There has been a number of fires in silos used for storing wood pellets in recent years. An example of a large, difficult to extinguish fire of this type was that which occurred in Härnösand in the autumn of 2004, involving fires in three of five 35 m high silos, each containing about 5000 m³ of wood pellets. The fire had started by spontaneous combustion.
Fires in silos are normally difficult to extinguish, can expose the fire-fighters to unpredictable risks and are difficult to assess. SP has previously investigated present knowledge and existing methods of fighting silo fires, described in Report no. SP-AR 2004:16 (only in Swedish). The report describes the dangers of silo fires and reveals the lack of knowledge of how to tackle such fires.
Silo tests at SP
Experiments have recently been carried out in a model scale silo at SP, loaded with wood pellets, in order to improve knowledge of methods of tackling silo fires. The project was financed by the Swedish Fire Research Board, the Thermal Engineering Research Institute, Lantmännen and the insurance company, If. The experiments were extended and complemented by detailed measurements of emitted substances in order to provide a better basis for detection of such fires. This additional work was financed by the Swedish Energy Agency (STEM) through the CECOST centre of expertise.
The model silo that was constructed for the test, was made of concrete, with a height of 6 m and a 1 m internal diameter. For each test, it was filled with about 2.4 tonnes of wood pellets, to a height of about 5 m. A local heat source was used to establish a smouldering fire, to represent spontaneous ignition. This smouldering fire was then allowed to spread inside the silo until a metre wide and deep section of the contents was pyrolysed. The growth of the pyrolysis zone was monitored by about 100 thermoelements positioned inside the silo. The gases evolved, both in the body of the pellets and at the top of the silo, were continuously analysed.
Each test took several days to carry out: not only the spread of pyrolysis, but also the subsequent extinguishing, are slow processes. Four tests were carried out, including investigation of smothering the fire, using both carbon dioxide and nitrogen.
Extinguishing a silo fire requires the correct tactics
Experimental investigations of extinguishing tactics for silo fires have never before been carried out using such a relatively large test facility. It is therefore satisfactory to be able to report that the extinguishing tests were successful and that, for example, the repeatability of growth of the pyrolysis phase before extinguishing started was very high. In each of the tests, the pyrolysis zone migrated downwards in the silo from where it had started, with a wave of moisture and pyrolysis gases moving slowly upwards. At the time of writing this article, the results from the tests have not been completely evaluated, but it is possible to give some general observations on detection and extinguishing:
- Early detection is desirable, but difficult to achieve using conventional methods. By the time that smoke or elevated concentrations of (for example) carbon monoxide are detected at the top of the silo, the pyrolysis process has probably been going on for several days.
- The best method of extinguishment is to smother the fire with inert gas, using carbon dioxide or nitrogen (in the gas phase) delivered as close to the bottom of the silo as possible.
- Water must not be used, as this causes the material to swell considerably, presenting a risk of bursting the silo.
- Seal all openings and leaks in the silo as tightly as possible. Each leak increases the intensity of the pyrolysis process, making the situation more difficult to tackle. The worst possible approach is to attempt to open and empty a burning silo.
- Start controlled emptying when (for example) temperature measurements and gas analyses indicate that the intensity of pyrolysis has been substantially reduced.
- Expect the process to take a long time (days or weeks).
The results from the tests have been published in i SP rapport 2006:47 (in Swedish).
The RISE institutes SP, Swedish ICT and Innventia are merging in order to create a unified institute sector and become a stronger innovation partner. At the beginning 2017 we changed our name to RISE.