Engineering Structures 30 (2008) 22532264 www.elsevier.

com/locate/engstruct

Dynamic analysis and seismic performance of reinforced concrete minarets

Halil Sezen a, , Ramazan Acar b , Adem Dogangun b , Ramazan Livaoglu c
a Civil & Environmental Engineering and Geodetic Science, The Ohio State University, 470 Hitchcock Hall, 2070 Neil Ave., Columbus,

OH 43210-1275, United States

b Department of Civil Engineering, Karadeniz Technical University, 61080, Trabzon, Turkey c Department of Civil Engineering, G m shane Engineering Faculty, 2900, G m shane, Turkey u u u u

Available online 20 February 2008

Abstract An unusually large number of minarets, which are slender tower structures, collapsed during the 1999 Kocaeli and Duzce, Turkey earthquakes with resulting damage to surrounding buildings and loss of life. The potential effects of the subsequently observed poor reinforcement detailing on the dynamic response is discussed. The probable cause of the extensive damage to reinforced concrete minarets is investigated by studying the observed failure modes and their seismic performance, and through the dynamic analysis of a representative minaret. The effects of spiral stairs, door openings, and balconies on the dynamic behavior are examined. The maximum dynamic internal force demands were compared with the calculated capacities. The locations of the maximum axial, shear, and exural demands predicted from the nite element analysis of the minaret model were consistent with the earthquake damage observed at those critical locations. c 2007 Elsevier Ltd. All rights reserved.
Keywords: Minaret; Reinforced concrete; Finite element analysis; Earthquake damage

1. Introduction A minaret is a slender tower built next to a mosque. While most historical minarets were constructed using reinforced or unreinforced stone or brick masonry, the majority of minarets recently constructed in Turkey are reinforced concrete (RC) structures. As shown in Fig. 1, a typical minaret structure comprises a base or boot on top of its foundation, a tapered transition segment, a circular body or shaft with one or more balconies, and a spire at the top. The base or boot is usually square or polygonal, and is sometimes called the pulpit by architects. The minaret can be free standing or the boot may be attached to the mosque structure. The minaret contains interior spiral stairs running all the way up to the highest balcony level which are not externally visible. Historically the balconies are built so that someone could climb up the stairs and call for prayer. With the advent of loudspeakers, these balconies are not needed; however, one or more balconies are built in each
Corresponding author. Tel.: +1 614 292 1338; fax: +1 614 292 3780.

E-mail addresses: sezen.l@osu.edu (H. Sezen), racar@ktu.edu.tr (R. Acar), adem@ktu.edu.tr (A. Dogangun), rliva@ktu.edu.tr (R. Livaoglu). 0141-0296/$ - see front matter c 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.engstruct.2007.11.005

minaret mainly for architectural reasons. Balconies create mass concentrations along the minarets height and affect its dynamic structural response. Currently, there are no structural code requirements or guidelines for the design of reinforced concrete minarets, or minarets in general, in Turkey. As a result, these slender structures have been built, for the most part, by experienced contractors and construction workers with no engineering knowledge. In most cases, each contractor constructs a typical minaret with the same structural and architectural features regardless of the local soil conditions or seismicity of the region. Turkey is located in one of the most seismically active regions of the world. Fifty-seven destructive earthquakes have struck Turkey in the twentieth century, resulting in the destruction of infrastructure and more than 90 000 deaths. During these earthquakes, many minarets were damaged or have collapsed. Sezen et al. [14] documents and discusses vulnerabilities and damages to 64 masonry and RC minarets after the 1999 Kocaeli (Mw 7.4) and Duzce (Mw 7.2) earthquakes. As a result of these two earthquakes, the collapse of 115 minarets in the city of Duzce alone was

2254

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

Fig. 2. Damage to the transition segment (photos by (a) Firat [8] and (b) Scawthorn [13]).

