Mitigating the Impact of Post-Earthquake Fires on Steel Structures

By Gabriel–Victor Risco,¹ Luisa Giuliani,² Varvara Zania²

Structural engineering design has been continuously challenged by natural disasters occurring on a random basis, with little to no possibility of prevention. Moreover, an equilibrium must be reached between a robust design and an economical solution for all construction projects. In addition to load combinations for normal usage of the building (serviceability and ultimate limit state), accidental load combinations have been introduced in design codes to account for unforeseen events such as earthquakes, explosions, fires, hurricanes, etc.

Several historic events in highly seismic active areas (i.e., California, Tokyo) have shown that structures affected by earthquake before a fire are more likely to experience a degradation of the fire response, as opposed to an undamaged structure. These high-intensity earthquakes followed by fires result in numerous structural failures and both material damage and loss of human lives. Scawthorn¹  has documented most of these post-earthquake fires and the resulting damage.

The Eurocode Standard, which is used in most European countries, offers different load combinations covering accidental actions, but does not consider the effect of two consequent actions. The extent to which an earthquake can affect the fire response of a structure and how this impact can be mitigated is still unclear to designers.

Post-earthquake fires have been the cause of structural collapse of both reinforced concrete and steel structures. The latter has been of more interest, since concrete generally possesses sufficient inherent fire resistance, without needing additional insulation. Multiple scientific studies^2-4  have investigated the effect of post-earthquake fires on uninsulated steel structures, most of them concluding that the reduction in fire resistance is rather insignificant, if the structure is designed to an earthquake of intensity similar to the one applied. However, it should be stressed that the majority of steel structures are designed with insulated elements, which results in different responses of the structures when affected by shaking.

 Research on insulated structures^5-6 has considered partial insulation loss from both beams or columns. The results indicated that the seismic impact should be carefully considered because a larger effect was noticed. Since both of these studies were investigating the loss of spray-applied insulating material, it was seen as necessary to look at insulating panel boards, which may detach and expose larger unprotected areas of the steel surface.⁸

 We carried out more-recent numerical investigations^7-8 in an attempt to extend these findings by looking into the post-earthquake fire response of steel frames, with and without fire insulation, further and comparing it to the resistance of the undamaged structure to fire. There is no clear differentiation between the behaviour of insulated vs. uninsulated identical frames, so a comparison of the two is appropriate.

This study used a commercial FEM software to model two moment-resisting steel frames of different heights (5 and 10 floors respectively), first as uninsulated frames and then by assuming a fire protection designed to ensure a 120 min fire rating of the elements (R120), in accordance with what is prescribed by the Danish regulation for buildings taller than 12 m.⁹ We applied nine earthquake accelerograms on the frame and selected two of them for the post-earthquake analysis, on the basis of the largest inter-storey drifts recorded and largest strain.

We considered several fire scenarios, depending on the location and extension assumed for the fire and a standard fire exposure¹⁰  for the structural elements affected by the fire scenario. For each scenario, we assessed and compared the fire resistance of the structure first to fire only and then to post-earthquake fire, using two criteria to quantify the fire resistance: the time of fire exposure at which the first structural element fails and the time of fire exposure at which a major collapse occurs.

While the first is closer to the design criterion of common structural fire design, the second allows observing the difference of the structural behaviour in term of collapse mode and not just resistance time, thus favouring a deeper understanding of the influence of the seismic damage on the fire response.

In this respect, the analysis of the uninsulated frames⁷ was deemed most useful for investigating the intrinsic response of the structure to post-earthquake fire, while the analysis of the insulated frame is more realistic, but dependent on the insulation and the damage assumed to affect it.

It was assumed that post-earthquake fire resistance is affected by either the steel material properties degrading, due to cyclic loading or residual stresses induced in the elements resulting from the plastic hinges developed during the seismic motion. The former is assumed to be insignificant in properly designed buildings to current seismic standards.²

The investigations carried out on the uninsulated frame showed a rather limited effect of design earthquakes on fire resistance. In particular, the reduction in global failure times was less than 3%, while the resistance in terms of single element failure was reduced up to 17%. The rather insignificant reduction was attributed to the drift limitation imposed by the Eurocode design criteria, which is in accordance with the results of previous literature.^2-4

In terms of failure modes, a similar progression of the collapse was noted in most of the scenarios examined. The heated elements elongate due to the rising temperature, leading to failure of the beams in the first instance, followed after by the columns. A difference was observed in case of fires affecting the entire ground floor of the tall structure: the columns failed first, leading to sudden collapse of the structure.

