TSINGHUA SCIENCE AND TECHNOLOGY ISSN 1007-0214 17/25 pp681-688 Volume 13, Number 5, October 2008

Simplified Beam Design for Semi-Rigid Composite Frames at the Serviceability Limit State*

WANG Jingfeng (王静峰)1,2,**, LI Guoqiang (李国强)3

1. School of Civil Engineering, Tsinghua University, Beijing 100084, China; 2. School of Civil Engineering, Hefei University of Technology, Hefei 230009, China;

3. School of Civil Engineering, Tongji University, Shanghai 200092, China

Abstract: This paper presents a simplified beam design method for semi-rigid composite frames with vertical

loading at the serviceability limit state. Equations were developed to determine the deflections of the com-

posite beam allowing for both joint flexibility and beam sectional properties, along with a formula for the

connection secant stiffness. The equations for the connection stiffness are more accurate than previous

equations used because it considers the beam-to-column stiffness ratio and the beam-to-connection stiffness

ratio. The equations were validated by the experimental results for two semi-rigid composite frames. The

equations agree well with the experimental data because they take into account the actual beam-to-column

connections and the composite action between the steel beam and the concrete slab.

Key words: semi-rigid; composite frame; composite action; deflection; connection stiffness; serviceability

Introduction

Traditionally, composite beam-to-column joints are designed either as perfectly pinned or rigid connections. However, the actual behavior of the composite joints is semi-rigid with a range of moment-rotation character-istics. The benefits of adopting the semi-rigid joint design have become more obvious in recent years. Most recently built steel buildings have used concrete floor slabs designed to act as composites with steel beams by means of shear connectors[1]. There are also economic and structural benefits to utilize the partially restrained composite connections with some degree of continuity and without the disadvantages associated with the fully rigid approach. Thus, the semi-rigid composite frames are very useful and the effects of the semi-rigid connections and composite action of the

slab should be properly considered in the design of steel frames.

When building frames are subjected to vertical and lateral loads, the distribution of the bending moment in the composite beams varies along the member length. In the negative moment region, the concrete is in ten-sion and may crack and the steel reinforcement in the slab may be yielded. In the positive moment region, a large bending moment may cause yielding of the steel section and crushing of the concrete. Thus, the effects of the location vary as the moment varies along the beam.

Over the past thirty years, extensive studies[2-4] have been carried out to understand the actual behavior of semi-rigid connections. Many studies[1] have used the nonlinear numerical analyses of frame systems with semi-rigid connections. However, little effort has been devoted to developing simplified methods to calculate the deflection of semi-rigid composite frames. Al-though Wong et al.[5] proposed a set of equations and design charts to calculate the deflections of the com-posite beams according to the beam line theory, they

Received: 2007-04-06; revised: 2007-06-21, 2008-04-08

* Supported by the National Science Fund for Distinguished YoungScholars (No. 50225825)

** To whom correspondence should be addressed. E-mail: jfwang@hfut.edu.cn; Tel: 86-551-2919884

Tsinghua Science and Technology, October 2008, 13(5): 681-688 682

did not give an equation to determine the connection rotational stiffness at the serviceability limit state (SLS). This paper presents a simplified practical beam design method for semi-rigid composite frames at the serviceability limit state with vertical loading which takes into account both the joint flexibility and the beam sectional properties. An equation is given for the appropriate formula of connection secant stiffness and the design procedure is developed.

1 Simplified Deflection Equation

The internal forces and deformations in each individual beam of a multistory frame with vertical loads are usu-ally evaluated by considering a plane frame (Fig. 1a).

The column ends far from the beam under investiga-tion can be assumed to be fixed. Each beam is assumed to have an effective value of the second moment of area, Ib, which approximately simulates the sectional properties of the composite beam. Due to the flexible natural of semi-rigid connections, the contributions to the end restraint from the beams on either side of the beam under consideration are generally insignificant. Consequently, a simplified subframe may be used to determine the moments and deflections of the composite beam with semi-rigid connections as shown in Fig. 1b. The simplified subframe can then be transformed to a beam line model shown as in Fig. 1c where Rks and Rc are the rotational stiffnesses of the semi-rigid beam-to- column connections and the columns.

bL

ksR ksR

q

(b) Simplified subframe model

(a) Frame model

bL

ksR ksR

qq q

bL

q

0=l 0=l

cR ksR cRksR

(c) Beam line model

c2h

c1h

c2h

c1h

Fig. 1 Calculation model

The rotation of beam-to-column connection, θr, based on the slope deflection method and the symmet-rical properties of the beam line model, can be ex-pressed as