Fig. 1. Typical reinforced concrete minarets in Turkey.

reported [6]. Sezen et al. reports that approximately 70% of the RC and masonry minarets surveyed in Duzce sustained severe damage or collapsed. Even though the minarets are hardly ever occupied, they are located mostly in residential areas or shopping districts, and their collapse sometimes causes loss of life. It is extremely important to regulate the construction and design of these slender structures for safety reasons in anticipation of future earthquakes. This study attempts to identify the structural vulnerabilities of minarets based on their past seismic performance. In addition to widespread earthquake damage and collapses, some reported failures of minarets due to wind loading indicate that most of these tower structures are vulnerable to lateral loads. A large number of research studies investigating the seismic response of historical masonry minarets and towers are available [3,7,11,1517]. However, there are only a few studies investigating the lateral response of RC minarets [8,14]. Dogangun et al. [4,5] investigate the architectural and structural properties of these slender structures. The description of each minaret segment and the associated observed damage are presented below. 2. Observed damage and implications The type and distribution of damage in a structure varies greatly depending on many factors, including the detailing and properties of the structure and its components, soil properties, and the magnitude of the earthquake. The effect of local soil conditions on the seismic response of RC minarets is investigated by Acar et al. [1]. Observations from recent earthquakes suggest that the damage in the minarets is usually concentrated in a few specic locations. These observed local damage concentrations and vulnerabilities of minarets are presented here.

The relatively stiff boot or base of the minaret normally suffers no damage. The stiffness and strength of the minaret are reduced over the height of the tapered transition segment with a larger square or polygonal shape near its bottom and circular shape near the top. In a few cases, damage over the transition segment was observed. Fig. 2 shows two such cases where the concrete cracking or spalling was either spread over the segment or concentrated near the top just below the cylindrical body. Horizontal circumferential cracks and concrete spalling near the bottom of the minaret cylinder or body were the most common types of damage, leading to the collapse of RC minarets (Fig. 3). There are two main reasons for this type of failure. First, the cross section size becomes smaller, which results in reduced lateral and exural strength. Second, as shown in Fig. 3 in most cases at that location all longitudinal steel bars were lap spliced, creating a discontinuity. Prior to 1999, smooth reinforcing bars were commonly used in Turkey because they are less expensive, more readily available than ribbed bars, and easier to bend and cut on site compared with ribbed bars. Considering that the anchorage length required for the smooth longitudinal bars is signicantly larger than that of deformed bars, it is most likely that the lap spliced longitudinal bars failed before the full exural strength could be developed. However, many other minaret collapses, e.g., top two pictures in Fig. 3, suggest that failure may have occurred simply because of insufcient exural strength near the bottom of the cylinder. The minaret shown in Fig. 4 survived after the August 17, 1999 Kocaeli earthquake with some apparent distress causing light cracks and concrete spalling near the cylinder base. After nearly three months, during the November 12 Duzce event, the minaret collapsed at the section near the bottom of the cylinder where the smooth longitudinal reinforcing bars had been spliced. The lap splice length was approximately 800 mm (Fig. 4b). The ends of the longitudinal bars had 180 hooks (Fig. 4c). It appears that the combination of smooth bars with 180 end hooks, and the existence of short lap splices,

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

2255

Fig. 3. Minaret failures near the bottom of the cylinder bodies.

Fig. 4. Minaret in Kocyazi after the August 17 earthquake (minor cracks, [8]) and November 12 earthquake (collapse).

