Read Microsoft Word - 29 - Hitaka.doc text version



Toko HITAKA1, Keiichiro SUITA2, and Mikiko. KATO3 SUMMARY Currently, the CFT column base is designed in a similar manner as the steel column base in Japan. There are three column base designs used in Japan. Of the three, the base plate connection has been used for the longest time. The Japanese steel column base design was changed twice in its past after 1978 and 1995 earthquakes. After the 1995 earthquake, the moment resisted by the base plate is estimated considering its rotational stiffness. Ductile design is possible if ductile anchor bolts are used. If conventional non-ductile anchor bolts are used, the design is similar to the AISC design. The embedded column connection was introduced after 1978 earthquakes to increase the rigidity of the column base. The behavior of the embedded column base is reliable, but the base plate connection is preferred if the construction period must be short. Behavior of the CFT column base is similar to that of the steel tubular column base. The purpose of the most researches on the CFT column base is to enhance workability. The semi-embedded column base is one such example. This is a combination of the base plate and the embedded column connection. While the construction process is similar to the base plate connection, the performance is the same as the embedded column connection. Keywords: concrete filled steel tube, column base, connection, design method, ductility. INTRODUCTION In Japan, behavior of the CFT columns and connections were actively studied during the past decades, but little attention was paid to the CFT column bases. This is due to the perception, common among the researchers and engineers, that the performance of the CFT column base will be equal or better than the steel tubular column base, because the inner concrete may prevent local buckling of the steel. The steel tubular column on the other hand has been widely employed for building frames since long time ago in Japan, where the space frame system is the norm in the practice. Behavior of the steel tubular columns, including its column bases, was thoroughly studied and the tubular column bases are designed using an established design procedure. CFT column bases are currently designed using similar procedure. The objective of this paper is to introduce the design procedure for the steel column base most frequently used by the Japanese engineers. Three types of column base design exist in Japan; base plate connection, encased base plate connection, and embedded column connection, as shown in Figure 1. Among the three, the base plate connection has the longest history in Japan. Encased base plate connection is becoming unpopular these days. The embedded column connection is used mostly in the large-scale buildings. Most CFT frames used to be employed only for such buildings, where this connection is used. Recently, the increasing demand for shorter construction period is pushing the employment of the base plate connection for CFT columns. When the building has basements, the section of the column is switched from tube to a cross-H section at the ground level, which is embedded in the concrete wall. So far, the CFT buildings, and therefore their column bases, have not been challenged by large earthquakes. Many steel base plates, however, were damaged due to large earthquakes. The design methods were revised after some of these events. Benchmarking earthquakes are Izu-Ohshima inshore and Miyagi-ken offshore earthquakes in 1978 and Kobe earthquake in 1995. Many base plate connections were found severely damaged in both cases.

1 2

Reserch Associate, Faculty of Human-Environment Studies, Kyushu University, Japan, e-mail: [email protected] Associate Professor, Div. of Earthquake Disaster Prevention, Disaster Prevention Researvh Institute, KyotoUniversity, Japan, e-mail: [email protected] 3 Engineer, Nikken Sekkei Co., Japan, e-mail: [email protected]


Behavior of the steel column bases was actively studied after the 1978 earthquakes. The findings still are the basis of the current design methods to this day. In response to Kobe earthquake, design method for the base plate connection was reviewed. The results were presented in the publication from Building Center of Japan in 2001. Recent studies on CFT column bases are to develop better column base designs, rather than to grasp the behavior of traditional column bases. Ideas of several such studies are, for example, to develop column bases that are physically very close to a pin, damage-control type, or Semi-Embedded type. In the following, the second chapter introduces a brief history of steel base plate connection design. Damaged steel column bases found after Kobe earthquake are also shown. The structural design is reviewed by building officers of the city/prefecture government in Japan. Engineers commonly design buildings using procedures recommended by Building Center of Japan (this procedure is called BCJ design procedure hereafter). The third chapter describes the current BCJ design procedure of the base plate and the embedded column connection, and compare the two connections from construction and other aspects. The last chapter introduces two recent experimental researches on CFT column bases in Japan.

