مقاومت در برابر آتش سیستم های کف پوش نسوز قرن 19: تجزیه و تحلیل حساسیت

Typical fireproof flooring systems of the 19th century comprise of metal beams embedded within insulation materials that span between them, sometimes in the form of arches. The limited or non-existent fire resistance requirement of that era demands a thorough understanding of their structural fire response when dealing with their conservation. This requires suitable material property models. Historical records from different sources contain large variations in the thermal (insulation and metals) and mechanical (for metals) properties of the materials. In this research, the variations were placed within lower and upper boundary curves. A sensitivity study of the thermal behaviour of typical flooring systems was conducted. The results of this study were used to indicate the level of uncertainty in the thermal properties of the metals (cast iron, wrought iron and mild steel) and the “insulation” materials (“early concrete” and masonry) that may be tolerated without introducing large inaccuracy (>10%) in the structural temperature results. To assess the applicability of the proposed boundary curves for the mechanical properties of the metals, the second series of sensitivity analyses of structural performance was performed, using the temperature profiles from the thermal sensitivity study.

In typical 19th century fireproof flooring systems, metal beams (made out of cast iron, wrought iron or mild steel) were used as the main load-bearing elements. To provide the necessary fire protection, a common practice was to encase the metal members of the floor with insulation materials (“early concrete” or masonry). It should be noted that “early concrete” was made from various constituents (broken brick, crushed tile, cinders, limestone etc.) different from those in modern concrete production. The two predominant types of such flooring systems are the “filler joist” (Fig. 1a) and the “arch jack” (Fig. 1b) floors. In a typical “filler joist”, wrought iron or mild steel girders are completely encased in “early concrete”. The “arch jack” floor commonly consists of asymmetric cast iron girders embedded in “early concrete” and masonry in an arch form supported by the lower flanges of the metal beams which remain unprotected. Full-size image (12 K) Fig. 1. Typical (a) “Filler joist” and (b) “Arch jack” floor [1]. Figure options In order to investigate the behaviour of 19th century fireproof flooring systems under fire exposure and evaluate their fire resistance, it is critical to establish the thermal and mechanical properties of the constituent materials (metal beam and insulation) at elevated temperatures. The authors [2] have recently presented the results of a thorough literature review which has yielded an extensive experimental database of the required thermal and mechanical properties. The collected results showed that considerable scatter existed in some cases. However, it can be argued that if the structural performance of the metal beams is not sensitive to the scatter, it would be acceptable to use some nominal (such as the mean) values for the material properties. Whether or not this is the case can be answered by performing a sensitivity study and this is the overall aim of the this paper. Specifically, this research will investigate the sensitivity of the fire resistance performance of the metal beams to variations in the following relevant material properties: (1) Thermal properties (thermal conductivity, thermal capacitance = specific heat ∗ density) of the metals. (2) Thermal properties of the insulation materials. (3) Mechanical properties of the metals. The variations in the relevant properties collected from literature by the authors [2] are represented by the lower bound and upper bound values fitted to the collected data. These boundary curves (usually straight lines) provide an envelope to the collected experimental data. Because the collected data were from different sources, some dating back a long time ago, they do not always follow a consistent pattern. Hence some boundary curves contain spikes within some specific temperature regions. The criteria for deciding sensitivity of the fire resistance performance results to variations in the material properties are: for the thermal properties, the fire performance of the metal beams is considered not sensitive to the variations of the properties if the calculated structural temperature does not vary by more than 10% from the mean value; for the mechanical properties of the metals, the fire performance of the metal beams is considered not sensitive to the properties if the load carrying capacity of the metal beam does not vary by more than 10% from the mean value.

The thermal and mechanical properties of the metals used in 19th century fireproof flooring-systems have different degrees of uncertainty. However, uncertainties in some of these properties in most cases have only minor influence on the structural performance of the flooring systems. Based on the sensitivity study results, the following recommendations may be made: • Due to heavy insulation, the structural behaviour of this type of construction is mainly influenced by thermal bowing. Therefore, the lower bound thermal conductivity values should be used to maximise the temperature gradients. These values for cast iron, wrought iron and old mild steel are given in Table 1. However, since the average curves for wrought iron and mild steel and the upper bound curve of cast iron (Fig. 2a) are higher than that of steel, the steel values can be used. The specific heat of steel in EN 1993-1-2 can also be used for these metals. • The structural behaviour of the fireproof floors was found to be sensitive to variations in the mechanical properties (coefficient of thermal expansion, stress–strain–temperature relationships). Because of the high temperature gradients in this type of construction, thermal expansion coefficient is a particularly important value influencing the beam deflections. The stiffness of the metals has a relatively minor influence. On the other hand, the ultimate tensile stress of the metals is highly influential. Due to the large scatter in historic data for these two properties, it is not possible to identify one unique value for each. One possibility is to select one set of values for the thermal expansion coefficient and the stress–strain–temperature curves (e.g. based on those in EN 1993-1-2 for steel), but carry out reliability based simulations to establish appropriate material partial safety factors to ensure that the probability of failure is within acceptable limits. This is now being pursued by the authors. • Although there may be considerable variations in the thermal properties of the insulating materials from different sources, these variations have only minor effects on the load carrying capacities of the structure in 19th century fireproofing structural systems because of the heavy insulations to the metal sections. Consequently, the elevated temperature thermal properties of the insulating materials commonly encountered in such 19th century fireproof flooring-systems can be sufficiently described by the relevant expressions given in the Eurocodes for similar contemporary materials. • As further simplification, the thermal properties of concrete in EN1992-1-2 (specific heat and lower bound thermal conductivity) can be used for the insulating materials of both fireproof flooring systems studied here.