created a vulnerable region near the bottom of the cylinder. At that location, light or insignicant damage was observed after the rst earthquake, and collapse occurred after the second event. Anecdotal evidence and the picture of the survived minaret (Fig. 4a) indicate no signicant damage or permanent deformation after the Kocaeli event. This shows that the minaret probably stayed elastic during that earthquake. The typical failure mode (as in Fig. 3) after the second event suggests that the minaret was vulnerable near the bottom of its cylindrical body and had very little or no inelastic strength and deformation capacity to resist strong lateral forces during the latter event. Many similar post-earthquake reconnaissance observations provided evidence for the probable cause of failure, which is typically a result of sudden lateral and exural strength

reduction due to a combination of several factors, including the use of smooth rebar leading to weaker bond between concrete and steel, transverse hoops with hooks rather than continuous spiral reinforcement, short longitudinal lap splices, and the choice of lap splice location where the cross section is reduced to a circle with a smaller size. Furthermore, discontinued longitudinal rebar with 180 end hooks seemed to contribute to the sudden stiffness and strength change near the bottom of the cylinder. The collapsed minaret shown in Fig. 5 is a good example illustrating that the transverse reinforcement had 180 end hooks and all smooth longitudinal bars with 180 end hooks were cut at the same location where the failure occurred. At this cross section, due to longitudinal bar end hooks, the amount

2256

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

Fig. 5. Splicing of transverse reinforcement and longitudinal bars with 180 end hooks.

Fig. 6. Failure around mid-height of a minaret.

or effective area of concrete is reduced with both possible poor concrete connement and potential unnecessary steel congestion. The current Turkish Standards Code (TS 500, 2000) does not allow 180 end hooks at the end of longitudinal bars in RC components or structures. Similarly, the 180 hooks at the end of the transverse reinforcement or hoops (e.g., Figs. 4c or 5) may open up under cyclic loads, and do not conne concrete as effectively as spirals. In minarets which have a ring shaped cross section, the effective connement of concrete is a challenge. The use of 180 hooks at the end of both longitudinal and transverse steel exacerbates the problem near the bottom of the minaret cylinder where the longitudinal rebar is usually lap spliced. Structural failure or damage to the upper portion of the cylindrical body of the minarets was observed less frequently. If and when structural damage occurs, it is typically associated with some irregularities such as larger mass or stiffness concentrations around balconies. Longitudinal rebar often may be lap spliced and not anchored well at those locations. Fig. 6 shows one such case where lap spliced longitudinal rebar with 180 end hooks exist. No sign of distress or damage at other locations, including the bottom of the cylinder, suggests that longitudinal reinforcement discontinuity created by the lap splices may have been the primary cause of this specic failure. No damage was reported to the spires that are RC and monolithically connected to the minaret body. Metal sheet is commonly used for spires because of its lightweight and easy installation. There were few instances of metal spire failure over the virtually undamaged minaret body. If the metal spire

is anchored to the top of the minaret body properly, no damage should be expected. In almost all cases, as shown in Figs. 2 through 6, the stiffer minaret base or boot is not damaged. Also, the boot is usually attached to the mosque structure, making it relatively rigid. However, if the boot is not attached to the mosque or if nearby structures of part of the mosque structure hit the boot, the rigid body rotation of the boot and minaret failures may be observed as shown in Fig. 7. The above discussion of minaret failures focused on how and why minarets may have failed during the recent earthquakes in Turkey. In many cases minarets fell on top of the mosque creating a potential structural hazard and causing casualties (Fig. 3). It is also possible that the minaret may fall on nearby buildings. As shown in Fig. 8, during the 17 August earthquake one minaret fell on an otherwise virtually undamaged building causing damage in the upper stories. It is recommended that there should be a safe distance between the minaret and surrounding buildings. To the authors knowledge, no safe distance requirement exists in the current Turkish design and construction regulations. A practical safe distance from the minaret to the nearest structure could be the cylindrical body length between the top of the transition region and spire (Fig. 1). This assumes that the minaret is not going to fall on the mosque structure and the spire is made up of light metal sheet material. 3. Dynamic analysis of a representative minaret The architectural, geometrical, and material properties of minarets vary widely. For example, the height of a typical minaret can be between 10 m and 55 m. The minaret may

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

2257

Fig. 7. Rotation and failure of boots (photos by: (a) Anatolianquake [2], and (b) Locke [10]).

