THe OKLAHOMA City BOMBING: ANALYSIS OF BLAST DAMAGE TO THE MurRAH BUILDING By Paul F. Mlakar Sr.,' W. Gene Corley,’ Mete A. Sozer Charles H. Thornton, Fellows, ASCE Apsrnact: The truck bombing of the Murrah Building caused significant damage to this structure. From the ‘characteristics of the bomb crate, it was determined thatthe explosion yielded energy comparable to that from the detonation of 1,814 kg (4,000 Ibs) of winitotoluene (TNT). The blast directly removed a principal exterior column, and the associated arblast caused the failure of two others. The airblast also destroyed some of the floor slabs in the immediate vicinity. This paper describes the blast loading and its effect on the building. INTRODUCTION ‘The writers represented ASCE on the Building Performance ‘Assessment Team (BPAT) that investigated the collapse ofthe ‘Murrah Building (Corley et al. 1996) In the report of this work ("The Oklahoma"? 1996), the design, construction, and Condition of the Murrah Building prior to the tragic bombing hnas been described in detail. The structure was a nine-story reinforced concrete frame with three rows of columns spaced at 6.1 m (20 ft) within each row. A large transfer girder at the third floor permitted the elimination of alternate exterior col- ‘umns below, The building was designed and constructed in accordance with the applicable codes, but did not provide any deliberate resistance against a vehicular bomb attack. Addi- tional information is given in companion papers (Corley etal 1998; Sozen et al. 1998). “This paper describes the estimation of the blast loading and. its direct effect on the structure ofthe building. First, the blast loading is inferred from the properties of the erater formed by the explosion. Then, the response of critical structural elements to this loading is computed using approximate methods ap- propriate to the purpose of the assessment. These include the principal exterior columns supporting the transfer girder and the floor slabs of the building Further implications of these direct responses are examined ina companion paper (Sozen etal. 1998), Of particular interest is the integrity of the structural frame with these damaged columns. Measures 10 mitigate ths situation in other buildings are discussed in another companion paper (Corley etal. 1998). BLAST LOADING ‘The explosive device was contained in an enclosed truck parked on the paved street along the north side of the Murrah Building (N.W. Fifth Steet). The blast caused the collapse of part of the nine-story portion of the Murrah Building, as well fs damage to a number of nearby structures. In this section, the effects of the blast are quantified for engineering analysis. FFor the purposes of this report, the quantification is performed with methods used for the analysis of conventional weapons effects on structures ("Fundamentals”” 1986), ‘The calculation of blast loading begins with the estimation "Div. Chk, US Army Engi Winey. Experiment Station, Vicksburg, Ms. ‘ice Pres. Const Technol. Lab, In, Skokie, I 60077-1030 2ketehut Distinguished Prof, Purdue Univ, West Laayete, IN “Pi, Thorton Tomasets Engrs, New York, NY. [Note Discussion open until January 1, 1999. Separate discussions sould be submited for ue individual papers the symposium. To ‘extend the closing date one mont, a writen request mist be led with {he ASCE Manager of Journals The mandserp for this paper was sb Ite for review and posable pubicaton oa February 19, 1998. Tas paver is pat ofthe Journal of Performance of Consruced Facies, Nol 12, No.3, August, 1998, ©ASCE, ISSN O887-382898/0008.0113~ (0119158.00 ~ 5-30 per page. Paper NO. 17777 of the yield or quantity of explosives detonated. For bursts near the ground surface, the yield or quantity is usually in- ferred from the dimensions of the crater formed (see Fig. 1). In this report, the engineering survey of the crater forms the basis of this inference ‘The crater was approximately 8.5 m (28 fo in diameter and 2.1 m (68 