The analysis of the insulated frames⁸ was conducted by assuming plasterboard insulating panels boxing the I-profiles of both columns and beams. One scenario for the uninsulated frames applied fire within one lateral ground-floor compartment (one bay with a width of 6 m and a height of 4.5 m). Different percentages of insulation loss were simulated on one exterior lateral column in the fire compartment by reducing the thermal resistance of the insulation from 100% (fully insulated column) to 0% (no insulation on the column). We chose this scenario based on the longer columns and higher load-to-resistance ratios, while the lateral column has a lower degree of restraint. To simplify the model, the insulation degradation was assumed to occur uniformly on the entire column length.

The results show a clear reduction of fire resistance, unlike where no insulation is present. For example, insulation damage corresponding to 50% reduction of the thermal resistance of the insulation anticipates structural collapse in about 40–50 minutes in the fully insulated case. This decrease is even more significant in the case of a full insulation loss on the column, where a decrease of approximately 80 minutes was observed.

Regarding the collapse mode, the progression tends to follow the same trend as with uninsulated frames. However, for large amounts of insulation loss, the collapse mode changes result in a sudden failure: The column temperature increases rapidly, while the beam does not undergo large displacements that would act as a warning, thus sudden failure occurs. This is considered to lead to global collapse of the frame.

Based on these results, it seems reasonable to assume that the main cause of premature structural failure can be attributed to the loss of insulation. It can be seen that early collapse can be caused even by the insulation loss of a single element, which therefore must be avoided.

Furthermore, further studies are needed to correlate the amount of damage to the deformation endured and potentially impose limitations on displacements with respect to the insulating material applied. These could then be introduced in the standard seismic design. An appropriate example is the existing drift restriction for non-structural damage (e.g., curtain walls). This work assumed that insulation would be damaged by either detachment of the panel or failure of its support.

Intumescent paint is another concern, because it may crack or detach more easily with column deformations. Different amounts of steel would be exposed compared to the panel insulation.

The sudden failure observed in certain scenarios (e.g., fire affecting the ground floor) is a concern that has not been highlighted before and poses a risk for the safety of occupants or firefighting and rescue crews, who might enter the building after evacuation has been performed.

Gabriel-Victor Risco is a graduate fire engineer with Trenton Fire Ltd., London, United Kingdom. Luisa Giuliani and Varvara Zania are associate professors in the Department of Civil Engineering, Technical University of Denmark, Kongens Lyngby, Denmark.

References

¹Scawthorn, C., Edinger, J.M., and Schiff, A.J. Fire Following Earthquake. 2005. American Society of Civil Engineers Publications.

²Della Corte, G., Landolfo R., and Mazzolani, F.M. 2003. Post-earthquake fire resistance of moment resisting steel frames. Fire Safety Journal, 38(7):593–612.

³Zaharia, R., and Pintea, D. 2009. Fire after earthquake analysis of steel moment resisting frames. International Journal of Steel Structures, 9(4):275–284.

⁴Jelinek, T., Zania, V., Giuliani, L. 2017. Post-earthquake fire resistance of steel buildings. Journal of Constructional Steel Research, 138, 774–782.

⁵Arablouei, A. and Kodur, V. 2016. Effect of fire insulation delamination on structural performance of steel structures during fire following an earthquake or an explosion. Fire Safety Journal, 84(1):40–49.

⁶Tomecek, D.V., and Milke, J.A. 1993. A study of the effect of partial loss of protection on the fire resistance osf steel columns, Fire Technology, 29(1):3–21.

⁷Risco, G., Giuliani, L., and Zania, V. 2018. Failure mechanism of steel frames subjected to post-earthquake fires, in A. Nadjai, F. Ali, J-M. Franssen, O. Vassart, The Proceedings of the 10^th International Conference on Structures in Fire, 803–810. Belfast, UK.

⁸Risco, G., Giuliani, L., and Zania, V. 2018. Numerical Study on the Effect of Post-Earthquake Fires on the Resistance of Fire-Insulated Steel Frames. The Proceedings of the Nordic Fire and Safety Days Conference, Trondheim, Norway.

⁹2012. Eksempelsamling om brandsikring af byggeri, Klima-, Energi-og Bygningsministeriet Energistyrelsen.

¹⁰ISO 834-1.1999. Fire resistance tests — Elements of building construction — Part 1: General requirements for fire resistance testing. Switzerland: International Organization for Standardization (ISO).