucial parameters are in many cases unknown. New industrial settings with a potential for smoldering keep emerging. In this pro ject, we propose a concerted effort to improve the knowledge both on smoldering mechanisms and on properties of relevant materials: We will study experimentally the influence of geometrical details on the transition from smoldering to flaming fire, and w e will measure characteristic quantities for several materials of importance industrially. Self-ignition is a major concern. Based on project results, we will make systematic predictions for cases that hardly are accessible experimentally, due to sample sizes and time scales. Finally, we will study emergency management both to smouldering fires and worst-case scenarios with transition to extensive fires in large storages and in buildings.