Fig. 8. Collapse of a minaret onto a nearby building [9].

have one or more balconies. The representative RC minaret investigated in this study is assumed to be 30 m high, including a 6 m boot or base, 2 m transition segment, 17 m cylindrical body, and 5 m spire. The assumed outer diameter and thickness of the cylinder are 1.76 m and 0.18 m, respectively. The plan view at the balcony level and elevation of the representative minaret are shown in Fig. 9. The assumed material properties are as follows: specied concrete compressive strength, f c = 16 MPa (design compressive strength is 11 MPa according to the Turkish codes), modulus of elasticity of concrete, E c = 27 000 MPa, concrete unit weight, = 25 kN/m3 , limiting concrete compressive strain cu = 0.003, smooth reinforcing bars with minimum specied yield strength, f y = 220 MPa, strain at hardening = 0.002, and fracture strain = 0.12. It is assumed that the minaret is located in a high seismic region with a soft soil site (Subsoil class Z4, and Zone 1 in Turkish Earthquake Code, [18]). According to the TEC, the structural behavior factor, R is 3, and importance factor, I is 1.2 for such structures. In this study, R = 3 is used in the response spectrum analysis as required by the TEC for the design of such structures. On the other hand, elastic material properties are used in all dynamic analyses presented below. This is mainly because the vast majority of RC minarets failed to develop plastic hinges during recent earthquakes. They either failed without any indication of ductile response (Fig. 3), or remained elastic with virtually no visible damage. The damage observed in Fig. 2b is an exception to more than 40 RC minarets surveyed by Sezen et al. [14]. It is usually burdensome to include all components in a structural model and consider their effect on the total behavior.

In this study, four nite element models representing the same minaret will be used to show how certain structural components affect the dynamic response, and thus to decide whether they should be included or can be ignored in a simplied model. The rst model (Model 1) includes all components of the minaret. The interior spiral stairs are ignored only in Model 2, the two balconies are ignored only in Model 3, balconies and door openings at the balcony levels are ignored in Model 4, and stairs, balconies, and door openings at the balcony levels are all ignored in Model 5 (Fig. 10). The computer program, SAP2000 [12] was used to analyze the models shown in Fig. 10. The response spectrum analysis of each minaret model is carried out using the design spectrum specied in the TEC [18] as shown in Fig. 11. Two ground motions recorded during the 12 November Duzce and 17 August 1999 Kocaeli earthquakes (Fig. 12) are also used in the dynamic modal time history analyses of the minaret models. The minarets are essentially slender towers with almost uniformly distributed mass, including additional concentrated masses if there are balconies. Thus, the dynamic behavior of these towers is strongly inuenced by the higher mode affects. All modes with total modal mass participation of 90% or more are included in the time history analysis. Periods, participation factors and directions for the rst seven modes estimated from the analyses are given in Table 1. As shown in the table, the dynamic response participation of the rst few modes of the minarets are relatively small compared with those of a typical frame structure. The total participation factor for the rst few modes is usually larger than 90% for typical frame structures, indicating that the effect of higher modes on the dynamic response of minarets is signicant. Also shown in the table are the periods for the rst eight and their percentage participations in the total response. First two modal shapes in the two directions, and the 7th mode shape for Model 1 are shown in Fig. 13. The maximum lateral displacement distribution over the height of the full minaret model (Model 1) is shown in Fig. 14. The maximum lateral displacements at the top of the transition segment or bottom of the cylindrical body are 0.007, 0.008, and 0.010 m, as calculated using the TEC design spectrum and Kocaeli and Duzce ground motions, respectively. The overall lateral displacement distribution shows that most of the exural deformations occur over the height of the relatively slender cylindrical minaret body.