3b b

r b cb b c24 2

qL ML MEI EI R

θ θ θ= − = − − (1)

where θb and θc are the support rotation of the beam and the column rotation at the joints, q is the uniformly distributed load, E is Young modulus of elasticity, M is the support moment of the beam, Ib is the effective second moment of area of the composite beam and

1 c1 2 c2c

c1 c2

a EI a EIRh h

= + (2)

where Ici is the second moment of area of the column at level i, hci is the story height for level i, and ai is the column stiffness coefficient at level i. For the first story, ai = 3 for columns fixed and equally for columns pinned at their remote ends, while ai = 4 for other stories.

In Eq. (1), when the support moment of the beam is equal to zero, the support rotation of the beam is θpin, i.e., the support rotation of the beam with pinned con-nections; when the support rotation of the beam is zero, the support moment of the beam is Mrigid, i.e., the sup-port moment of the beam with rigid connections. Mrigid and θpin are expressed as

3b b

rigidb b c

124 2

qL LMEI EI R

⎛ ⎞= +⎜ ⎟

⎝ ⎠ (3)

WANG Jingfeng (王静峰) et al：Simplified Beam Design for Semi-Rigid Composite Frames at the ... 683

3b

pinb24

qLEI

θ = (4)

At the serviceability limit state, the beam-line equa-tion and the moment-rotation curve of the connection are assumed to determinate the tangent connection stiffness, Rkt,E , and the secant connection stiffness, Rks,E, as shown in Fig. 2. The linear secant connection stiffness, Rks,E, can be used to represent the connection characteristics. The moment and rotation of the con-nection are utilized by

r ks,E/M Rθ = (5)

Moment-rotation curve

rθEθ

M

rigidM

EM

kiR

pinθ

Beam line equation

Eks,R Ekt,R

O

Fig. 2 Connection stiffness at SLS

From Eqs. (1) and (5), the moment of connection is 3

rigidb

rigidb b

pin ks,Eks,E b c

124 1 1 1

2

MqLM MEI LRR EI R θ

= ⋅ =⎛ ⎞ ++ +⎜ ⎟⎜ ⎟⎝ ⎠

(6)

A dimensionless factor μ is defined to characterize the connection rigidity.

rigidrigid

pin ks,E

/ 1 1M

M MR

μθ

⎛ ⎞= = +⎜ ⎟⎜ ⎟

⎝ ⎠ (7)

When the beam-to-column connection is pinned, μ = 0; when it is rigid, μ =1; and when it is semi-rigid, 0<μ <1.

The composite beam with semi-rigid connections with a uniformly distributed load (UDL) is shown in Fig. 3. According to the slope deflection method and the integral method, the deflection of the beam with semi-rigid connections is

b b

b

/ 2

1 20 / 2b b

1 1( ) d ( ) dL L

LM x M x M x M x

EI EIδ = + =∫ ∫

4 2 2b 1 b 2 b

b b b

5384 16 16

qL M L M LEI EI EI

− − (8)

where

2b 1 21

b b

1( )2 2

qL M MM x x qx ML L

⎛ ⎞= + − − −⎜ ⎟⎝ ⎠

,

1 / 2M x= , b0 / 2x L< < ;

2 (1 ) / 2M x= − , b b/ 2L x L< . For a composite beam with semi-rigid connections

and a uniformly distributed load, the left and right support moments of the beam can be assumed to be M1= M2 = M. Thus, Eq. (8) can be simplified to

4 2b b

b b

5384 8

qL MLEI EI

δ = − (9)

when the beam-to-column connection is pinned, pinδ = 4b

b

5384

qLEI

, and when it is rigid, 24

rigid bbrigid

b b

5384 8

M LqLEI EI

δ = − .

Thus, pin

pin rigid rigid

MM

δ δδ δ

−=

− (10)

From Eqs. (7) and (10), the deflection of a semi-rigid composite frame is approximately

rigid pin rigid(1 )( )δ δ μ δ δ= + − − (11)

where δ is the deflection of the semi-rigid composite frame with the vertical loads, rigidδ is the deflection of

the corresponding rigid composite frame with the ver-tical loads, and pinδ is the deflection of the corre-

sponding simply supported composite frame with the vertical loads.

q

bL

1M 2Mq

1 2

bEI)(xM

x

bL

bL

2M1M

Fig. 3 Semi-rigid composite beam under UDL

For an interior span beam, the support moments of the beam are approximately equal. For an exterior span, the connection moments of the external columns are approximately 25%-30% of the connection moments of internal columns. Therefore, the average value, μave, of the connection moments of the external and internal columns is used in Eq. (11).

