With more than a thousand fires a year, prisons in England and Wales need very specific means to suppress fires and allow the safe but secure evacuation of cells. Kelvin Annable reports on some ground breaking research.

Prisoners housed in cells are different from most other occupants of buildings in the event of a fire, as they are unable to evacuate themselves away from the heat and smoke generated. Currently, there is only a small number of cells fitted with automatic fire detection systems – most fires are therefore usually detected either by automatic detection systems external to the cells, or directly by prisoners or prison warder staff. Firefighting operations are currently conducted manually by staff applying water from fixed hose reel systems through the inundation point in the door to a cell.

In 2007 the Ministry of Justice commissioned BRE Global to help develop new safety strategies for the fire protection of prison cells. The primary focus of the work was an experimental evaluation to investigate the effectiveness of water mist suppression systems in tackling cell fires. Additionally, testing was conducted to assess the potential fire size possible in a cell with a high fuel loading.

There is a significant number of fires each year in prison cells and most are started deliberately. The Ministry of Justice supplied BRE Global with the following statistics:

• 139 prisons, six more being constructed
• Approximate 85,000 offender population
• About 1200 fires annually; many deliberately started
• About 250 offenders or staff officers attend hospital each year suffering burns or smoke inhalation
• It has been estimated that the fire service only attends around 25% of incidents
• There have been two deaths in the last five years; 19 deaths in the last 29 years.

Since the introduction of the Regulatory Reform (Fire Safety) Order 2005 (FSO) in October 2006, prisons have had to comply with the legislation (which covers ‘all relevant persons’). Previously all Crown prisons were exempt from all fire safety statutory duties.

Experimental programme
Due to the size and diversity of prison buildings, a ‘tool kit’ of fire protection options was required. It was recognised that different systems would be of particular benefit for particular establishments. The following fire protection systems were considered and evaluated in the BRE Global burn hall:

• In-cell detection systems
• Traditional water hose reels for use through a cell door
• Fixed ‘institutional’ sprinkler system with heat activated automatic operation of frangible glass bulbs
• Fixed water mist system (with both heat activated automatic operation and actuation linked to in-cell detection systems and triggered via a control panel )
• Manual water mist systems for use through a cell door.

In the tests, the suppression systems needed to be capable of effectively suppressing a fire and maintaining tenable conditions for cell occupants for a 20 minute period, and prison warder staff for a 31⁄2 minute ‘snatch rescue’ period.

The following suppression system attributes were recognised as generally desirable for the prison estate:

• Low water flow rate systems with limited total water quantity (due to the prison cell population being close to maximum levels, the potential for quick cell re-instatement after a fire is a major issue)
• Low system pressure which could be maintained ‘in-house’
• Water mist discharge characteristics with good ‘total flooding’ capability (cells sizes and geometries vary significantly)
• Economic, reliable, ergonomic, and easy to use, install and retrofit
• Robust to malicious tampering

This was a challenging and potentially contradictory list and it was therefore recognised that other types of systems – for example high pressure water mist – could also be appropriately used within the ‘tool kit’ of options.

Tests conducted
BRE Global conducted a series of 29 fire tests in a test ‘cell’ measuring 3m x 4m x 3m high, including:

• A ‘baseline’ fire scenario with no suppression system (this fire scenario was then subsequently used for further tests)
• A cell ‘burnout’ test with a high fuel loading to assess the potential fire size and heat release (with the cell door left open)
• Eight ‘feasibility tests’ on identified ‘tool kit’ options
• 19 tests with industry provided commercial systems (12 different system providers).

The cell burnout test with a high fuel load and open door resulted in a large fire

The BRE Global fire scenario selected was intended to be a ‘bad’ but also ‘typical’ fire. The materials used were not considered to present a worst case in terms of toxic combustion product release (there was a relatively low level of plastic materials used). There was a significant degree of shielding from direct water discharge spray for both fixed and manual systems.

