A Comparison of the Fire Performance of Prestressed Concrete Double T-beams using Design Provisions and SAFIRThe Station Nightclub Fire Evaluation
CE 808: Structural Fire Engineering
FridayMonday, AprilFebruary 814, 2008
Nickolas Hatinger
Megan Vivian
23
Table of Contents
I. Introduction 1
II. Literature Review 2
A. General 2
B. Experimental Studies 2
C. Analytical Studies 6
III. Structural Models 8
IV. Design Codes 10
A. General 10
B. American 11
C. Canadian 12
D. Eurocode 13
V. SAFIR 14
VI. Results/Discussion ……………………………………………………16
VII. Conclusions 17
VIII. Acknowledgements 18
IX. References 20
APPENDIX A: Sample Calculations 22
APPENDIX B: SAFIR Input Files 40
I. Introduction
II. Literature Review
A. General
B. Experimental Studies
C. Analytical Studies 5
III. Structural Models
IV. Design Codes
A. General
B. American 10
C. Canadian 11
D. EuroCode
V. SAFIR
VI. Results
VII. Conclusions 2
VIII. Acknowledgements 3
IX. References 4
I. Introduction 1
II. The Station Nightclub Fire 1
A. Structural and Material Features 1
B. Ignition of Fire 2
C. Fire Growth 2
D. Fire Protection Features 3
E. Fire Service Response 3
F. Structural Effects Due to Fire 3
G. Aftermath 4
III. Lessons to be Learned 4
A. NIST Investigations 4
B. Human Behavior 5
C. Recommendations 5
IV. Conclusions 6
V. References 7
23
Abstract
The fire performance of two prestressed concrete double T-beams were assessed using the tabulated data and simplified calculations under the American, Canadian, and European design provisions, as well as with a performance-based approach, SAFIR. A 10DT24+2 and 12DT32+2 simply supported beams were selected as representative beams typically incorporated into retail stores or parking structures, with spans of 40 and 50 ft., respectively. The design provisions coincided well with one another, but the tabulated data fire ratings under the Canadian code proved to be slightly conservative. Similarly, the SAFIR results were consistent with the design provisions for the 10DT24+2, but diverged slightly for the 12DT32+2 and were contributed to lack of available temperature distributions for 90 minutes. Overall the prescriptive-based design provisions provided reasonable results, but it is been suggested that American and Canadian codes incorporate an advanced calculation method to promote accurate and efficient fire safety designs.The purpose of this paper is to present a case study to compare current structural fire safety design provisions in the United States, Canada, and Europe to finite elements methods using SAFIR.The purpose of this paper is to present an overview of The Station nightclub fire that led to death of 100 and injured 200 patrons on February 20, 2003. The overview consists of a description the building’s features and materials, fire ignition and growth, fire protection systems, response of emergency units, and aftermath of the fire. In addition, it highlights the several simulations and tests conducted by the NIST investigation to identify any negligence in safety of the nightclub and further work that could be done to ensure safety among similar public structures. It was conclude that fire severity or the event as whole could have been avoided if proper building materials, sprinkler systems, and considerations of egress would have been accounted for in the buildings design.
23
IntroductionIntroduction
Fires are among earthquakes, hurricanes, floods, tornadoes, and blasts as one of the most destructive forces subjected to a structure. In 2006, fire’s caused 3,245 civilian fatalities, 16,400 civilian injuries, and $11,307,000,000 in property loss throughout the United States, alone. A total of 524,000 structural fires contributed to $9.6 billion dollars in property damage (Karter, 2007). The devastation and unpredictability of such catastrophes is the reason why design provisions demand structures to meet minimal fire safety requirements.
The philosophy of most design provisions around the world is to ensure life safety and minimize property damage in the event of a fire. Commonly this is accomplished through a prescriptive-based approach, which assigns structural elements fire ratings through tabulated data derived from standard fire tests. Although widely practiced it is limited in scope and restrictive in application. Therefore, countries such as New Zealand, United Kingdom, Australia, Japan, Sweden, other European countries have moved towards a performance-based methodology. The new method is intended to capture the complexities of the event through the introduction of realistic fire scenarios (variation in fuel loads and ventilation), loading regimes, and restraint, in attempt to develop a rational and economic design alternative.
