Temperature distribution of intumescent coated steel framed connection at elevated temperature.

Transcription

1 NSCC2009 Temperature distribution of intumescent coated steel framed connection at elevated temperature. R. R. Krishnamoorthy 1 & C.G. Bailey 2 1PhD Student, Extreme Loading & Design Group, The University of Manchester, UK 2 Professor of Structural Engineering, Head of MACE School, The University of Manchester, UK ABSTRACT: A series of ten small-scale experimental works was conducted in a gas furnace. Temperatures were measured in a number of protected and unprotected steel members at various locations. This paper compares the test data against the predicted values computed by finite element analysis package, ABAQUS version 6.7. The paper considers the geometry of the members, the exposure conditions and the thermal property of the fire protection material. The thermal predictions using ABAQUS agreed well for unprotected members, however, in protected members the results were reasonable good although it did not exhibit a good match in the initial time-temperature stage. This mainly attributes to the thermal behavior of the intumescent coat. 572

2 1 INTRODUCTION 1.1 Steel in Fire The use of naked steel elements in both structural and architectural is growing incessantly. Nevertheless, with its ability to deform under huge stress and resisting fire up to 550ºC, steel became the choice of modern architecture, mainly for aesthetic appearance, easy erection, availability in various steel sections, and most important of all, it s fire resistant capabilities. The latter became one of the prime emphases in modern structures to withstand aggressive fire at elevated temperatures. The drawback is that, naked structural steel members may need additional fire protection. The type of fire protection selected depends on the fire resistance requirements, expected use, environment, appearance, and cost. Conventionally, the performance of structural elements has been assessed in a range of standard fire tests [1]. The assessments were based on individual elements based on ISO 834 standard fire curve. The performance based approach involves the assessment of three basic components comprising the likely fire behavior, heat transfer to the structure and the structural response [2]. 1.2 Intumescent Coating Over two decades, intumescent paints have been available in Europe. With their capability to resist fire at high temperatures, it is widely used in the United States, UK and European countries. The products are broadly similar in terms of testing, performance and applicability among manufacturers. Intumescent paints have two key components: a resin binder and a mixture of chemicals that decomposes and releases a gas when heated. During a fire, the material melts. A gasproducing reaction is triggered at a temperature corresponding to an appropriate resin melt viscosity, and the release of gas causes the resin melt to foam developing an insulating layer. This then produces a thick char, which insulates the steel from fire. Intumescents may typically expand approximately 15 times to 40 times their initial thickness during a standard fire test. Intumescent coatings comprise ingredients such as catalyst which produces mineral acid during decomposition period, carbonific substance which produces char once combines with the mineral acid, resin to soften the intumescent at elevated temperature and finally spumific agent which release gas during chemical reaction [3]. Solvent-based intumescents are typically used for exterior applications, and are tested against weather and temperature variations. They are also used for interior applications. Water based intumescents have less odor; however they are less tolerant of humidity and low temperatures. Intumescents are available in liquid form and are typically applied via airless spray equipment. Smaller areas may be rolled or brushed, however this often does not leave as smooth of a finish. The required thickness of paint is dependent on the size of the structural element (i.e., structural elements with larger, heavier cross sections may require less insulation than lightweight members). Thickness of the applied intumescent paint materials is typically 0.5 mm to a few mm but can be as much as 5 mm. The two-part epoxy system is typically used in more harsh environments, including the chemical industry and offshore operations, in areas that may be difficult to access for maintenance, or where high levels of impact damage may occur, and they are more expensive than the other intumescent paints. Additionally, they perform well with regard to hydrocarbon type fires. They are not grouped with the thinner-film intumescent materials due to their epoxy binder; however, the behavior during a fire is similar. The char formed however, is thinner, though mechanically it is much stronger in order to withstand the higher heat flux and erosive gases. For the purpose of this research, the technical specifications of intumescent coating were obtained from a well known intumescent manufacturer in UK. 573