ft) in depth, as shown in Fig. 2. The center is about 2.1 m7 ft) east and 4.3 m (14 f) north of Column G20, as shown in Fig. 3. According to the design drawings and ob- servations on site the thickness ofthe pavement was 460 mm (18 in.) and the underlying soil was dry, sandy clay. From information about the truck reported to have contained the explosive device, the center of the explosive is estimated to hhave been 14 m (4.5 ft) above the ground surface, as shown in Fig. 2 AAs listed in Table 1, the detonation of a spherical charge of winitrotoluene (TNT, the standard by which the energy of var- ious explosives is measured) weighing approximately 1,800 kg. (4,000 bs) at 1.4 m (4.5 f) above 460-mm- (18-in) thick pavement on Soil results ina crater whose dimensions are con- fistent with those measured at the Murrah Building site. In ‘Table 1, the crater dimensions for pavement on soil are an wverage of those for massive concrete and for the dry, sandy clay alone, weighted in proportion to the depths of the two materials inthe crater. This weighting is substantiated by the results of ongoing research concerning craters in pavements based on soil ‘Such a detonation also produces an airblast wave that prop- ‘gates radially from the burst point. This wave is characterized by an instantaneous rise to a peak value, termed the incident pressure, from which it decays to ambient conditions. Fig. 4 Shows the contours of this incident pressure comesponding to the surface detonation of 1,800 kg (4,000 lbs) of TNT at the location of the bomb crater. These contours correlate approx- imately withthe level of damage shown for buildings in the neighborhood. However, this damage is also a function of load FIG. 1. Bomb Crater (Covered by Tarp) JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES / AUGUST 1998/113,ee / NN. Fin Stet Sion FIG. 2. Proximity of Column G201o Location of Bomb (Pian View) ‘TABLE 1. Estimate of Vield from Crater Dimensions (“Funda smentals” 1986) Depth [Diameter Consiton (teat) | (eet) o @ | @ ‘ooo pounds of THT on massive concete 26 | ‘$000 pounds of TNT on dy sandy clay | ‘4000 pounds af TNT" inches of pavemeat| ‘on ail 2 | om ‘Mensred at Marah Building os | modification by nearby buildings and the resistance of the buildings themselves. In particular, buildings shown in the up- per part of Fig. 4 were not shielded by the Murrah Building I can be seen that damage in the upper part of the figure is heavier than that in the lower par. (Of particular interest is the loading on the nine-story portion ‘of the Murrah Building, When a blast wave impinges on a ‘structure, a higher pressure is developed, termed the refected pressure. The calculated ("Fundamentals"” 1986) peak over pressures on the north elevation are shown in Fig. 5. These Tange from a maximum of over 69 MPs (10,000 ibs/in’) at the point closest othe detonation to a minimum of 62 kPa Tps/in#) at the upper west comer, with an equivalent uniform value of approximately 965 kPa (140 Ibw/in’). While these pressures are extremely lage, they act for a limited duration, a shown in Fig. 6. The duration ranges from a maximum in the upper west corner to a minimum at the point closest to the ‘last and has an equivalent uniform value fora tiangular pulse ‘of about 5 milliseconds. Any explosive detonation near the ground surface causes a ‘ground shock motion in addition to the aiblast loading (“Fun- damentals” 1986). For the surface detonation of 1,800 kg (4.000 lbs) of TNT, the free-field motion at the center of the building is shown in Fig. 7. There is a rapid rise after arrival {114/,JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES / AUGUST 1998— Building Inspection Area Legend [CE] Ae. aurran Federal Butsng HMI cotapsee Seucure EE) ruc! Damage TS xen Gace Door Cine Fam Pao etre Foot FIG, 5. Peak Overpressures on North Elevation of NineStory Portion of Murrah Bullding to.a peak value of 230 mm/sec (9 in/sec), followed by a grad- ual decay over the next 270 milliseconds. In comparison to the extreme airblast loading, this ground shock motion is of litle consequence to the structure. BLAST ANALYSIS OF ADJACENT COLUMN G20 ‘As shown in Fig. 3, the bomb crater places the explosive device in close proximity to Column G20. The response of structural elements in such proximity is a function ofthe range divided by the cube root of the explosive mass or scaled range. For Column G20, the scaled range is only UW = 4.785 mo 814 kg) = 677 -RIW = 15.6 £4(8,000 1b)" = 1.0 Fvib!