Fire prevention through oxygen reduction at Forsmark19 May 2014
Faced by new regulations relating partly to fire safety, Forsmark NPP has brought an industrial safety system into a nuclear environment. The oxygen reduction (or hypoxic air) system reduces the oxygen in protected environments to eliminate the risk of combustion. By Patrik Lundgren, Susanna Backström and Stefano Chiti.
Certain areas at Forsmark Unit 1 and 2 do not comply with new regulations from the Swedish Radiation Safety Authority regarding physical separation of components. One purpose of the physical distance is to make sure that if a fire occurs, it will not be able to spread to a redundant system in the same room. To meet these new regulations, fire protection in these rooms needs to be improved. Initially, conventional fire suppression systems like gas extinguishing systems were considered. However, during the early system design phase, risks were identified: they would require complex installation in rooms with sensitive equipment, and could pose risks to personal safety. So alternative solutions were considered. The use of an oxygen reduction system (ORS) was one of them.
Oxygen reduction systems
ORSs, also known as hypoxic air fire prevention systems, are able to prevent fires by reducing the oxygen concentration in the protected enclosure to a level where ignition and flame-spreading via common materials is impossible. Common applications for ORS are data centres, electrical appliance rooms, large automated storage and archives.
ORS was first used in the 1990s, but it has recently become more popular. Unlike traditional fire suppression systems, which detect and then extinguish a fire, ORS employs a semi-active approach, which reduces the oxygen available for combustion and ignition. Usually 5.5 - 6.5% of the oxygen in the protected enclosure is replaced by nitrogen to produce and maintain an atmosphere comprising 14.5-15.5 vol.% of oxygen and 85.5-84.5 vol.% of nitrogen. This is a similar oxygen partial pressure as air at an altitude of around 2400-2900 m, but there the ratio of gases is the same as at sea level (referred to as normobaric) so ignition and combustion can occur.
Generally speaking, if there is an opening to air in the protected enclosure the oxygen level is kept constant either by overflow from adjoining rooms or by replenishing the enclosure using OR generators. Even if a breach is so large (say, a door or gate open), that it exceeds the replenish capacity, many hours might pass before common combustible material can ignite inside the compartment. Ignition is also prevented for several days in the event that all OR generators fail, but there is no breach of integrity to the outside. Even when the oxygen level reaches a level where ignition occurs (about 17%), the heat release rate and rate of fire spread is still severely reduced.
At 14.5-15.5 vol. % there is no health hazard for healthy people and personnel can safely perform simple tasks. However, national regulations typically limit the time exposure of personnel in the protected enclosure to several hours, according to the current draft of the European standard on ORS, TC 191 WI 00191236 (2013).
To determine what level of oxygen concentration provided the required level of fire protection, ad hoc fire tests were performed. The fire tests were based on the method described in oxygen-reduction standards VdS 3527en 2007 and BSI PAS 95:2011.
SP, the Technical Research Institute of Sweden, performed the tests. A third-party pre-evaluation study concluded that for a more conservative fire scenario, instead of testing single components, the fire tests should be carried out with the components assembled in a model representing the actual geometry and relation between them. The ignition source, in this case an oxy-acetylene torch, must be applied to the components in the most favourable position for fire spreading.
The materials used in the fire tests were spare components from Forsmark representative of typical materials found in the rooms to be protected: relays, signal and power cables, and circuit boards. The components included different plastic materials, for example polycarbonates and PVC.
The fire tests were performed in a range of oxygen concentrations from 20.9% (normal air) to 15.0%. During the tests, measurements of oxygen levels, burning times, mass reduction and adiabatic surface temperatures were taken, smoke gases were analysed and visual observations and video recordings were made. Analysis of the results found that below 17% oxygen the risk of fire spreading was reduced. Based on that analysis, and adding a safety margin, the target value for the oxygen content was set at 15%.
To reduce the oxygen concentration in the protected rooms, around 6% of oxygen is replaced by nitrogen. A full-scale test was performed to verify whether the ventilation system was able to distribute nitrogen evenly to the protected rooms. Oxygen sensors were installed to monitor oxygen concentration from the main control room. A temporary nitrogen injection system was used to lower oxygen concentration in the rooms to 15%. The test showed that the ventilation system was suitable for implementation of an ORS. (The test was also used to verify the placement of the oxygen sensors).
