JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126 119

Research Paper

Reliability assessment of frame steel considering semi-rigid connections

TrongHa Nguyen *
Department of Civil Engineering, Vinh University, Vietnam , 182 Le Duan street, Vinh city, Vietnam

ARTICLE INFO ABSTRACT

Article history :

Received : 10 July 2018

Initial stiffness and moment resistance are two parameters that formed the hardness of
Revised : 9 January 2019 the connection between columns and beams. This problem has been studied based on
empirical of many authors in the world. In this research been used initial stiffness and
Accepted : 11 January 2019 moment resistance models the proposed model of Aleksander K., for reliability
assessment of frame steel under the internal force of semi-rigid joints when random
input parameters. A stochastic model for the design of frame steel considering semi-
Keywords: rigid and Monte Carlo reliability analysis. This program is written by using the Matlab
programming language. Then, various numerical tests are performed to illustrate the
Semi-rigid connections
studied method.
Steel frame

Reliability assessment

Monte Carlo simulation

1 Introduction

Steel frame with semi-rigid connections is always an interesting topic of the scientists all over the world. In 1934,
Batho and Rowan proposed the straight-line common beam method to classify semi-rigid connection [1] Rathburn, in 1936,
considered the stiffness of the connection in his research on the moment distribution method [2] Baker, William and
Sourochnikoff, in [3] investigated the influence of semi-rigid connections on the frame subjecting to the shake. Monforton
and Wu [4, 5] were some of the first researchers applying the stiffness matrix method to analyze the plane frame with semi-
rigid connection. In which, the stiffness matrix and nodal force vector of each element depend on the linear stiffness of the
connection. Kim, S.E. and Choi, S.H proposed a novel method to analyze the spatial frame considering the material
nonlinearity and large deflection effect [6]. The linear and nonlinear semi-rigid connection models were also interested by
Hadianfard and Razani in 2003 [7].
In 1990, Kiureghian et al. assessed the reliability of the steel frame under the dynamic loads generated from the

* Corresponding author. Tel.: +084 942809698.

E-mail address: trongha@[Link]

e-ISSN: 2170-127X,
120 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126

earthquake El Centro in 1940 using β probability index method [8] The same method was used in [9] by Hong et al. to
study the steel frame according to the plastic limit state. Reliability of the steel structure under the corrosion was considered
by Melchers in [10]. Hadianfard and Razani used the Monte Carlo simulation method to investigate the steel frame with
flexible beam-column joint. In 2012, to study the reliability assessment of steel scaffold shoring structures for concrete
formwork was considered HaoZhang et al in [11]. The rotational stiffness and the moment resisting of the connection were
referred in [12] of Chen et al. This model of flexible joint, according to recommendation of Aleksander in [13], is not
correct. This author proposed then a novel model based on the empirical tests and the Eurocode 3.
So can be noticed that reliability assessment of steel frame considering semi-rigid connections are important. However,
semi-rigid models and reliability assessment methods are relatively limited. Therefore, this research aims to study
reliability assessment of steel frame considering semi-rigid connections under durable conditions of the beam to column
connections (joint) are elastic-plastic models, this value is determined through research of Aleksander K, [13] by Monte
Carlo simulation.

2 Initial stiffness and moment resistance of joints steel

Initial stiffness and moment resistance are two parameters that formed the hardness of the connection between columns
and beams. These two values were calculated using the proposed model of Aleksander K. et al [13]. The reliability of this
model has been verified by comparing the experimental results according to Euro code 3 standard (EC. 3) (Figure 1).

Fig. 1 – Validation of the Aleksander's model by comparison with EC.3 and experimental results [13]-

The moment resistance of steel M Rd and the initial stiffness of joints steel S j ,ini are determined as the following.

M Rd = 7, 4 × 10−5 × hc0.62 × hb1.2 × t 0.4

p ×d
0.85
(1)

S j ,ini =K1 × hc0.44 × hb1.2 × t 0.35

p × d 0.005 − K 2 (2)

where S j ,ini is initial stiffness ( kN .m / rad ); S j is the elastic stiffness of the connection ( kN .m / rad ) which is
determined as:
S j ,ini
Sj = (3)
η

hc (mm) is the height of the column section (HEB); hb (mm) is the height of the beam section (IPE); t p (mm) is the
thickness of the end plate and d (mm) is the bolt diameter;
= K1 1.5
= and K 2 19211 are identified from experimental
results [13].

3 Internal forces analysis frames of semi rigid connections

Consider a semi-rigid connections plane frame as shown in Figure 2. The internal forces of the frame are determined
using the displacement method. The basic system, the position of considered sections are also presented in Figure 2 and the
 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126 121

equilibrium equations are shown by (4).

q Z3 Md1 Md3
P Z1 Z2 Mc1 Mc3
Md2

H
H
Mc2 Mc4
l l l

Load-bearing frame Basic system Position of considered sections

Fig. 2 – Semi-rigid connections plane frame, basic system and position of considered sections

 r11 r12 r13   Z1   R1P 

r    
 21 r22 r23   Z 2  +  R2 P  =
0 (4)
 r31 r32 r33   Z  R 
 3   3P 

r11 =
r22 =
4i +
4 S j 3i + S j ( ;
) r12 =
r21 =
4i +
2 S 2j i
;
EI b
i=
12i 2 + 8S j i + S 2j 12i 2 + 8S j i + S 2j L
6i 6i 12i
where r13 =
r31 =
− ; r23 =
r32 =
− ; r33 =
l l l2

R1P = −
2
(
1 ql S j 6i + S j
;
) R2 P =
2
(
1 ql S j 6i + S j ); R3 P = − P
12 12i 2 + 8S j i + S 2j 12 12i 2 + 8S j i + S 2j

Moments of plane frame M p and moments of joints M 1 , M 2 are determined as follows.