* in which R = horizontal range (see Fig. 3); and W = equivalent ‘TNT charge mass, jased on bomb damage reports from the Second World War Effects” 1946), the destruction of frst-story reinforced con- crete columns by the brisant effects of blowing out, severing, and undermining occurs at scaled ranges of 2,032 mnvkg"™ (3 JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES / AUQUST 1996/ 115dase ex Pao Dietance From Pont of Detonation in Feet we PPP ER TERRE? 10 tbo FIG. 6. Duration of Loading (In Miliseconde) on North Elevation of Nine-Story Portion of Murrah Bullding Tie (rican) FIG. 7. Free Fleld Ground Shook at Center of Murrah Bullding fib) from cased charges such as bombs or artillery shells From contemporary research (MeVay 1988) on the difference between the breaching of reinforced concrete walls by uncased and cased charges, the scaled range for destruction by uncased charges is estimated to be 1,016 mim/kg"” (1.5 fub™) Thus, in all Likelihood, Column G20 was abruptly removed by bri- sance. This conclusion is supported by the fact that no one ‘whom the team interviewed found any evidence ofthis column in the debris or in the crater caused by the explosion BLAST ANALYSIS OF NEARBY COLUMNS G26, G16, AND G12 Column G24 was located outside the range of brisance for tuncased charges, but was highly loaded by the detonation. As indicated in Fig. 8, ils response (o this load is approximated ‘as a simply supported beam between the first- and third-floor elevations (Biggs 1964). The column did extend below the first-floor elevation to ils supporting caisson, but it was not Toaded by the blast below this level and received some support from the surrounding soil. The column was also laterally con- nected at the second floor by a transfer strut, 2B-13, but this feature provided litle restraint in the east-west direction ex- cited by the blast. ‘As indicated in Fig. 9, the column resisted this loading about its weak axis. This cross section was 914 mm (36 in) wide by 508 mm (20 in.) deep. It was reinforced with 20 #11 vertical bars and 2 #4 horizontal ties at 406 mm (16 in.) on center. The strength was limited by the shear resistance atthe ‘ends of the column, From the material strengths measured in this study and the estimated axial prestress from dead and actual live loadings this limiting capacity, Vs, corresponds to 359 kPa (52 Ibs/n-) uniformly distributed on the 914 mm (36 in) face Fig. 10 shows the blast loading on Column G24. On the front face, the load rises abruptly to the reflected pressure, 10 FIG. 8 Model of Column G24 (Single Degree of Freedom) Cross Section of Column G24 116 /JOURNAL OF PERFORMANCE OF CONSTAUCTED FACILITIES / AUGUST 1998MPa (1,400 Ibs/in2). When the blast clears this face, it falls TABLE 2._ Blast Response of Intermediate Columns Support- to the stim of incident and dynamic pressures. The later stems !8GNorthTranefer Girder fom ihe parle or wid ey seid wih te ast oeeinerer trave ab i impinges on sroctoes fn ip as ube a sequently arrives at the tear face and rises gradually to the — coe oat ate aie ‘sum of the incident and dynamic pressures at this range and cu} )_|_(@)_} 4) | 6) Grienation, The effective tangular duration of the net lading Sak ange G0 afar] so] SS only about 1 mllisecond Peak pete Couns per sso] | ot say | ais ‘The response of Column G24 to this loading, ilustrated in ps8, caitisecondsy 00 | S600 | gat | 118 Fig Tiss specifically the laeral defection ofthe column at Buenas a: | tz | 0 the midpoint or second-tory level asa function of time mea- Show svopparaine caciy | t8 | + | to | 01 sured after the detonation of the bomb. Notice that most of this response occurs after the net load has diminished to zero, so that i is an impulsive structural event. When this deflection reaches 25 mim (1.0 in), the shear atthe supports, V, exceeds Vi, and the britle failure of the element occurs. As the axial prestress and corresponding shear capacity are greater at the fist floor than at the third floor, this failure is expected atthe top. Immediately after the blast (Fig. 12), the upper portion of this column was missing, in agreement with the results of this, analysis "Column destroyed by trisance Presse (ones pr sci) 5 Tone recon) FIG. 10. Blast Loading of Column G24 5 ie i: A Fa. 14. Column Grint i As shown in Tbe 2 he slat rage (stage rom te ". ° 7° . 7” ey explosion to the column midheight) is greater for Column G16 Ted ty thin for Column Gas, Av th tan of 13 (0%, the Fo. 11, set Response cf cotumn 240 Midpin (Secon ek eure sil # MPa (tT). Acoring to an Floor Elevation) tedahee hed limels analysis similar to that performed for Column G24, this load- ing just reaches the shear capacity. This implies an incipient britle failure, which is consistent with the conditions shown in Fig. 13 However, Column G12 did endure the direct blast effects ‘of the bomb, It was located ata slant range of 27.1 m (89 f0), as indicated in Table 2. Here, the loading was 793 kPa (115 Tpsfin#). The associated response is only 0.1 of the capacity to resist. The results of this analysis are consistent with the {ntact condition of this column after the bombing, as shown in Fig. 14 BLAST ANALYSIS OF SLABS. "The floor slabs in close proximity o the bomb were directly : loaded by the blast. The facade of the north elevation consisted - of 1,520 x 3,050 mm (5 X 10 ft) glass panels restrained by FIG. 12. Damage at Column Line! aluminum channels. This glazing offered insignificant resis JOURNAL OF PERFORMANCE OF CONSTRUCTED FAGILITIES AUGUST 1998/117tance tothe propagating blast wave. Upon the failure of the fling, the blast fled the structural bays above and below Each floor slab. The filling pressures below the slab were freatr than the fling pressures above and caused an upward Toad on each sab “This net upward loading is shown in Fig. 15 as a spatially uniform pressure forthe purposes ofthis tepor. The slab i ‘modeled «simply supported element spanting from est ‘west between floor or roof beams (Biggs 1968) The length of this span between the supporting beams is 44m (16 f). ‘A rons section of «pica floor sla is shown in Fg. 16 [At midapan, the sla is 152 mm (6 in) deep and reinforced ‘with #4 bars at 229 mm (in) on center inthe east-west Sirecion and 57 mm (18 in) on cemer inthe north-south direction, Because of ther locaton near the bottom ofthe sab, these bars provide lite resistance to an upward loading, and the capacity is a uniform load of only 3 kPa (0.40 Ibs/in.’). ‘he woot slabs are typically 132 mm (6 in) thick as well bt are reinforced with fa bars at 406 mo (16 in) on center in the east-west direction and therefore provie resistance of 2 Pa (028 Ibsfin®) at midspan. Fig. 17 shows the loading onthe fifth-lor slab between Column Lines 20 and 22, which i considered here for ilu tration. The Toadings are asnumed to be the incident overpres- ~ | FIG. 15. Model ot Siab (Single Degree of Freedom) ry 5 AA FG. 16. Types ection of lor Sib at Cnterineof Span Be is Pras pao oe aes Tee (ilacons) FIG. 17. Blast Loading of Fifth-loor Siab between Column Lines 26 and 22 sures at the range of the midpoint of the structural bays above and below each slab. These loadings are further represented by twiangular pulses, as shown in Fig. 17. In this particular ‘ase, the load from below has a peak of 1 MPa (154 Ibsfin’), whereas the load above is only 8 kPa (87 lbsin#). In both cases, these loads act for relatively short durations, ‘The response of this