2258

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

Fig. 9. Geometrical and cross-sectional properties of the representative minaret. Table 1 The rst few modes and their participation factors Modes Direction Model 1 Model 2 Model 3 Model 4 Model 5 Period (s) Participation (%) Period (s) Participation (%) Period (s) Participation (%) Period (s) Participation (%) Period (s) Participation (%) 1st y 0.643 46.3 0.627 48.5 0.608 42.8 0.594 42.1 0.578 43.6 2nd x 0.626 0.0 0.611 0.0 0.592 0.0 0.594 0.2 0.578 0.4 3rd y 0.136 22.7 0.127 21.4 0.128 24.0 0.122 10.2 0.113 21.6 4th x 0.131 0.0 0.122 0.0 0.123 0.0 0.122 13.1 0.113 0.2 5th y 0.056 9.0 0.052 8.9 0.052 10.9 0.051 5.6 0.047 8.4 6th x 0.055 0.0 0.051 0.0 0.052 0.0 0.051 5.6 0.047 3.0 7th Torsion 0.053 0.0 0.052 0.0 0.046 0.0 0.044 0.0 0.043 0.0

The lateral displacements at the top of the minaret, as calculated from the modal time history analysis of Model 1 using the 1999 Duzce and Kocaeli input motions, are shown in Fig. 15. The maximum calculated displacements were 0.14 m and 0.10 m for the Kocaeli and Duzce earthquakes, whereas the maximum displacement calculated using the code design spectrum was 0.08 m. Post-earthquake investigation of minaret failures revealed that the majority of minarets failed near the bottom of the cylindrical body immediately above the transition segment

(Fig. 3). The variations of axial and shear stresses at the intersection of cylinder and the transition segment are shown in Figs. 16 and 17. The maximum calculated compressive axial stress is 14 MPa, which is lower than the specied concrete compressive strength of 16 MPa. The maximum calculated tensile axial stress is about 16.5 MPa, indicating that the concrete has already cracked in tension. Considering that the specied tensile strength of steel rebar is 220 MPa, a tension failure should not occur as long as the bars are not cut or do not have short lap splices at that location. It is noted that the

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

2259

(a) Model 1: Includes all components.

(b) Model 2: Stairs are ignored.

(c) Model 3: Balconies are ignored.

(d) Model 4: Balconies and door openings are ignored.

(e) Model 5: Stairs, balconies and openings are ignored.

Fig. 10. Finite element models for the representative minaret with and without spiral stairs.

Fig. 11. Design spectrum recommended by TEC [18].

maximum displacements and maximum stresses occur at about the same time, at 4.1 s and 6.1 s for the Duzce and Kocaeli earthquakes, respectively. The axial compressive and tensile stress contours for the basic minaret model (Model 1) are shown in Fig. 18 when the maximum stresses and displacements were calculated for the two earthquake motions. Consistent with the observed performance, the stress contours show that both the largest compressive and tensile axial stresses are concentrated in the cylindrical body within a few meters above the transition segment. The stress distributions were very similar for both earthquakes with larger stresses measured for the Duzce earthquake. It should be noted that in the minaret model analyzed, the effect of longitudinal rebar lap splices or bar pullouts were not considered. The existence of short lap splices within the highly stressed minaret body probably intensied the problem and led to collapses during the earthquakes. Even though no RC minaret failures were observed by the authors in or around the balconies, the calculated stresses were high but still much smaller than those below the lower balcony.

2260

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

Table 2 The maximum internal forces at the top of the transition segment (height = 8 m), and at the bottom of the minaret or ground level Height Model 1 Model 2 Model 3 Model 4 Model 5 8m ground 8m ground 8m ground 8m ground 8m ground Shear forces (kN) DSa THA-Db 204 214 173 187 172 187 174 189 145 160 370 465 303 349 296 365 298 371 251 322 THA-Kc 272 304 249 285 241 275 246 281 231 281 Bending moments (kN m) DS THA-D 2490 3902 2210 3145 2151 3351 2179 3394 1901 2913 4695 7295 4056 6184 3657 5970 3746 6018 3724 5345 THA-K 3371 5414 3076 4956 2897 4742 2954 4834 2821 4628 Axial forces (kN) DS THA-D 592 960 541 860 507 873 514 888 458 785 593 966 541 860 508 878 515 893 458 785 THA-K 591 962 541 860 507 876 514 891 458 785

a Analysis using design spectrum. b Modal time history analysis using Duzce record. c Modal time history analysis using Kocaeli record.