The purpose of this paper is to present a case study to compare the fire performance of two precast prestressed concrete double T-beams through various methods. The methods consist of design provisions from American, Canadian, and European codes, as well as a finite-element methods using SAFIR to predict the members’ fire response. The goals of the study are to illustrate the limitations of the prescriptive-based approach and to promote the performance-based movement in an attempt to accurately and efficiently predict the fire response of structural elements.
The following paper consists of seven main sections. First, is the introduction discussed previously. Second, is the literature review, which identifies similar analytical studies conducted in the past. In addition, a few experimental studies to explore how the current provisions were established for the fire performance of prestressed concrete beams. In the third section, a description of the structural models and loading are presented. Overviews of the relevant codes are discussed in the fourth section. The computer program SAFIR, used in the comparison, is detailed in the fifth section. Sixth, are the results of the comparison. Finally, in the seventh section the relevant conclusions are addressed.
In the Unites States alone, fires contribute to 3,245 civilian fire fatalities and approximately $11,307,000,000 in property loss in 2006. Of this, about 524,000 fires took place in structures; which accounted for $9.6 billion dollars in property damage (Karter, 2007). A very small percentage of these structural fires typically occur in nightclubs or similar venues, yet these occurrences lead to remarkably high numbers of casualties.
One such structural fire, on the evening of February 20, 2003, at The Station nightclub in Warwick, Rhode Island, took the lives of 100 civilians (FEMA, 2004). The pyrotechnic display from the nightly entertainment created a spark causing ignition of surrounding foam insulation and subsequently engulfed the entire building in flames. Due to overcrowding and panicked egression of The Station, patrons were unable to evacuate from the quick spreading fire and many perished that evening. Investigations have been completed to determine any negligence in safety of the nightclub and further work that can be done to ensure safety among similar public structures.
I. Literature Review
General
In the following discussion a series of experimental studies are reviewed to briefly illustrate how the current prescriptive-based provisions were developed from standard fire tests for prestressed concrete beams. In conclusion, the similarities among the studies are highlighted and their limitations are identified. Subsequently, several analytical studies are presented to provide similar research findings.
Experimental Studies
In general, most fire tests of precast prestressed concrete beams have been performed to establish prescriptive-based codes under standard fire exposure. Most of the ratings depend on concrete cover, cross-sectional area, reinforcement axis distance, restraint, or insulation thickness. Many of these investigations were used not only to establish design provisions, but also to validate models and identify prevalent failure modes under elevated temperatures. The following are a series of fire test conducted to establish current prescriptive-based design codes in the United States.
Gustaferro and Carlson (1962) compiled 50 standard fire tests on precast prestressed concrete building components and conducted tests on a variety beams and slabs to determine the factors which affect fire resistance. The tests were conducted from 1953 to 1961 collectively by the National Bureau of Standards, Underwriters Laboratories, Portland Cement Association, and Fire Prevention Research Institute. An assortment of span lengths, insulation thicknesses, aggregate types and cross sections of members such as I-shaped, double-tee, and single-tee beams, as well as flat hollow-core, solid, and stemmed floor assemblies were tested in accordance with ASTM E 119. The factors which affected the fire resistance of the precast prestressed components were concrete cover, degree of restraint, cross-sectional area, member geometry, aggregate type, and concrete moisture content. The main factors which affected the strand temperature were concrete cover, cross-sectional area, member shape, type of aggregate and insulating protection. Although, all of the details of restraint were not provided for each study, it proved to significantly increase the member’s fire resistance and typically failed due to heat transmission. Conversely, the failure mode for unrestrained beams primarily was due to prestressing strands exceeding their critical temperatures (850 to 950°F). Other key observations were that lightweight aggregates provided better fire resistance than normal weight aggregates. The moisture content only became an issue once 70% relative humidity was exceeded. Similarly, the vermiculite insulation has proven to provide greater fire ratings as long as it remains bonded to the concrete. Finally, a comparison of one studies results tabulated in Table 1 were deemed consistent with the other tests, but was considered slightly conservative.
Table 1: Cover for various fire resistances (Gustaferro and Carlson 1962).