At Forsmark 1&2, there are two fire compartments consisting of five rooms that are protected by ORS. The two fire compartments in each unit have an identical layout and are located on opposite sides of the control room. The rooms are distributed over three stories and each fire compartment has a total volume of approximately 3200 m3.
A pre-study concluded that an ORS would be able to prevent a fire from spreading to a redundant system in the same room. It also found that the existing ventilation system was suitable for implementing an ORS: the ventilation system is common for all the rooms in the same fire compartment and no adjustments to the ventilation system were necessary.
The oxygen concentration in the rooms is reduced by injecting nitrogen into the ventilation system and maintained at that level by setting the ventilation system to recirculation mode. Tests were performed to ensure that the rooms maintained a good environment with regard to temperature, CO2 level, moisture content and material emissions. The test showed no negative effects. The CO2 level and temperature are continuously monitored. Procedures were also introduced to measure the ozone level in the rooms before and after implementation of the ORS.
Oxygen sensors installed in the rooms monitor the oxygen concentration and control the operation of the nitrogen generator so that an average oxygen concentration of 15% is maintained in the protected areas. These sensors trigger an alarm if the oxygen concentration goes above or below the set values. To control the operation of the nitrogen generator, one oxygen sensor is installed per room, with the exception of one room where the layout makes two sensors necessary to ensure accurate measurements.
An additional oxygen sensor is installed in every room to monitor concentration for health and safety purposes. These sensors trigger an automatic shutdown of the nitrogen generator if the oxygen concentration goes below a set value determined by the lowest allowed concentration. They also trigger audio and visual alarms advising evacuation of the rooms.
The nitrogen is produced by a membrane-based generator. The existing compressed air system at Forsmark provides the necessary compressed air to the generator, which splits the compressed air into two different flows: nitrogen-enriched air with about 95% nitrogen and 5% oxygen, which is injected into the protected rooms, and oxygen-enriched air which is discharged to atmosphere. The pressure and flow of the compressed air feed are monitored, as are the flow and oxygen content of the nitrogen enriched air. These values are needed to monitor the performance of the nitrogen generator.
The existing compressed air system was tested with regards to oil content and dew point to ensure that the compressed air is of sufficient quality for use in a nitrogen generator. The membranes inside the nitrogen generator are protected by integrated carbon filters to further ensure their lifetime.
The nitrogen flow level was determined by measuring the leakage rate of the rooms. After initial leakage tests performed with tracer gas, measures were taken to seal the rooms and make them more airtight: the critical areas were doors and wall penetrations. Leakage tests were then performed again and the results were used to determine the required nitrogen flow.
After the system was installed, the nitrogen generator was adjusted to find the optimal working mode with regards to system availability and compressed air consumption. As a consequence, the nitrogen flow was reduced, to lower the consumption of compressed air while maintaining system availability. This is beneficial as there are other process systems that rely on the availability of the same compressed air system. The availability of the system has proven to be relatively high: most of the outages were for planned activities and not system faults.
Employees need to enter the rooms where the oxygen reduction system is in place on a regular basis, to take measurements, do routine tests, perform diagnostics and maintain components. Typical working times are usually 5-15 minutes.
In Sweden, there is currently no general rule for human occupation in spaces with an ORS system. However, a provision from the Swedish Work Environment Authority requires breathing apparatus in confined spaces with oxygen levels below 18%. At Forsmark, it was deemed that breathing apparatus would be impractical for the frequent short visits to these areas. To apply for an exemption, Forsmark sent technical reports, a risk analysis and an early project description, together with literature studies regarding the effects of low oxygen environments on people, to the Swedish Work Environment Authority.
The Swedish Work Environment Authority granted Forsmark an exemption allowing the workers to be exposed to a minimum oxygen concentration of 14.5%, but with a number of restrictions. Before anyone is allowed to enter the rooms, a medical examination must be undertaken to verify that the heart, circulatory and respiratory functions are normal. Working hours are restricted to three two-hour shifts per day and capped at 24 hours per week. Minors, smokers and pregnant women are not covered by the exemption. All the work performed is logged.
About the authors
Patrik Lundgren Forsmark Kraftgrupp AB project manager; Susanna Backström, Vattenfall AB process engineer; Stefano Chiti fire safety specialist at COWI AS