3 2 THERMAL PROPERTIES The thermal properties considered are thermal conductivity, specific heat, density. Moisture content of 3% assumed for fire protection material. Intumescent type fire protection exhibits varying thermal properties at elevated temperature. The thermal conductivity of intumescent coating were computed using experiment data from X.H Dai et al [4] and is given by with and (1) where: A p /V is the section factor for steel members insulated by intumescent coat A p is the area of intumescent per unit length of the member [m 2 /m] V is the volume of the member per unit length [m 3 /m] c a is the temperature dependent specific heat of steel [J/kgK] c p is the temperature independent specific heat of intumescent coat [J/kgK] d p is the thickness of intumescent (m) t is the time interval(s) a,t is the steel temperature at time t [ºC] a,t is the temperature increase of insulated member g,t is the ambient gas temperature at time t [ºC] g,t is the increase of ambient gas temperature during the time interval t [K] p is the thermal conductivity of intumescent coat [W/mK] a is the mass of steel [kg/m 3 ] p is the mass of intumescent [kg/m 3 ] Figure 1. Thermal conductivity of ten different test specimens and the average curve. 574

4 Figure 2. Thermal conductivity of ten different test specimens from temperature range 150ºC-850ºC Table 1. Thermal properties of thin film intumescent coating. Temperature Density (ºC) (kg/m 3 ) Independent Specific Heat (J/kgK) A proposed equation to compute the thermal conductivity of intumescent with equivalent material properties can be adapted based on the test results. The relative intumescent thermal conductivity can be determined from the following: - for 20 C p 350 C : = -2x10-8 p 3 + 1x10-5 p p for 350 C 1000 C : = 3x10-10 p 3 + 5x10-8 p p Where intumescent thermal conductivity p intumescent temperature 2.1 Test Data Test data for model validation were obtained from X.H Dai et al [4]. The test data for ten test specimens for coated and unloaded steel are SP1, SP2, SP3, SP4, SP5, SP6, SP7, SP8, SP9 and SP10. The thicknesses of coating varied considerably from 1.02mm-1.43mm on beams and on columns. The thicknesses were recorded manually using a digital meter and it is named as dry film thickness (dft). Each specimen s dft s will be shown in the respective figures below. Thermal conductivity of intumescent coating varies significantly at elevated temperature and an average mean value must be found before the final thermal conductivity value can be used for analysis purpose in ABAQUS. Since it is impossible at this moment to measure the intumescent coat temperature using thermocouple, a simplified approach to determine the temperature was done by averaging the temperature of furnace and the temperature of steel, both which are measurable using the thermocouples. The thermal conductivity is then computed using equation (1). The final thermal conductivity value that will be used for analysis would be the average each thermal conductivity specimen. 575

5 3 HEAT TRANSFER ANALYSIS Average thermal conductivity (Figure 1 and 2) values and the ISO 834 standard fire exposure were implemented in ABAQUS [4] for the transient heat transfer analysis to determine the thermal response of intumescent protected steelwork. For this research, the convection coefficient, h c, for the gas furnace fire according to ISO is taken as 25 W/m 2 K [6]. The intumescent coating temperature is taken as the average value of its exposed surface temperature and the steel temperature which were measured by thermocouples. The surface radiation emissivity assumed is 0.8.The heat transfer analysis for both protected and naked steel was implemented using the two and three-dimensional finite element models. The two-dimensional models were developed on the basis of the steel cross section for both fully protected and naked steels, and the three-dimension models were developed for partially protected steel members. A 4-node linear heat transfer quadrilateral, DC2D4 and an 8-node linear heat transfer brickdc3d8 were engaged for the two models respectively. Figure 3. Steel beam with partial intumescent coating protection at the web and bottom flange section. The type and response of simulation used was transient heat transfer analysis. The surface exposed to the fire were carefully selected and interacted. The time period for fire duration is set at one hour. The minimum increment size used was and maximum of 100. The maximum allowable temperature change per increment is 10 ºC. A time-temperature relationship between test data and ABAQUS using the average thermal conductivity is shown to illustrate the behavior of intumescent coat at elevated temperature. Figure 4: Test result validation of unprotected steel temperature against ABAQUS version