Non-Shrinkage Mortar Anchor Bolt

Reinforcing Bar Base Plate Hoop

Base Plate Base Plate Base Beam Anchor Bolt Base Beam

Base Beam Base Beam

Anchor Bolt

a) Base Plate Connection

b) Encased Base Plate Connection Figure 1 Column Bases in Japan

c) Embedded Column Connection

BRIEF HISTORY OF STEEL COLUMN BASE DESIGN IN JAPAN Two major events changed the Japanese steel column base design since the steel members started being used for buildings. They are the establishment of the new seismic design standards in 1981 and the revision of the Japanese building code in 2001. Before 1981 The structural steel, in its earliest days, was mainly used for the factories in Japan, which at that time was enjoying an economic boom. In those days, built-up members consisting of angles and rivets were used for building frames. The column base was base plate connections, reinforced by ribs and shear panels placed between the built-up components. Tension and shear were resisted by anchor bolts. Shear lags were used when large shear was expected. In the structural design, the connection was modeled as a pin or a rigid end depending on the designer's strategy. At this time, it did not make much difference whether the connection was regarded as a pin or a rigid connection, because the story drift was not regulated by the code. In 1966, Japanese Industrial Standard was revised, which included the wide flange shapes. This caused great reap for these steel members to be used for buildings. The design approach for column bases remained the same as for the built-up members. Normally, for the connections modeled as pins, the anchor bolts were located between the flanges. The earthquakes around Miyagi prefecture in 1978 caused severe damages to the base plate connections. The 282

column base design became an issue for the first time. Many researchers were motivated to study behavior of the steel column bases, and its effect on the overall building behavior. The damages proved that the rigidity of the actual base plate connections is not to be relied on. The smaller rigidity of the column base meant larger strength demand for the second floor beams. After 1981, Pre-Kobe Earthquake Reflecting the observations of the damages, the new seismic design standard was endorsed in 1981. However, the column base design practice itself remained the same as before. Rather, it was the story drift limit endorsed in the new standard that drew designers' attention to the column base design. In the new standard, the story drift angle of a building under design seismic force should not be more than 1/200 rad. Whether the column base is modeled as a pin or rigid end now mattered. The reports of the damaged base plate connections generated the view that the performance of such connections is no better than pins. The legislation announced that only the column bases with embedded length of more than 30 cm were to be regarded as rigidly connected to the ground. The base plate connections were thus modeled as pins after the incident. The encased base plate connections and the embedded column connections were invented because of this change. The embedded column connections were mostly used for large-scale buildings, and base plate connections, for mid-to-small buildings. In the meantime, some companies developed base plate connection packages for wide flange shape and tubular members used for building frames. Combinations of cast-steel base plate, anchor bolts and washers were sold with a design manual. The performance (i.e. rotational stiffness, strength and ductility) of such connections is pre-qualified by the authority. Non-commercial base plate connections, thus regarded as pin connection, were designed to resist only to shear and tension. Steel Column Bases Damaged during Kobe Earthquake Some details of column base damages found in 426 steel buildings are reported in Reconnaissance (AIJ, 1995) after Kobe earthquake. The damaged buildings were mostly low-rise. Eighty-two percent of these buildings had less than 6 stories (CFT was still quite new structural member at this time. Hence no CFT buildings damaged in Kobe). Base plate connection was found in 63% of the 426 buildings. Sixty-seven percent of the unbraced frames with base plate connections (112 buildings) collapsed or were severely damaged. In almost all the damaged base plate connections, anchor bolts had fractured or elongated severely. Other forms of damage were also found in a few cases (e.g. large plastification of base plate in 6 buildings, fracture at welds between the column and the base plate in Figure 2 Damaged Base Plate 4 buildings). Twenty-six percent of unbraced frames and 29% of braced frames were damaged in the column base. No damages in the embedded column connections were reported. Figure 2 shows an example of a damaged base plate connection. The column of this four-story building is a 200x200 wide flange shape member. The damaged connections were mostly such small base plate connections that had been modeled as pins. Post-Kobe Steel Column Base Design Reflecting the damages observed after the event, efforts were made to investigate proper seismic design force into the base plate connection. The results were reflected in BCJ design procedure (explained in the next chapter). The procedure for the embedded column design remained unchanged. In the new procedure, strength demand for the base plate connection is estimated considering the rotational stiffness of the connection. Another major change is the introduction of a design method relying on ductility of the anchor bolts. Engineers, however, are still careful in using this method. CURRENT STEEL COLUMN BASE DESIGN IN JAPAN Background Researches Most of the findings about steel column bases, i.e. the rotational stiffness, yield and ultimate strength, types of damages, histeresis, were made in late 1970's and 1980's. The rotational stiffness of the connection varies 283