Fig. 12. Ground motions recorded (a) at DZC station in Duzce during the 12 November Duzce, and (b) at YPT station in Yarimca during the 17 August Kocaeli earthquakes.

Fig. 13. First two modal shapes in the two directions, and the 7th mode shapes for Model 1.

Fig. 19 shows the distribution of maximum shear stresses along the minarets height. The existence and location of spiral stairs noticeably inuence the magnitude and distribution of shear stresses. Fig. 20 shows the maximum shear stress contours calculated using the code design response spectrum. In the same gure, the locations of spiral stairs inside the minaret are also shown. Evidently the stairs increase the shear strength and reduce the stress signicantly.

Fig. 14. Lateral displacement distribution over the height of minaret calculated using (a) TEC design response spectrum, and (b) modal analysis using Kocaeli and Duzce motions.

The maximum internal forces calculated from the dynamic analysis of the representative minaret models using the Duzce and Kocaeli earthquake records and code design spectrum are summarized in Table 2. The internal forces calculated from the

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

2261

Fig. 15. Lateral displacement time histories at the top of the minaret subjected to (a) Duzce, and (b) Kocaeli earthquakes.

Fig. 16. Time history of normal stresses at the top of the transition segment for the (a) Duzce, and (b) Kocaeli earthquakes.

Fig. 17. Time history of shear stresses at the top of the transition segment for the (a) Duzce, and (b) Kocaeli earthquakes.

code design spectrum seem to be relatively small compared to the forces from modal time history analyses. Although the two ground motions result in fairly close shear and axial forces, the moments, and hence the axial bending stresses, appear to increase signicantly at the bottom of the minaret compared to the top of the transition segment located at 8 m above ground. 4. Demand versus capacity Both the dynamic analysis results and observed minaret failures indicate that the bottom of the cylindrical body or the top of the transition segment is the most vulnerable section in RC minarets. As shown in Table 2, the maximum internal forces at that level were calculated for Model 1, which includes all structural components of the minaret. These maximum elastic moment and shear force demands will be compared with the predicted moment and shear strengths here. During the post earthquake reconnaissance visits, the authors observed that the most common steel rebar used in the earthquake affected region was S220 with a yield strength

of 220 MPa. For easy workmanship, small size bars with diameters of 12 or 14 mm were commonly used. In many cases, it was found that the longitudinal reinforcement ratio was close to or smaller than the minimum code specied ratio of 0.01. Many studies carried out after the 1999 earthquakes found that the concrete strength could be as low as 10 MPa. In this study, a concrete design strength of 11 MPa is used. It should be noted that according to the current Turkish building code [19], the minimum concrete strength is 20 MPa. Although the critical section includes two layers of longitudinal rebar (Fig. 8), a single layer of steel is assumed at the centerline of the section (Fig. 21). The calculated axial loadmoment interaction diagram is plotted in Fig. 21 for a typical reinforced concrete ring section with a 0.18 m thickness and 1.76 m outer diameter (do ) and 1.40 m inner diameter with a corresponding cross sectional area, Ac of 0.9 m2 . The diagram also shows the elastic axial loadmoment demands calculated from the dynamic time history analysis and code design spectrum. The axial load, N is assumed to be 592 kN (N /Ac f c = 0.06).

2262

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

Fig. 18. Maximum tensile stress contours, Smax for (a) Duzce, and (b) Kocaeli records, and compressive stress contours, Smin for (c) Duzce, and (d) Kocaeli earthquakes.

Fig. 20. Effects of spiral stairs on shear stress distribution in Model 1.