Tsinghua Science and Technology, October 2008, 13(5): 681-688 684

2 Main Parameters 2.1 Moment-rotation relationship of the

connections

Modeling of the beam-to-column connections requires representation of the nonlinear moment-rotation, M-θr. The Kishi-Chen’s three-parameter power model[4] of semi-rigid connections can be used to represent the nonlinear M-θr relation of a composite connection. This model includes the initial connection stiffness, Rki, the ultimate moment capacity of the connection, Mu, and the shape parameter, n. The moment used for rela-tive rotation, θr, is

ki r1/

r 0[1 ( / ) ]n n

RM θθ θ

=+

(12)

where θ0 is the reference plastic rotation given by θ0 =

Mu/Rki . This power model is an effective tool for designers

to quickly and accurately execute the second-order nonlinear structural analysis. The tangent connection stiffness, Rkt, can be expressed as

kikt ( 1)/

r r 0

dd [1 ( / ) ]n n n

RMRθ θ θ += =

+ (13)

Figure 4 shows the calculation models of the com-posite flush endplate connection. With the applied forces based on the component’s resistance defined in EC3[6] and assuming that the neutral axis lies in the beam web (the most common situation), the connec-tion moment resistance can be obtained by taking mo-ments with respect to the midpoint of the beam’s bot-tom flange[2]:

w fbu tr r b w wb ywb

0

( )2 2

m

i ii

h tM F L F L h t f=

⎛ ⎞= + − +⎜ ⎟⎝ ⎠

∑ (14)

where Ftr is the tensile force in the reinforcing steel, Fbi is the tensile force at bolt row i, hw is the height of the beam web in compression, twb is the beam web thick-ness, tfb is the beam flange thickness, Lr is the distance from the reinforcing steel to the center of the beam’s bottom flange, Li is the distance from bolt row i to the center of the beam’s bottom flange, and fywb is the yield strength of the beam web.

The initial connection stiffness is given by[6,7] 2

kir

1i i

M EzR

kθ

= =∑

(15)

where z is the level arm, ki is the stiffness coefficient of

component i, and E is Young modulus for steel.

Reinforcement

bft

trF

b1F

cF uM

(a) Moment capacity

rθ

z

rk

bok tcw,kcfk epk

ccw,k vcw,k

(b) Initial rotational stiffness[7]

Fig. 4 Calculation models of the composite connection

2.2 Effective composite beam stiffness

The analysis of composite frames must consider the composite effect of the steel beam and the concrete slab on the frame behavior. Because the moment in the frame beam varies at different locations, the effective stiffness of the composite beam also varies for loca-tions where the moment puts the concrete slab into compression or tension. Despite this apparent com-plexity, Ammerman and Leon[8] proposed an effective second moment of inertia that gives acceptable predic-tion of the beam behavior in a composite frame:

b pos neg0.6 0.4I I I= + (16)

where Ib is the effective second moment of area of the composite beam, and Ipos and Ineg are the second mo-ments of area of the composite beam section in the positive and negative regions.

2.3 Connection rotational stiffness

The simplified deflection equation also includes the connection stiffness, Rks,E. A linear tangent stiffness is normally used to represent the connection characteris-tics with the initial rotational stiffness, Rki , frequently

WANG Jingfeng (王静峰) et al：Simplified Beam Design for Semi-Rigid Composite Frames at the ... 685

used in analyses of flexible frames. However, the ini-tial rotational stiffness, Rki , can produce unacceptable overestimates of the connection stiffness[9] and uncon-servative underestimates of the fame deflection.

Six methods have been proposed to determine the connection stiffness: (1) the initial stiffness of connec-tion, Rki , (2) the connection secant stiffness, R2.5, pro-posed by the ASCE Task Committee[10] for the seismic design of semi-rigid composite connection, which cor-responds to a rotation of 0.0025 rad, (3) the connection secant stiffness, R10, proposed by Bjorhovde[11] , which corresponds to a rotation of 0.01 rad, (4) the connec-tion secant stiffness, Rk0, based on the power function model[9], which corresponds to the reference plastic rotation, θ0, (5) the connection secant stiffness of RY, defined in Eurocode 3[6], which corresponds to the design moment capacity MY = 2Mu/3, and (6) the half rotational stiffness, Rhalf = Rki / 2, given in Eurocode 3 Appendix J[12]. The various connection stiffnesses are compared in Fig. 5.