depending on the construction detail or quality. The currently used equation for stiffness (the equation (1) in the next section) tends to underestimate the stiffness, especially in the case where specimens are not loaded axially. The ultimate strength of the steel column base connections is well predicted using the current design formula (the equation (2) in the next section). Figure 3 compares the calculated strength and test results of the connections (AIJ, 2001). The horizontal axis is the axial load ratio for the base plate connection (Left), and the embedded column length for the embedded column connection (Right). In Fig. 3b, it is observed safe to expect hinging of the column at the base if more than twice the column depth of the lower end of the column is embedded.

Embedded Length / Column Length

Figure 3 Comparison of Measured Strength and Calculated Strength (Left: Base Plate Connection, Right: Embedded Column Connection) The slip model is appropriate as the hysteresis model for base plate connections. The histeresis curve slips at a moment estimated as the moment resisted by concrete bearing as observed in Figure 4 (Akiyama, 1985). The shape of the hysteresis loops varies according to axial load, but the building's overall behavior is not much affected (Kawano, 1998). The engineers have so far avoided relying on the ductility of base plate connections, because of its thinner hysteresis loops. However, analytical studies have shown that the story drift and hinging of frame members may be distributed along the height more equally if the column bases are semi-rigid rather than perfectly-rigid (Yamada, 1997 and Kawano, 1998). In another analytical study, ductility demand for the base plate connections was investigated. The results showed that the plastic rotational capacity of 0.03 rad. of the column base is sufficient for buildings of varied base shear strength ( 0.25), beam-column strength ratio ( 1.2), and column base ­ column strength ratio (more than 1/3) (Hasegawa, 2000). Base Plate Connection A typical configuration of the base plate connection is shown in Figure 5. Reinforcement in the base concrete is designed so that the force transferred from the steel column will be resisted by the RC column under the base plate, and then the base concrete beams. The anchor bolt embedment length must be larger than 20 times the diameter of the anchor bolts (same as for the reinforcing bars in the AIJ RC design standard). The anchor bolts' diameter is typically less than 50mm. Anchor bolts are usually tied to each other by so-called anchor beams shown in Fig. 5. The upper limit of the marginal clearance for bolt holes is around 5mm. Nowadays, using shear lags is rare in Japan. The anchor bolt's shear strength is used to resist shear force. Leveling nuts are not used. The major changes in the BCP design procedure after Kobe 284

Column Base plate Mortar Anchor bolt Base Concrete

Figure 4 Hysteresis of Base Plate Connection


Anchor Beam

Figure 5 Typical Base Plate Connection

earthquake are 1) estimation of design moments for the frames and the column bases considering the rotational stiffness of the base plate connection, and 2) different design approaches depending on ductility of the anchor bolts. If "ductile anchor bolts" are to be used, yielding is allowed for the column base (anchor bolts) under large earthquakes. Yielding of base plates is not favored. Rotational Stiffness and Design Moment of the Base Plate Connection The rotational stiffness of a base plate connection, K BS , is calculated by the following equation (1).