Fig. 19. Maximum shear stress contours for (a) Duzce, and (b) Kocaeli records.

The normalized moment demands (M/Ac do f c ) are 0.14, 0.27, and 0.19 for the code design spectrum and dynamic response with Duzce and Kocaeli motions, respectively. The comparison of axial loadmoment demands and the capacity curve suggests that the exural capacity of the minaret model considered

here is not sufcient, and smaller than the inelastic demands calculated using the code design spectrum and elastic demands from the dynamic time history analysis. It should be noted that the exural demand calculated from the code design spectra is signicantly lower than the demand imposed on the minaret during the two earthquakes. This, and the displacement response plotted in Fig. 14, imply that the strength and displacement capacity required by the code design spectra may be much lower than the expected demand during large events such as Kocaeli and Duzce earthquakes.

H. Sezen et al. / Engineering Structures 30 (2008) 22532264

2263

Fig. 21. Axial loadmoment interaction diagram and predicted demands.

The shear strength of the cylindrical minaret body is predicted from the column shear strength equation provided in the Turkish Building Code, TS500 [19]. Vcap = 0.8 0.65 f ctd Ac 1 + 0.007 N Ac + Asw f ywd d s (1)

where Ac is the gross cross-sectional area, f ctd is design tensile strength of concrete, N is the factored axial force calculated under simultaneous action of seismic lateral and axial loads, Asw is the transverse steel area, f ywd is the design yield strength of transverse reinforcement, d is the effective depth of the section, and s is the transverse reinforcement spacing. Ignoring the contribution of shear reinforcement, the shear capacity of a typical minaret cross section is estimated as Vcap = 423 kN with Ac = 900 000 mm2 , fctd = 0.9 N/mm2 , N = 592 kN. The predicted shear strength, even without the strength contribution from transverse steel, appears to be much larger than the calculated shear force demands listed in Table 2. 5. Conclusions Observations from the minarets that collapsed during recent earthquakes and the analyses of a representative minaret showed that the bottom of the cylindrical minaret body immediately above the transition segment is the most vulnerable section under seismic loading. The poor design practices include the use of: (1) smooth steel rebar, (2) 180 end hooks at the ends of both the transverse and longitudinal reinforcements, (3) unstaggered and short longitudinal lap splices, (4) inadequate transverse hoops instead of a spiral reinforcement, and (5) a short transition length between the square boot and cylindrical body. These practices exacerbated the problem of insufcient bending strength and deformation

capacity near the bottom of cylindrical part and increased the susceptibility of this section to failure. Practicing engineers and contractors can improve the design of minarets by providing a more gradual change from a square or polygonal section to a smaller circular section using a longer transition segment, and by eliminating lap splices near the critical section and using instead staggered lap splices over the height of the minaret body. When either stairs or balconies are ignored in the analysis, the maximum shear and exural demands were underestimated by approximately 20%. It was found that the shear strength of the minaret was larger than the maximum shear demands calculated from the dynamic analysis, indicating that shear was not the likely cause of failure. In addition, the shear stress demands were reduced at locations where spiral stairs existed. The results of the elastic time history analyses have shown that the exural capacity at the critical section would be exceeded when the representative minaret is subjected to ground motions recorded during recent earthquakes. This is partially because the exural strength is smaller under relatively small axial loads. The strength and displacement capacities calculated using the inelastic code design spectra may be much lower than the elastic demands imposed during large seismic events such as the Kocaeli and Duzce earthquakes. Almost all minarets surveyed after the recent earthquakes either behaved elastically or collapsed (Fig. 3). The existing design and construction practices should be improved to provide sufcient ductility. Otherwise, it will be misleading to use the inelastic response spectrum analysis prescribed by the current code. References
[1] Acar R, Livaoglu R, Dogangun A, Sezen H. The effects of subsoils on the seismic response of reinforced concrete cylindrical minarets. In:

