The fire risk tool

Background

Experiences of museums, libraries and archives - which carried out a risk assessment indicate that fires, despite of significant technical progress in the field of fire protection, is among the greatest risks to long-term preservation of collections. Managers responsible for preservation of collections should also take into account fact that the priority of fire protection resulting from applicable legal regulations and standards is primarily protection of human life and health, and then building.

To address the lack of an appropriate methodology of selecting an optimal and cost-effective strategy for preservation of collections, Jean Tétreault of the Canadian Conservation Institute has developed a method of quantitative assessment of the risks caused by fires in museums, archives and libraries, taking into account the statistical frequency of fires and their estimated effects. The analysis, which Tétreault based on statistical data from ​​Canada, was published in 2008 and is the basis of this tool assessing risks caused by fires [ref. 1]. He demonstrated that both the incidence of fire, its extent and the resulting material damage depended primarily on the type and effectiveness of the fire protection used in the institution. Tétreault proposed six control levels for fire risk (CL). Reaching given CL in a building depends on meeting a number of requirements in six areas: Prevention (location and purpose of the building), Blocking (use of certain passive fire protection measures), Detection (use of certain active fire detection measures), Firefighting activities, Procedures and Training of facility personnel. CL 1, 2 and 3 apply to buildings without automatic extinguishing systems (AES), CL 4 applies to buildings where AES is installed in rooms with high accumulation of flammable materials, and CL 5 and 6 apply to buildings where AES is installed in all rooms.

In the quantitative assessment of fire risk, the occurrence of fires and their anticipated effects should be taken into account. Mathematically, fire risk is expressed as the product of the probable number of fires from the present to the assumed future point called the time horizon and its effects, which are understood as the percentage loss of value of the collection.

The impact of fire depends on several factors:

The formula defining magnitude of risk for a given group of objects takes form:

\( MR \) - magnitude of risk expressed as the expected loss of collection value. It is assumed that the value of the collection is currently 1. \( MR \) will be a fraction of this value. The greater the risk, the greater the loss of value,

\( P \) - likehood of fire in the selected time horizon,

\( j \) - summation index for three factors causing damage: fire - combustion and high temperature, smoke and water as a result of the extinguishing operation,

\( k \) - summation index for fires of various sizes: limited to material of origin, a room, a level or covering the entire building,

\( f(V_j) \) - correction function depending on material sensitivity \( V_j \) to various agents,

\( EF_k \) - contribution of fires of various sizes,

\( PM_{kj} \) - part of the material (collections and other museum equipment) affected by the fire for fires of various extent and for various agents,

\( ML_{kj} \) -loss of the material caused by combustion and high temperature, or by smoke or the effects of fire extinguishing action causing material damage equivalent to the loss of material, for fires of varying sizes,

\( PC_{kj} \) - the probability that it is the collections and not other material which is affected by the effects of fires of various sizes, for various agents.

Likehood of fire (\(P\)) and Distribution of extent of fire (\(EF\))

The likehood of fire (probable number of fires) is estimated as a ratio of the chosen time horizon and the average period between fires per an institution preserving collections in your country, area or organization. The choice of different time horizons affects the magnitude of the risk. Table 1 shows the average time between occurrences of fires in Canadian museums given in ref. 1. These periods depended on the museum's control level.

The conclusions of such analysis will not be exact as it does not take into account the likelihood that, in the chosen time horizon, a fire will break out two or maybe three times. On the other hand, if the chosen time horizon were 200 years, the probable number of fires is greater than 1 and it needs to be taken into account that objects damaged at the first fire cannot lose their value again during the second fire. The above formula does not take into account such cumulative effects and therefore should be used to estimate risk magnitude only for time horizons shorter than the average period between fires. Nevertheless, such an estimate is sufficient for heritage managers. If risk analysis for museum collections is conducted in the perspective of 30 or 50 years and the probable number of fires is more than one, it should be a ‘red light’ for managers of the collection indicating that solutions reducing the risk of fire should be immediately implemented.

Fractional shares of fires of various extents – in terms of fire spread - in the total number of fires, can be derived from the fire statistics for institutions preserving collections in your country, area or organization. The fire extents are defined as limited to material of origin, a room, a level or spreading to the entire building. The fractional shares must add up to 1. They are different for six different control levels for fire risk. Fractional shares for Canadian museums are provided in Table 2 [ref. 1] .

Loss of the material due to combustion and high temperature

To start with, the loss of material of medium sensitivity to combustion and high temperature is estimated depending on the size of the fire and the control level for fire risk. As the control level increases, the fire will be extinguished faster and the material loss will be less. The relevant data for Canadian museums are provided in Table 3 [ref. 1].

The effects of fire depend also on the sensitivity of objects to various agents - fire, smoke and water. The ML values in Table 3 were estimated for materials of medium sensitivity to combustion and high temperature. The formula for the risk magnitude MR also includes a correction of values estimated for materials of medium sensitivity depending on the real sensitivity of objects to various factors \( f (V_j) \), listed in Table 4 after ref. 1. In this tool, the correction in the form \( (ML_{medium Sensitivity}) ^\wedge (1 / (f(V_{fire})^{1/2}) \) was applied. For materials with high damage sensitivity, the proposed corrections will increase ML to 1 (total loss of material), and for decreasing sensitivity the parameter will asymptotically approach 0 (no loss of material), while for medium sensitivity it will take the value from Table 3.