K BS =

E nt Ab (d t + d c ) 2lb



E : Young's modulus, nt : number of bolts on the tension side, Ab : anchor bolt's section area, lb : embedded

length of the anchor bolt. Other geometric notations are defined in Fig. 7. This equation gives a conservative estimation of the stiffness. A linear analysis considering this rotational stiffness gives larger moments for the first story columns and the second story floor beams. In the case of a base plate connection with configuration as shown in Figure 8, the distance between the flexural point of the first story column and the ground level generally is 0.4 to 0.45 times story height. This value is not much affected by the detail of the connection (e.g. anchor bolt location, stiffeners around the column end). In most cases, it is only for the second floor beam sections that larger section is required if the rigidity of the column base is reduced (Hasegawa, 2000). Strength of Column Base and Anchor Bolt's Ductility Figure 6 schematically shows the column base design procedure for regular buildings of more than certain volume. In Japan, designing of such a building is a two-step procedure. First, the design moment, axial and shear forces for each member are obtained by linear static analysis considering the rotational stiffness of the column base, K BS . The base plate connection design is determined at the second step. Depending on the ductility of the used anchor bolts, the ultimate or yield strength of the connection is compared with either (=1.3) times the plastic strength of the column (Mpc) or the force that exist in the column base when the building is resisting with its full capacity (Mla). Mla is obtained by multiplying the forces due to the earthquake load in the connection by (=2). The value for , 2, was determined because the ultimate base shear of steel frames is usually about twice the design seismic load. A "ductile" anchor bolt is defined by the yield ratio of the steel. The yield ratio of a ductile anchor bolt is no more than 0.75 (0.65 for some special anchor bolts). These values are determined so that the yield strength of the anchor bolt's shank exceeds the ultimate strength of its thread section. The base plate connections with such anchor bolts possess the plastic rotational capacity of more than 0.03 rad.

1st Step: ASD (Rotational Stiffness Considered)

2nd Step If Ductile Anchor Bolts Used Yes Yes If M M u pc No No

M u M la


M y M la


G Figure 6




BCJ Base Plate Connection Design Method

In the case of using "ductile" anchor bolts, the ultimate strength of the base plate connection, M u may be used.

M u is estimated assuming stress distribution as shown in Figure 7. The compression strength of the concrete is 0.85 times its mix strength. M u is calculated by the equation (2),


( N u - N ) d t (N + Tu ) D 1 - N + Tu M u = Tu d t + 2 Nu ( N + 2T ) d u t

(N u N > N u - Tu ) (N u - Tu N > -Tu ) (- Tu N -2Tu )


where N : axial force in the connection, N u : maximum compression strength of base concrete under the base plate ( = 0.85 BDFc ), Tu : maximum tensile strength of anchor bolts on the tension side, Fc : mix strength of concrete. Other geometric notations are defined in Fig. 7. If non-"ductile" anchor bolts are used, the connection must be designed such that neither the concrete nor the anchor bolts will be damaged by the force, Mla. This is similar to comparing Mla with the yield strength calculated using the ASD procedure shown in AISC Design Guide (1990). In AISC Design Guide (AISC, 1990), a base plate connection under axial force with large eccentricity as shown in Figure 8 is designed as an example using ASD and LRFD methods. The same connection is designed here using Japanese procedure for a comparison. For the Japanese case, the anchor bolts are assumed non-ductile. Assuming that the moment is generated by the seismic load and the axial load is caused by the gravity, the design moment for the Japanese designing was multiplied by 4/3 (=2/3x2), reflecting the first step of BCJ design procedure. Considering the column base rotational stiffness in the Japanese case, the design moment for the column base is reduced to about 2/3 of the American case, where the column base is assumed perfectly rigid. The design moment caused by the seismic force is then multiplied by (=2) at the second step. Minimum required anchor bolt diameter and base plate thickness are calculated for the connection under axial load, P


In Concrete

In Anchor Bolts (two cases)

Figure 7

Stress Distribution in Base Plate Connection





Anchor Bolt Diameter Base Plate Thickness (mm)


tb la


JP ASD LRFD B.P. Thickness

D D=B=14in.(356mm) de=1.5in.(35mm) tbB.P.thickness laEmbedded length dbDiameter of A. Bolt


ntBolts on the tension side =2

P: Axial force 120kip (267kN)

Connection is designed for varying design moment, M.