slab to the loading is shown in Fig. 18. In this calculation, the static dead and actual live loadings are included as well as the blast loading from Fig. 17. Notice that the upward response of the slab has a long period, and the blast event represents an impulsive loading condition. In this ease, the maximum deflection is 236 mm (9.3 in.). This Aeflection exceeds the ultimate capacity of the floor slab and also represents a rotation of 5.3 degrees over the 4.9 m (16 f®) span. Under these conditions, the collapse of the slab from the direct blast loading is expected, 5 anette . awrite “Time (raeconee) FIG. 18. Blast Response of Fitth-Floor Siab between Column 2 ]) { ea FIG. 19. Schematic of Blast Response, North Elevation of Ning Story Portion of Murrah Building at Column Line G FIG, 20. Cross Section of Nine-Story Portion of Murrah Bulld- Ing Showing Blast Response of Slabe 118 /JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES / AUGUST 1998,A similar analysis was performed for the other floor and roof slabs in the building. The results are shown in Fig. 19. Tn particular, the floor slabs in the fifth floor and below be- tween Column Lines 18 and 24 were sufficiently loaded by the blast to fail as shown. However, the other slabs responded elastically to the differential blast loading and in some cases were not loaded above the static downward loads. Tig. 20 approximates the inward extent of this directly in- duced slab failure. The results show a penetration of failure to 12 m (40 fi) atthe second floor, which diminishes to zero at the sixth floor. CONCLUSIONS ‘The Murrah Building suffered important structural damage sa direct result of the bombing. This included the failure of three intermediate principal columns supporting a third-loor ‘wansfer girder on the north side. Some floor slabs in the prox- imity of the bomb were also directly demolished. As shown in the photographs, visible damage was in good agreement ‘with that predicted from calculations. ‘ACKNOWLEDGMENTS. The investigation was done by the American Soclety of Civil Engi neers under the sponsorship of the Federal Emergency Management ‘Agency (FEMA). Ciford Otve, FEMA Senior Engineer, wa project, officer and team leader. Ouitanding asstance was also provided by Stanley C. Woodkon, Research Structural Engineer, US. Army Compe of Enpigers (USACE); Richard D. Marshall, Resewch Seuctral Engineer, "National Institute of Standards and Technology (NIST), and David Kos tover Stturl Engineer, General Services Administration (GSA) APPENDIX. REFERENCES Biggs, 5. M, (1964), Inroduction 1 structural dymamles, McGraw Hil Ihe, New York, NY. Covi, W. G., Miakr. PF, Sozen, M. A. and Thoenton, CH, (1998), “The Oklaboma City bombing: Summary and recommendations for smolhazard mitigation." J. Pry. Conte Pac ASCE, 12), 100-112 ‘Coley, W. G., Sozen, M.A. Thornton, C. Hand Mahar PF (1996) “The Oklahoma ‘City bombing: improving building. performance through moki-haard mitigation.” FEMA Bulletin 277, Fedeel Eat gency Management Agency, Washington, D.C. “Elects of impact and explosion” (1946), Summary Tech. Rep. Dix 2. ‘Vol. 1, National Deenee Research Commie, Washiagton BC. “Fundamentals of protective design fr conventional weapons effects." (1986). Publ. No. TMS-835-1. Deparment of the Army. Washington, De. -MeVsj, M.K. (1988). “Spall damage of concrete rrvctores.” U.S. Army ‘Engr Wire Experiment Saton Teck Rep. SL-88-22. US, Army Ea sgineer Waterways Experiment Station, Vicksburg, Mis. “The Oklahoma City bombing: improving building performance trough ‘mult-haand mitigation" (1996). Fed. Emergency Mart. Agency Rep 277, Building’ Performance Assesment Team, Federal Emergency Management Ageney. Washington, DC ‘Sozen, M, A. Thornton, C. H., Corley, W. G., and Mlakar, P.F (1998), “The Otlahoma City bombing: Stcural detals and possible mech: anims forthe Murra Building.” J Perf. Const: Fact, ASCE, 120), 120-136. JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES / AUGUST 1998/19