These various connection stiffnesses have been de-termined according to experience, without considering the beam-to-column stiffness ratio and the beam- to-connection stiffness ratio. Therefore, these methods may inaccurately predict the deflection of composite beams.

For a composite beam with semi-rigid connections under a uniformly distributed load, the rotational stiff-ness of the connection at the serviceability limit state

M

uM

YM

0θ

kiRk0R

2.5

2.5R YR halfR 10R

10Yθ rθ0 (mrad)

Fig. 5 Various connection stiffnesses

of the beam, Rks,E, can be expressed as proposed by Wang[13]

ks,E 1 kiR Rη= (17)

where η1 is 0

1/ 21

0

1

0.5[1 ]

1 2 2

n n

α βηη α βη

+ +− =

+ + (18)

where

( 1) /2 (1 0.667 )n n nη += − , 0 b ki b/( )EI R Lα = , and β =

b c b/ ( )EI R L . For a flush end-plate connection, the shape parame-

ter, n , in Eq. (18) is equal to 1.5. The stiffness factor 1η is then obtained from Table 1 or Fig. 6. Equation

(18) can also be used for composite frames with one point load (1PL) or two point loads (2PL).

Table 1 Factor η1 (n=1.5)

β α0 = 0 α0 = 0.01 α0 = 0.02 α0 = 0.04 α0 = 0.08 α0 = 0.10 α0 = 0.25 α0 = 0.50 α0 = 0.75 α0 = 1.00 α0 = 1.50 α0 = 2.00 α0 = 10.00

0 0.748 0.729 0.711 0.678 0.625 0.604 0.522 0.479 0.464 0.456 0.448 0.443 0.434 0.10 0.748 0.732 0.717 0.688 0.641 0.621 0.536 0.488 0.470 0.461 0.451 0.446 0.434 0.25 0.748 0.735 0.723 0.699 0.658 0.641 0.556 0.501 0.479 0.468 0.456 0.450 0.435 0.50 0.748 0.738 0.729 0.711 0.678 0.663 0.583 0.522 0.494 0.479 0.464 0.456 0.436 0.75 0.748 0.740 0.732 0.718 0.690 0.678 0.604 0.540 0.508 0.491 0.472 0.462 0.437 1.00 0.748 0.741 0.735 0.723 0.699 0.688 0.621 0.556 0.522 0.501 0.479 0.464 0.438 1.25 0.748 0.742 0.737 0.726 0.706 0.696 0.635 0.570 0.534 0.512 0.487 0.474 0.440 1.50 0.748 0.743 0.738 0.729 0.711 0.702 0.646 0.583 0.545 0.522 0.494 0.479 0.441 2.00 0.748 0.744 0.740 0.732 0.718 0.711 0.663 0.604 0.566 0.540 0.508 0.491 0.443 3.00 0.748 0.745 0.742 0.737 0.726 0.721 0.684 0.635 0.598 0.570 0.534 0.512 0.448 4.00 0.748 0.745 0.743 0.739 0.731 0.727 0.697 0.655 0.621 0.594 0.556 0.531 0.453 5.00 0.748 0.746 0.744 0.741 0.734 0.730 0.706 0.670 0.639 0.613 0.575 0.548 0.458

10.00 0.748 0.747 0.746 0.744 0.740 0.738 0.725 0.704 0.684 0.666 0.635 0.609 0.482

Tsinghua Science and Technology, October 2008, 13(5): 681-688 686

Fig. 6 η1 and β relation (n=1.5)

3 Design Procedure

The deflections of a semi-rigid composite frame under a vertical load can be predicted as follows.

(1) Select the beam and column size and the connec-tion type according to previous experience.

(2) Calculate the connection moment capacity, Mu, and the initial rotational stiffness, Rki , using Eqs. (14) and (15).

(3) Calculate the cross-sectional properties posI and

negI of the composite beam according to the beam-end

constraints and the beam details. The effective second moment of area of the composite beam, Ib, is given by Eq. (16).

(4) Calculate the service load (1.0gk+1.0qk), where gk is the service dead load and qk is the service live load.

(5) Calculate the beam deflection, δrigid, and the connection moment, Mrigid, by the elastic analysis of the rigid composite frame.

(6) Calculate the beam deflection, δpin, and the con-nection rotation, θpin, by the elastic analysis of the sim-ple composite frame.