A. Bolt Diameter JP ASD LRFD






Figure 8

Design Comparison (US-Japan) 286

(=267kN) and varying moment, M. The design results for the required base plate thickness and anchor bolt diameter are compared in Figure 8. The horizontal axis of the graph is the moment divided by P and half the depth of the column section. Despite the difference of material strength used in the calculation in the US and Japan as shown in Table 1, the results are similar. Ductility of anchor bolts does not affect the result unless the design sequence A in Figure 6 is taken. Table 1 Material Strength Used for Design

Elements Anchor Bolt Base Plate Concrete


J apan 1st Step 2nd Step Fy 2Fc/3 Fc

US ASD LRFD 0.33Fu 0.56Fu 0.75Fy 0.9Fy 0.350.7Fc 0.511.7Fc

Fy and Fu: nominal yield and tensile strength of steel, respectively, Fc: mix strength of concrete Embedded Column Connection Figure 1c shows a typical configuration of the embedded column connection. The embedded column length is determined such that the moment and shear will be resisted by the bearing of concrete on the embedded column. In the case of steel tubular column, about twice the column depth is generally sufficient for the embedded column length. Filling concrete in the steel tube embedded in the concrete has been one of the solutions to avoid kink near the concrete surface since early days. The problem in the embedded column design arises if the column is located at the fringe of the building and the base concrete cannot be extended outward. Larger rotational stiffness is expected for the embedded column base than the base plate connection. In the linear analysis, the rigid point of the column is assumed to be 1.5 times column depth below the concrete. Other Issues In using the base plate connections, engineers have two options, i.e. to use commercial connections or to design and fabricate the connections using ordinary material. Though expensive, the commercial connections possess reliable rigidity and large rotational ductility. If the base plate connections are to be fabricated, the engineers estimate the stiffness using the equation (1). Although not readily available yet, the "ductile" anchor bolts were specified in the Japanese Industrial Standard in 2002. The circulation of such anchor bolts will increase, and so will the fabricated "ductile" base plate connections in future. In terms of construction, the embedded column connection is not too desirable. In the case of employing this connection, steel members must be erected before the base concrete casting. This requires that preparation and fabrication of steel members and selection of the construction workers should all be finished before casting the base concrete. One major attraction of using steel system is that its construction period is shorter. Replacing the embedded column connection with the commercial base plate connections, despite the expense, is sometimes a reasonable solution. It is a difficult task to place anchor bolts correctly to match the column location. The poured concrete some times washes the anchor bolts away from its original position. If large stress is expected in the connection (e.g. braced columns) complete joint penetration groove weld is used between the steel column and the base plate. Fillet welding is used if the strength is sufficient to resist design force. As to possible problems in using CFT instead of steel columns, more reinforcement is needed in the base concrete beam because the column has larger strength than steel columns. Likewise, the required embedded column length is larger for the CFT columns. SEMI-EMBEDDED COLUMN CONNECTION TEST (This research is a work of Morino, et. al. (2003)) Objective: Studying the behavior of Semi-embedded CFT Column Base. According to studies on the embedded column bases, the embedded column length is determined by the moment generated by the bearing of the 287

concrete, which is roughly two times the depth of the column section (2D). Semi-embedded CFT Column Base consists of a CFT column embedded over the length shorter than 2D in the base concrete, a base plate and anchor bolt. A concrete pocket similar to those required for shear lags is required. But it is not necessary for the column to be placed before the base concrete casting, thus the construction is similar to that for the base plate connection. The construction procedure of the semi-embedded column base is schematically illustrated in Figure 9. Firstly, concrete is cast for the foundation beam with anchor bolts arranged, leaving a void area for the steel erection. Then, the steel is erected, and finally the concrete is post-cast into the pocket beneath the column base. In this manner, a mixed construction of steel and concrete is avoided. Specimens and other details: a total of eight specimens were tested under horizontal load at the column top. Table 2 summarizes the test specimens. The first alphabet indicates monotonic or recursive loading. The numerals e.g. 22 indicate width-thickness ratio of the column tube. The last alphabet in the designation of the specimen signifies F: full, H: half, Q: quarter, Z: zero length of the required embedded length for the embedded column connection by AIJ (2001). No axial loading. The base plate is thick enough not to yield due to tension in the anchor bolts. The base plate is fillet welded around the steel tube. Anchor bolts are designed for two specimens (Z and Q) such that the ultimate moment resisted by concrete baring and the anchor bolts will be equal to the yield strength of the CFT column. Anchor bolts are tied together by so-called "anchor beams" near the bottom. Table 2 Dimensions Related to Test Parameter