2264

H. Sezen et al. / Engineering Structures 30 (2008) 22532264 Fifth international conference on seismology and earthquake engineering. 2007. SE-24. Anatolianquake. A curated repository of data on the Kocaeli-G lc k and o u Bolu-D zce earthquakes of 1999. http://anatolianquake.org/; 1999. u Dogangun A, Sezen H, Tuluk OI, Livaoglu R, Acar R. Traditional Turkish monumental structures and their earthquake response. International Journal of Architectural Heritage 2007;3(1):25171. Dogangun A, Tuluk OI, Livaoglu R, Acar R. Traditional Turkish minarets on the basis of architectural and engineering concepts. In: Proceedings of the rst international conference on restoration of heritage masonry structures. 2006. P34, p. 110. Dogangun A, Acar R, Livaoglu R, Tuluk OI. Performance of masonry minarets against earthquakes and winds in Turkey. In: Proceedings of the rst international conference on restoration of heritage masonry structures. 2006. P32, p. 110. Duzcedamla. Mosques expose the effects of earthquake. http://www. duzcedamla.com/text/index.dwx?TextID=2811; 2006 [in Turkish]. El-Attar AG, Saleh AM, Osman A. Seismic response of a historical Mamluk style minaret. In: Earthquake resistant engineering structures III. WIT; 2001. p. 74554. Firat YG. A study of the structural response of minarets in the 1999 Anatolian earthquakes. M.S. thesis. West Lafayette (IN): Purdue University; 2001. Also, http://cobweb.ecn.purdue.edu/anatolia/FIRAT/ 5adapazari/adapazari2.ppt#257,2,Slide2. H rriyet. Newspaper earthquake photo archive. http://arsiv.hurriyetim. u com.tr/x98/deprem/depremgaleri.htm; 2006. Locke J. Damage photos for 1999 Earthquakes. http://www.marin.cc.ca. us/jim/photos/turkey/duz6.jpg; 2006. Menon A, Lai CG, Macchi G. Seismic hazard assessment of the historical site of Jam in Afghanistan and stability analysis of the minaret. Journal of Earthquake Engineering 2004;8(1):25194. SAP2000. Integrated software for structural analysis & design. California: Computers and Structures Inc.; 2006. Scawthorn C. Preliminary Report Kocaeli (Izmit) Earthquake of 17 August 1999. http://mceer.buffalo.edu/research/Reconnaissance/ turkey8-79/ScawPrelim.pdf; 1999. Sezen H, Frat GY, Sozen MA. Investigation of the performance of monumental structures during the 1999 Kocaeli and Duzce earthquakes. Paper no: AE-020. In: Fifth national conference on earthquake engineering. 2003. Sykora D, Look D, Groci G, Karaesmen E, Karaesmen E. Reconnaissance report of damage to historic monuments in Cairo. Egypt following the October 12, 1992 Dahshur earthquake. Technical report NCEER-93-0016. National Center for Earthquake Engineering Research, State University of New York at Buffalo; 1993. Syrmakezis CA, Asteris PG, Sophocleous AA. Earthquake resistant design of masonry tower structures. In: Structural studies, repairs and maintenance of historical buildings. Advances in architectural series, vol. 3. Southampton (UK): Computaional Mechanics Publications; 1997. p. 37786. Syrmakezis CA. Seismic protection of historical structures and monuments. Structural Control and Health Monitoring 2006;13:95879. TEC. Turkish code. Ministry of Public Works and Settlement. 1998. Specication for structures to be built in disaster areas; Part III-earthquake disaster prevention, Government of Republic of Turkey; 1998. TS500. Requirements for design and construction of reinforced concrete structures. Ankara (Turkey): Turkish Standards Institute; 2000 [in Turkish].

[2] [3]

[12] [13]

[4]

[14]

[5]

[15]

[6] [7]

[16]

[8]

[17] [18]

[9] [10] [11]

[19]