The water mist systems submitted to BRE Global for evaluation represented a truly international market. The range of systems tested was extensive and comprised low, medium and high pressure water mist for both fixed in-cell systems and manual systems for use through a cell door. Two different types of ‘twin fluid’ system – which use high pressure compressed air to generate a very fine mist of small water droplets utilising a very low water flow rate – were also evaluated.

One of the key considerations was the assessment of toxic conditions after a fire and operation of a suppression system. The main toxic hazards in fire effluent atmospheres consist of smoke irritants (particulates and gases) and asphyxiant gases. In order to assess human tenability, the hazards from exposure to asphyxiant toxic gases and heat (hazard from convection due to contact with hot fire gases) were analysed using Fractional Effective Dose (FED) methodology.

The asphyxiant gases considered were carbon monoxide (CO), hydrogen cyanide (HCN), carbon dioxide (CO2) and low oxygen hypoxia (O2). These were measured at both mid level (approximately standing head height) and low level (representative of an occupant lying on the floor). The combined effects of these gases were estimated according to the method of Purser. The method is incorporated in a current British Standard (BS 7899-2: 1999 Code of practice for assessment of hazard to life and health from fire. Guidance on methods for the quantification of hazards to life and health and estimation of time to incapacitation and death in fires) and an International Standard (ISO 13571:2007 Life-threatening components of fire – Guidelines for the estimation of time available for escape using fire data).

The calculation method is used to predict the time at which a person subjected to toxic conditions would become incapacitated (FED equals 1) and also the time to reach fatal conditions (FED equals 2). For FED heat a value of 1 is a level described as ‘extreme pain’ and FED value of 2 is described as ‘severe burns, possibly death’. An example graph is shown in figure 1.

Figure 1: Example of FED graph

Findings
The cell ‘burnout’ test with a high fuel loading and open cell door resulted in a severe and large fire. Heat release rate measurements indicated a peak heat release in excess of 2.5 MW. As can be seen in figure 2, the heat release rate may not have peaked at the time that the fire was manually extinguished.










 

Figure 2: Heat release measurement from cell fuel loading test

The BRE Global developed fire scenario was burnt in the test room with no suppression system operation. Highly hazardous conditions resulted for both cell occupants and staff rescuers and conditions became life threatening after approximately 10 minutes.

All the suppression systems tested improved room tenability (asphyxia and heat conditions) compared with the unsuppressed fire. All but one of the systems submitted maintained survivable conditions in the test room for a period of 20 minutes, and tenable conditions for staff rescuers for a 31⁄2 minute period.

Low water flow manual systems, with both partially shielded and, in one instance, highly shielded fires, were demonstrated to provide effective fire suppression and maintain tenable conditions for a 20 minute period.

Very fine mist (very small water droplet size, understood to be of the order 25-40 microns), high pressure mist (of the order 100 microns) and low pressure mist (of the order of 250 microns and larger) systems were demonstrated in the test series to provide effective fire suppression and tenability control (note that BRE Global did not verify any of the actual droplet sizes). Each of the different types of systems had their own desirable attributes. However, testing has demonstrated that concerns remain regarding the ability of very fine mist systems to remove heat (due to a lack of a ‘wetting effect’) and prevent shielded fire re-ignition or development after operation is terminated.

The suppression systems tested ranged significantly in the amount of water discharged; this range was from 9% to 70% of the water used by the institutional sprinkler tested, or from approximately 5l/min to 40l/min (the institutional sprinkler used around 57 l/min).
It was generally observed in all tests that, when the cell door was opened, there was limited, or very limited, visibility in the cell. This would affect the ability of staff to carry out occupant rescue.

In many tests the gas conditions at low level deteriorated after the operation of water mist systems, as smoke and buoyant combustion gases were ‘brought down’ to floor level. However this finding needs to be considered in the context of the demonstrated fire suppression (reduction in fire size and heat providing improved tenability generally).