Figure 9 Construction of Semi-Embedded Column Connection

Figure 10 Specimen

Figure 11

Test Setup


Figure 12 Horizontal Force ­ Rotation Relation (: horizontal displacement of loading point, Hcu: plastic strength of CFT column)

Test Parameter: Embedded column length (le in Table 2). Results: The behavior of specimens is compared in Figure 12. Performance of the column embedded over 1D (R22H) is almost equivalent to those of the embedded column connection (R22F). The specimen with less embedded length (R22Q) behaves fairly similarly to the column with a base plate connection (R22Z). It was observed that the reinforcing bars also carried substantial amount of moment in R22H. The strength of the specimens R22Z, R22Q and R22F are estimated considering the moment resisted by anchor bolts and concrete bearing (Hbu1 in Fig. 12). The ultimate strength of R22F slightly exceeds Hbu1, which is substantially larger than the column plastic strength. Hbu2 in the figure for R22H is estimated considering the moment resisted by the reinforcing bars. A stiffness estimation method was proposed. The calculated stiffness showed good agreement for the specimens R22H and R22F. CONCLUDING REMARK CFT column base designs are similar to the steel column. For the steel buildings, the base plate connections are more frequently used among the three column bases used in Japan, because the required construction period for this connection is shorter. The situation is the same for CFT column bases. The design procedure for the base plate connection for steel buildings was revised after the discovery of many damaged base plates caused by Kobe earthquake. In the latest procedure, the strength demand for the frame is estimated considering the reduced rigidity at the column base. The base plate connection is designed in different manners depending on the ductility of the anchor bolts. Yielding of the anchor bolt is allowable if its yield ratio is no less than 0.75. The base plate connections with such anchor bolts possess plastic rotational capacity of more than 0.03 rad. The embedded column connection is mostly used for large-scale buildings. So far, no damages have been reported for this type of column base. The embedded column length of more than twice the depth of steel column 289

is required for the hinging in the column lower end. A new CFT column base, the semi-embedded column connection was proposed and tested by Morino, Combined with anchor bolts and a base plate, the semi-embedded column connection possesses equivalent structural performance as the embedded column connection. The construction process is similar to that for base plate connections, hence more economical than the embedded column connection. REFERENCES Akiyama, H. (1985), Seismic Design of Steel Column Base, ISBN4-8189-0531-3 C3052. (in Japanese)

AISC, Steel Design Guide Series Column Base Plates, Chicago, Illinois, 1990

Architectural Institute of Japan (1995), Reconnaissance report on damage to steel building structures. Architectural Institute of Japan (2001), Recommendations for Design of Connections in Steel Structures. (in Japanese) Building Center of Japan (2001), Commentary on structural design standard and techniques. (in Japanese) Hasegawa, T. (2000), "Seismic response behavior of steel rigid frames having exposed-type column base," Journal of Constructional Steel Research, Vol. 46B, 657-665. (in Japanese) Kawano, A. (1998), "On the effect of restoring force characteristics of column-base on inelastic response behavior of weak-beam steel frame under earthquake ground motion," Journal of Stract. Constr. Eng., AIJ, Vol. 507, 139-146. (in Japanese) Morino, S., Kawaguchi, J., Tsuji, A., and Kadoya, H. (2003), "Strength and stiffness of CFT semi-embedded type column base," Proceedings of the International Conference on Advances in Structure, Sydney, Australia, June 2003, 3-14. Yamada, T. and Akiyama, H. (1997), "Influence of the rigidity of column bases on the ultimate earthquake resistance of multi-story steel moment frames," Journal of Stract. Constr. Eng., AIJ, Vol. 496, 113-118. (in Japanese)



Microsoft Word - 29 - Hitaka.doc

11 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Microsoft Word - 2 bolted connections.doc
Fasteners for structural timber: nails, screws, bolts and dowels