6 Figure 5: Time-temperature test and computed, SP1 (Average beam dft =1.11mm) Figure 6: Time-temperature test and computed, SP2 (Average beam dft =1.02mm) Figure 7: Time-temperature test and computed, SP3 (Average column dft = 0.60mm) Figure 8: Time-temperature test and computed, SP4 (Average beam dft =1.28mm) Figure 9: Time-temperature test and computed, SP5 (Average beam dft =1.18mm) Figure10: Time-temperature test and computed, SP6 (Average column dft = 0.62mm) Figure11: Time-temperature test and computed, SP7 (Average beam dft =1.22mm) Figure 12: Time-temperature test and computed, SP8 (Average beam dft =1.28mm) 577

7 Figure13: Time-temperature test and computed, SP9 (Average beam dft =1.25mm) Figure14: Time-temperature test and computed, SP10 (Average beam dft =1.23mm) 4 SUMMARY AND DISCUSSION The predicted temperature rise of the protected bottom flange and web closely matches the test result. The initial rise of temperature is less predicted but the error band decreases to the test result after 600s. This discrepancy that appears in Figure 5 to 14 may be attributed to few reasons. The first reason will be associated to the assumption that intumescent temperature which were averaged throughout the burning process to determine the thermal conductivity. Secondly, during the test, the 200mm concrete topping covering the beam s top flange acts as a heat sink, thus absorbing the heat energy from the beam during heating and causing constant heat loss. Temperature dependant density loss which is assumed is another reason to be considered. Taking into consideration these reasons, with a marginal discrepancy with test result, it can be summarized that the numerical heat transfer model is valid. 5 REFERENCES [1] Fire Tests on Building Materials and Structures, BS 476. British Standards Institution, [2] Fire Saf. J. 41 (7) (2006), pp [3] Structural Fire Design: Off-site Applied Thin Film Intumescent Coatings, The Steel Construction Institute,1996 [4] X.H. Dai, Y.C. Wang and C.G. Bailey, Temperature distribution in unprotected steel connections in fire, Fire Saf. J. 44 (3) (2009), pp [5] ABAQUS. ABAQUS/Standard user s manual, vols. I-III and ABAQUS CAE Version 6.8 [6] EC1, Eurocode 1: Actions on Structures-Part 1-2:General Actions-Actions on Structures Exposed to Fire. [7] Y.Kang, G.V Hadjisophocleous and H.A Khoo, Effect of partial loss of spray-on protection on load capacity of steel beams during standard fire, Fire Protection J. (Vol.18-February 2008) [8] D.V. Tomecek and J.A. Milke, A study of the effect of partial loss of protection on the fire resistance of steel columns, Fire Technology 29 (1) (1993), pp [9] Wang, W.-Y. and G.-Q. Li (2009). "Behavior of steel columns in a fire with partial damage to 578

8 fire protection." Journal of Constructional Steel Research 65(6): [10] Omrane, A., Y. C. Wang, et al. (2007). "Intumescent coating surface temperature measurement in a cone calorimeter using laser-induced phosphorescence." Fire Safety Journal 42(1): [11] Kodur, V. K. R. and M. M. S. Dwaikat (2009). "Response of steel beam-columns exposed to fire." Engineering Structures 31(2): [12] Y.C. Wang, Steel and Composite Structures, Behaviour and Design for fire Safety, Spon Press, London(2002) [13] Wald, F., L. Simões da Silva, et al. (2006). Experimental behaviour of a steel structure under natural fire. Fire Safety Journal. 41: [14] Foster, S., M. Chladna, et al. (2007). "Thermal and structural behaviour of a full-scale composite building subject to a severe compartment fire." Fire Safety Journal 42(3): [15] Wong M.B. and Ghojel J.I., Sensitivity Analysis of Heat Transfer Formulations for Insulated Structural Steel Components, Journal of Fire Safety, Vol. 38, pp , [16] Y.C Wang, Fire Safety Design of Structures, The University of Manchester, United Kingdom. [17] Bailey C.G. One Stop Shop in Structural fire Engineering. The University of Manchester [18] Moore, D.B. and Lennon, T. (1997), Fire engineering design of steel structures, Progress in Structural Engineering and Materials,1, No.1, 4-9 [19] T.Lennon, D.B. Moore, Y.C.Wang and C.G.Bailey, Designers guide to EN , EN , EN and EN [20] C. G. Bailey et al The design of steel framed buildings without applied fire protection. The Structural Engineer, Volume 76/No 5, 3 March