Selvaggio and Carlson (1964) performed a study on the influence of aggregate type and load intensity on the fire resistance of twenty two prestressed concrete I-beams. The beams were tested under standard fire tests with three-sided exposure and spanned 20 ft. Three different normal (dolomitic or siliceous) and lightweight (expanded shale’s) aggregates were used. Two loading intensities were uniformly applied through different live and dead load combinations. The results yielded, lightweight aggregates exhibited the best fire performance and were remarkably similar. The dolomitic aggregates exhibited the best fire performance of the normal weight aggregates, through delayed heat transmission affects. Lightweight aggregate beams failed in a compressive manner, while the normal weight aggregate beams failed in tension. When considering varying load intensities, the heavier loading regime failed quicker with greater midspan deflections. A few beams were restrained and revealed a 22% better fire performance than simply supported beams. In addition, less thrust was observed in the expanded shale aggregates compared to the normal weight aggregates.
Abrams and Gustaferro (1972) conducted fire endurance tests of four prestressed concrete double T-beams with spray-applied insulation. Two different cross sections were tested under the ASTM E119 standard fire with unrestrained support conditions. A control specimen and two more beams with ½ and 1 in. of vermiculite acoustical plastic insulation, consisted of one cross section. The other beam section used ½ in. mineral fiber insulation. Their fire endurances were 1 hr. 2 min., 1 hr. 50 min., 3 hr. 6 min., and 2 hr. 28 min., respectively. Both types of insulation maintained adhesion throughout the tests. A prescriptive based tabulated approach was suggested for 2 and 3 hr. ratings, based on stem width, concrete cover, type and thickness of insulation.
The majority of the fire tests for the prestressed concrete beams were performed in the 1960’s and 1970’s. Although these results are invaluable in determining many of the factors affecting the members’ fire performance they are limited to standard fire exposure and are not applicable to other potential fire scenarios. Other issues are raised when considering the variability in size of the test specimens, because full-scale specimens fail in different mechanisms than large specimens. In addition, although some of the studies included restraint, it is only vaguely understood to improve the fire resistance without any rational understanding to quantify such a scenario. Therefore, crude tabulated fire ratings are assigned to a multitude of beams based on handful of fire tests. The tabulated data limits a designer’s ability to push the envelope for new designs and prevents new materials to be used without costly fire tests to assure fire ratings.
Analytical Studies
While reviewing the literature, no specific case studies were found that compared fire ratings among different countries design provisions. However, several papers were discovered in which finite-element methods were used to accurately estimate the fire performance of prestressed concrete beams. The methods used in the studies are examples of the types of procedures that are regularly used in a performance-based methodology. The following are significant findings of each study.
Anderson and Laurisden (1999) undertook a comparison of prescribed fire ratings to a series of calculations for prestressed concrete double T-beams roof assemblies. Four participants (DTI, PJK, DTU, and FSD) used different calculation methods to determine the fire ratings. Three used finite element programs (FEM), while the other performed simplified calculation methods. The fire endurance calculations and tests were in accordance with standard fire test method ISO 834. Three beams were tested and modeled. The beams differed through their location of the prestressing steel from the bottom flange (57, 73, and 83 mm). However, the second specimen temperature-time curve was greater than ISO 834; therefore it was excluded from the results. Table 2 illustrates the results.
Table 2: Analysis method and fire test results (Anderson and Laurisden 1999).
All three beams failed initially to bond failure followed by shear fracture. Spalling was observed in all three tests on the bottom surface of the slab between the stems. The average fire resistance time calculated for type 1 and 3 beams were 36 and 6% greater than the actual fire tests. Although the moment calculations were similar for DTI, DTU, and PJK, the shear and anchorage failure calculations were scattered. The variations of the results are attributed to different material property reductions and bond models incorporated in the different methods.
Franssen and Bruls (1997) performed a state-of-art analysis on a precast prestressed concrete double-T beam to verify the results of a previous experiment under the ISO 834 standard fire exposure. The manufacturer prescribed a 2 hr. fire rating for the beams based solely on its ultimate bending capacity; however the test specimen failed in 79 min. Therefore, the authors conducted a nonlinear thermo-mechanical analysis, using a computer program called SAFIR, to validate the test. The bending capacity analysis resulted in a fire rating of 91 min. and led to the conclusion that the beam failed in a shear failure mode. The ultimate shear resistance was calculated using the method proposed in Eurocode 2-Part 1-1. The study considered the contributions of the steel stirrups, prestressing tendons, and concrete to the beams shear strength. As a result, the calculations coincided very well with the tests at 80 min. However, the beam did not satisfy code requirements; therefore the beam was modified in SAFIR to provide a fire resistance for shear and bending moment of 135 and 130 min., respectively. A subsequent fire test yielded a fire endurance of 144 min.