For low pressure systems, observations of ‘cold discharge’ water spray patterns indicated that there were very distinct areas within the cell open area where there was only a very low water coverage and low levels of water mist in the air. This was much less so for the high pressure systems, where the entire room appeared to be filled with a fine mist.

Of particular interest were a number of tests where it was noted, subsequent to the actual test from cold water discharge testing, that there was very little water reaching the fire location. In these tests the fire began to increase in size (see figure 3) but was then suddenly suppressed or extinguished a few minutes later (whilst oxygen levels were still reasonably high and capable of sustaining combustion). This was likely to have been due to the steam produced by a larger fire and room humidity conditions not supporting flaming combustion. These test results indicate that it is difficult for fires to increase above a certain size in the presence of water mist in limited volume cells with limited ventilation.

 

Figure 3: Temperatures in the test room

Potential problems with the installation of water mist systems for prison cell use were observed, with instances of nozzles being incorrectly installed or becoming partially blocked and pump units not supplying the ‘correct’ or specified water pressure.

A range of nozzles were submitted to BRE Global for evaluation, some of which provided ostensibly robust anti-ligature and anti-tamper design. Others, for the test programme itself, provided what appeared to be an obvious ligature point and some were clearly vulnerable to malicious tampering.

Recommendations
The work undertaken in the project demonstrated that both fixed and manual water mist systems can provide an effective means for suppressing cell fires and maintaining tenable conditions, both for cell occupants and staff rescuers. BRE Global therefore recommends that the Ministry of Justice considers providing water mist systems, suitably specified and proven by fire testing, for the protection of prison cells.

It is almost inevitable that a cell occupant will suffer some degree of exposure to irritant smoke during any cell fire, but it is not inevitable that this will result in serious injury. However, both from a workplace ‘health and safety’ perspective, and in order for staff to be able to perform an effective rescue, it is likely that staff entering a cell during a fire incident would benefit from personal protective equipment. This would be defined based on an appropriate risk assessment but is likely to include some degree of respiratory protection (such as a particulate filter mask), and eye protection, especially if they might encounter such situations on a repeated basis.

The Ministry of Justice should consider the implications, for prison cell construction and management, of the cell fuel loading test which resulted in a fire size of approximately 2.5 MW.

The fire test scenario had relatively low levels of plastic materials

Performance specification
Subsequent to the initial research work, BRE Global was commissioned by the Ministry of Justice to produce a performance specification document. This is based primarily on draft British water mist standards (primarily DD8489) but also on publicly available international standard documents. The performance specification stipulates requirements for manual water mist systems specifically for custodial establishments relating to:




 

• Fire testing
• Detection
• Actuation
• Control
• Design and installation
• Components
• Water supplies
• Pressurisation systems
• Commissioning
• Inspection and maintenance
• Operability
• System documentation

The way forward
BRE understands from the Ministry of Justice that it is intended to introduce water mist systems for all ‘new builds’ and all major refurbishments, and also in existing buildings where exceptional requirements – such as for segregation units – have been identified. Many in the prison service have embraced the technology and are keen to extol the virtues of its ability to knock down fires and heat in the specific environment of a custodial establishment. Because the amount of water used is low, it is unlikely to have any adverse effect on the in-cell electrical systems, and the cell is unlikely to be put out of action for any significant period of time.

The programme of work conducted was an unparalleled detailed assessment of a diverse range of internationally provided water mist fire suppression systems for a particular application. The work has been of interest internationally and it may be possible for many of the findings to be extrapolated to prison environments in other countries. The project has provided the Ministry of Justice with a clear evidence base and accompanying documentation for appropriate specification and installation of water mist systems for prisons.

Kelvin Annable is senior consultant at BRE Global. The work reported on was carried out under a contract placed by the Ministry of Justice and this article is published with permission. Any views expressed are not necessarily those of the Ministry of Justice. The author wishes to acknowledge the assistance given by many individuals and organisations, and in particular the suppliers who contributed systems for assessment.