(7) Calculate the factor η1 by Eq. (18) or Table 1, and the appropriate rotation stiffness of the connection, Rks,E, by Eq. (17).

(8) Calculate the factor μ by Eq. (7) and the average μave can be determined.

(9) Calculate the deflections of a semi-rigid com-posite frame by Eq. (11). This equation is subject to the limit b / 400Lδ as specified by GB50017[14]. If this limit is exceeded, select revised beam size and/or con-nection type and repeat Steps (1)-(7) until the deflec-tion is satisfactory.

4 Experimental Validation

The proposed method was validated for a pair of full-scale semi-rigid composite frames with two stories and two bays[15]. The predicted moments and deflec-tions in the semi-rigid composite frames were then compared with test results.

The service loads on the test frames are shown in Fig. 7. The steel beams were connected to the column flanges by means of flush end plates 14 mm thick and two rows of Grade 10.9 M22 bolts. A 140-mm-high concrete slab was supported by profiled steel sheet decking placed longitudinally with welded-through stud shear connectors providing the composite action with the steel beam. The width of the concrete slab was 1.5 m as specified in GB50017[14].

190 kN 190 kN 187 kN 187 kN

1.7 m 1.7 m1.6 m 1.7 m 1.7 m1.6 m

(a) Frame A

1.8

m3.

0 m

185 kN 185 kN 85 kN 85 kN

1.7 m 1.7 m1.6 m 1.7 m 1.7 m1.6 m

(b) Frame B

1.8

m3.

0 m

Fig. 7 Test frames and service loads

The slab reinforcement ratio for each specimen was 0.98%. The steel columns were stiffened with trans-verse stiffeners welded to the webs of the columns level with the top and bottom beam flanges. The rein-forcements were high yield deformed bars. One layer of equally spaced 10-mm-diameter longitudinal rein-forcements was attached over the width of the slab beside the column. Two layers of 6-mm-diameter de-formed bars were used as transverse reinforcement to prevent longitudinal splitting failure of the concrete slab. They were deliberately cut off at the plane of the beam-to-column connection to prevent the bottom layer of the longitudinal bars from contributing to the moment resistance. The design of the transverse

WANG Jingfeng (王静峰) et al：Simplified Beam Design for Semi-Rigid Composite Frames at the ... 687

reinforcement and shear connectors for the two speci-mens were based on BS5950[16] and GB50017[14], with full composite design assumed.

The trough height metal decking was 76 mm with a width of 344 mm. The composite beams were designed for full shear interaction with four 120 mm×19 mm headed shear studs welded in each trough of the metal decking. Details of the test results were given by Wang and Li[15].

The deflections of the composite beams in the frame at service loads were characterized by the beam loads corresponding to mid-span deflections of 1/400 of the beam span length[14] and were given in Table 2. The

design method was then used to calculate the mid-span deflections for these codes for each of the beams as-suming semi-rigid connections with the results also presented in Table 2. For comparison, the mid-span deflections for each of the beams assuming either rigid or pinned joints are also listed in Table 2 for the same beam loads. The results confirm that the predicted de-flections agree quite well with the test results, while the rigid connection assumption underestimates the beam deflections by 47% and the pinned connection assumption overestimates the beam deflections by 26%.

Table 2 Comparison of predicted and measured deflections

Pi / kN δ t / mm δ / mm δ pin / mm δ rigid / mm δ / δ t δ pin /δ t δ rigid /δ t Beam-1 190 12.5 11.9 22.1 6.5 0.95 1.77 0.52 Beam-2 187 12.5 11.6 21.7 6.4 0.93 1.74 0.51 Beam-3 185 12.5 11.5 21.4 7.0 0.92 1.71 0.56 Average 0.93 1.74 0.53

Note: Pi is one of the two point loads which give a deflection equal to Lb /400, δ t is the test defection which is equal to Lb /400, δ is the predicted defec-tion for each of the beams with semi-rigid connections for the same load, δpin is the defection for each of the beams assuming pinned connections for the same load level, and δrigid is the defection for each of the beams assuming rigid connections for the same load.

5 Conclusions

A simplified design method was developed to predict the deflection of semi-rigid composite frames with vertical loads at the serviceability limit load, including the effecting semi-rigid joints and the beam sectional properties. Comparisons against deflection measure-ments taken on semi-rigid composite frames show that the method is sufficiently accurate. In addition, the ultimate limit state theory was used to develop an ap-propriate formula for the rotation secant stiffness of the connection at the serviceability limit state considering the beam-to-column stiffness ratio and the beam-to- connection stiffness ratio, which is more accurate than previous methods.