When considering the effect of fire, it is necessary to estimate what part of the material PM in the museum (objects and equipment) will be covered by the fire, i.e. in this case the combustion and high temperature. When estimating the consequences of fire covering the entire building, the PM is 1. Assuming a simple building layout, one can estimate the PM for the fire covering one level by dividing 1 by the number of levels, and for a room by the number of rooms. However, in the case of a fire limited to material of origin, it was assumed that the PM result for the room should be divided by 10.

In addition, the effect of the fire depends on the probability PC that heritage objects will be within its range. Tétreault reported that in Canadian museums only 3% of all fires began in rooms containing collections. This means that if the fire is limited to a room, then the probability that the objects will be within the fire range is 3%. In turn, it is known that if the fire covers the entire building, then the probability is 1. It is assumed that if the fire is limited to material, the probability that objects will be within it is 0.15%. However, if the fire covers one level then the probability that there will be objects within it is calculated by dividing the number of levels on which the collections are located by the number of levels in the building.

1. Material loss due to impact of smoke and water

In the proposed method, it is reasonable to assume that equipment damaged by fire cannot be destroyed a second time because of flooding or smoke. Therefore, the material loss resulting from the action of smoke as well as water should be estimated after deducting the part of material affected by fire (\(PM_{kj}\) - \(PM_{k,fire}\)).

In the event of a fire covering the entire building, the \(PM\) is 1, i.e. it is assumed that the fire will affect all the museum's equipment. Similarly, the entire building will be affected by smoke and water. However, damage from smoke or water is calculated only for the material that survived after exposure to combustion and high temperature, which as shown in the previous section depends on CL and material sensitivity.

When calculating the fractional material loss for a fire covering a level, first of all the loss of material due to combustion and high temperature should be deducted. In contrast to fire, whose operation is limited to the level, the smoke will also spread over the upper levels, and the water will flood the lower levels than those covered by the fire. For simplicity, it was assumed that for a statistical museum, smoke and water would affect half of the building, including the level covered by the fire.

where \(PM_{level,fire}\) = 1 / number of levels.

In the event of a fire limited to a room, it was assumed that the water effect would be limited to that room. However, the impact of smoke will depend on CL. In CL1, CL2 and CL3, there is no smoke-proof door or automatic air-conditioning switch, which will cause smoke to spread throughout the entire level.

\( PM_{room,water}\) or CL4, CL5, CL6 \( PM_{room,smoke} = (PM_{room,fire}– ML_{room,fire})* f(V_{fire}) * PM_{room,fire})\)
CL1, CL2, CL3 \( PM_{room,smoke} = (PM_{level,fire} – ML_{room,fire}* f(V_{fire}) * PM_{room,fire}) \)

When calculating the fractional material loss for a fire limited to material, first of all the loss of material due to combustion and high temperature should be deducted. Unlike fire, whose operation is limited to material, smoke and water will affect the whole room.

Material damages are additionally limited by systems and methods of storing collections or objects. Estimated material damage resulting from the effects of fire extinguishing action and smoke will be significantly smaller, if it is taken into account that some objects can be placed in showcases, or they can be framed or wrapped in foil. Frequently used solutions are shelves and wardrobes of high tightness, well protecting objects stored in them against water. Large objects are also often packaged in waterproof fabric. When preparing a fire risk analysis for a specific museum, this type of individual solution can be taken into account by introducing into the tool a parameter ‘part of the collection showcased’ expressed as a fraction.

As described earlier, the effect of fire depends on the probability PC that heritage objects will be within its reach. PC will have the same values ​​regardless of damaging agent and was given in the previous section.

In turn, the assumed categorization of water sensitivity of materials are presented in Table 5.

In turn, the loss of material of medium sensitivity to water was estimated to be 0.005, 0.05, 0.1, 0.2, respectively, for increasing size of the fire. However, material damage resulting from the firefighting operation will not depend on CL. It is recognized that regardless of how the fire is extinguished (using AES or by Fire Brigade units) a comparable amount of water is needed to extinguish it.

In contrast to fire and water, it is assumed that all surfaces of objects are damaged to a similar extent as a result of dirt associated with the deposition of smoke particles. This assumption is certainly not true for all objects, but due to the lack of relevant data it was decided to take it as a certain approximation.

For the purposes of this study and further analysis, it is also assumed that material damage resulting from smoke will be proportional to those in Table 3, because the amount of smoke produced during a fire depends on its size (including materials that have been burned) and the speed at which the fire will be extinguished. The proportionality factor has been chosen in such a way that the loss of value of the collection due to smoke, combustion and high temperature is 2/14.

References:

Tétreault J. 2008. Fire Risk Assessment for Collections in Museums. Journal of the Canadian Association for Conservation (J. CAC), Volume 33, 3-21.