Acknowledgements

The authors acknowledge the assistance of Dr. Liu Qingping of Tongji University and Dr. Hou Hetao of Shandong University.

References

[1] Liew J Y R, Chen H, Shanmugam N E. Inelastic analysis of steel frames with composite beams. Journal of Struc-tural Engineering, ASCE, 2001, 127(2): 194-202.

[2] Simoes D S L, Simoes R, Cruz J S. Experimental behavior of end-plate beam-to-column composite joints under monotonical loading. Engineering Structures, 2001, 23: 1383-1409.

[3] Xiao R Y. Analysis and design of composite connections and frames. Progress in Steel Building Structures, 2002, 4(3): 13-19. (in Chinese)

[4] Kishi N, Chen W F. Moment-rotation relation of semi-rigid connections with angles. Journal of Structural Engineering, ASCE, 1990, 116(7): 1813-1834.

[5] Wong Y L, Chan S L, Nethercot D A. A simplified design method for non-sway composite frames with semi-rigid connections. The Structural Engineer, 1996, 74(2): 23-28.

[6] CEN. Eurocode 3: Design of steel structures. Part 1.1: General rules and rules for buildings. ENV 1993-1-1. Brussels, 2005.

[7] CEN. Eurocode 4: Design of composite steel and concrete structures. Part 1.1: General rules and rules for buildings. ENV 1994-1-1. Brussels, 2004.

[8] Ammerman D J, Leon R. Unbranced frames with semirigid composite connections. Engineering Journal, AISC, 1990, 27(1): 1-10.

[9] Baraket M A, Chen W F. Practical analysis of semi-rigid frame. Engineering Journal, ASCE, 1990, 27(2): 54-68.

Tsinghua Science and Technology, October 2008, 13(5): 681-688 688

[10] ASCE Task Committee. Design guide for partially restrained composite connections. Journal of Structural Engineering, 1998, 124(10): 1099-1114.

[11] Bjorhovde R. Effect of end restraint on column strength-practical application. Engineering Journal, AISC, 1984, 20(1): 1-13.

[12] CEN. Eurocode 3, Annex J: Design of steel structures- joints in building frames. European Committee for Stan-dardization Document CEN/TC250/SC3, Brussels, 1998.

[13] Wang J F. The practical design approach of semi-rigidly connected composite frames under vertical loads

[Dissertation]. Tongji University, Shanghai, China, 2005. (in Chinese)

[14] GB50017. Code for design of steel structures. Beijing: China Construction Industry Press, 2003. (in Chinese)

[15] Wang J F, Li G Q. Testing of semi-rigid steel–concrete composite frames subjected to vertical loads. Engineering Structures, 2007, 29(8): 1903-1916.

[16] BSI. BS5950, structural use of steelwork in building. Part 1: Code of practice for design of simple and continuous construction. London, 2000.

Tsinghua University’s Youngest Professor

Born in 1977, Dr. Nieng Yan is currently Tsinghua University’s youngest professor. As a Tsinghua undergraduate, she developed a strong interest in science and was deeply influenced by Beijing’s unique civil milieu. After receiv-ing a bachelor’s degree in biology in 2000, she traveled to New Jersey to pursue graduate training in the Depart-ment of Molecular Biology at Princeton University. Under the guidance of Dr. Yigong Shi, she used structural bi-ology and biochemistry techniques to elucidate the molecular mechanisms of cell death regulation. After Dr. Yan received her Ph. D. in December 2004, she continued her research at Princeton University as a research associate, focusing on the structural and functional characterization of intramembrane proteases.

Dr. Yan accepted an offer to become a full professor in the School of Medicine from Tsinghua University in Oc-tober 2007. Her research centers on the structural and biochemical studies of disease related membrane proteins. Up to now, Dr. Yan has published 17 peer-reviewed articles. Among them, she was the co-corresponding author of a comprehensive review article published in the Annual Reviews of Cell and Developmental Biology and an invited preview published in Cell. She was the first author of eight research articles published in Nature, Molecular Cell, Nature Structural & Molecular Biology, etc. In 2005, Dr. Yan was the only representative from North America to win the prestigious Young Scientist Award co-sponsored by Science/AAAS and GE Healthcare.

(From http://news.tsinghua.edu.cn, 2008-07-04)