4 EI 6 EI C 4 EI 6 EI C
MP =∑ M j .Z j + M P0 ; M1 = C Z1 +
H H2
Z3 ; M 2 = C Z2 −
H H2
Z3 (5)

with E is the Young modulus of the steel and I b , I c are respectively the moments of inertia of the cross section of
beam and column.

4 Vadidation of design computer program

The design program consists of two parts: structural analysis and safety condition verification. The first part will be
validated by comparison with the result of SAP200 software. The second part was validated by the authors in [14]. The
validation is performed in two extremal cases: articulation joints (stiffness of the joint is zero) and rigid joints (stiffness of
the joint is infinitely high). Input parameters are shown in Table 1 and the results are shown in Table 2. We can observe
that the error is very small in the case of button joint (<0,2%) and slightly increases in the other case (<2,5%). This result
can be explained by the fact that in our program the rigid joint is modeled by setting a very high value to the stiffness of
the joints and it gives the error. This result confirms the reliability of our program.

Table 1 - Input parameters of the analysis

End
Bolt Material
Beam (cm) Column (cm) plate Load-bearing
(cm) (kN/cm2)
(cm)
L hwb bf tf tw H hwc bf tf tw tp d E f P(kN ) q(kN/cm)
500 30 20 2 2 400 30 20 2 2 2 1.6 2,1E4 21 100 0.05
122 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126

Table 2 - Validation of established program by comparison with Sap2000 software

Articulation joints Rigid joints
Element Section
SAP2000 Program error SAP2000 Program error
t t
Mc1 - - 7483,0 7531,8 0.65%
Columns Mc2 -20041,0 -20000 0.20% -11515,9 -11352,0 1.42%
Mc3 - - - -8900,0 -9019,9 1.35%
Mc4 19958,9 20000 0.21% 12100,7 12096,0 0.04%
Md1 - - - 7483,0 7531,8 0.65%
Beams Md2 - - - -8900,0 -9019,9 1.35%
Md3 1562,5 1562,5 0.0% 853,8 832,8 2.46%

5 Monte Carlo simulation method

Monte Carlo simulation is a method that uses the pseudo-random numbers to simulate the random property of the basic
variables and then directly estimate the reliability based on the concept of larger number law. If the safe region is defined
as f ( X ) > 0 , the failures probability of the system will be determined as follows:

=Pf I f ( X )<0 f X ( x ) dx
∫= E  I f ( X )< 0  (6)
 

1
 khi f (X) < 0
Where I f ( X )< 0 =  (7)
0
 khi f (X) ≥ 0

According to the statistical theory, if we have a number N of the sample of the random vector X, a sample that consists
of N values of I f ( X ) < 0 can be obtained, and the expectation of I f ( X ) < 0 can be approximated as the average of the sample as:

 N
= I
Pf E=  1 ∑ I if ( X )<0 (8)
 f ( X )< 0  N i =1

According Lemaire in [15], it is pointed out that the estimation in (8) is convergent and the confidence interval at 95%
of Pf value can then be calculated as follows:

 
  1 − Pf    1 − Pf 
Pf 1 − 200   ≤ Pf ≤ Pf 1 + 200   (9)
 NPf   NPf 
   

6 Reliability assessment joins of frame steel considering semi-rigid connections

6.1 Safe condition

According to Euro code 3 allows the uses initial stiffness for design structures. When used initial stiffness in joints
frame apparition moment resistance. Safety conditions of joints of frames then are written in the following.

M j , Rd
Mj ≤ (10)
n
where M j is the moment value of joints effect load-bearing; n is a safety factor

6.2 Deterministic model and uncertainty model

Deterministic model is the above internal forces analysis problem, in which the input parameters are those of geometry
 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126 123

( H , L , hc , bb , tb , hc , bc , tc , t p , d ), of Young modulus material (E), of load ( q ) and of flexible joint ( K1 , K 2 ). This model

can be written in form with X =  H , L , hc , bb , tb , hc , bc , tc , t p , d , q , E , α , K1 , K 2  .

M j = ℑ( X ) (11)

Uncertainty model is constructed based on the deterministic model by taking into account the randomness of some
input paramters. In this paper, we distinct two vector of input parameters: the first one of the parameters assumed to be
deterministic X1 =  H , L , hc , bb , tb , hc , bc , tc , t p , d , K1 , K 2  and the second one of the parameters assumed to be ramdom

X 2 (ω ) = [ q (ω ),α (ω ), E (ω ) ] with ω represents the randomness of the parameters. This model can be written in form.

M j (ω ) = ℑ ( X 1 , X 2 (ω ) ) (12)

6.3 Reliability assessment of the steel frame by Monte Carlo simulation

By introducing the uncertainty model in the Monte Carlo simulation method, we obtain the scheme of the reliability
assessment of steel frame as shown in Figure 3.

Fig. 3 – Scheme of the reliability assessment of steel frame using Monte Carlo simulation method

6.4 Convergence of the Monte Carlo simulation

Consider the steel frame as shown in figure 2 with the deterministic input parameters and the random input parameters
respectively presented in Table 3 and Table 4. The comppression load q (kN / m) , material ( E ) and α = qL P assumed to
be normal variable with mean µ and coefficient of variation CV = σ / µ .

Table 3 - Deterministic input parameters

End
Beam Column Bolt Material
plate
(cm) (cm) (cm) (kN/cm2)
(cm)
L hwb bf tf tw H hwc bf tf tw tp d E f
500 30 20 2 2 400 30 20 2 2 2 1.6 2.1E4 21
124 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126

Table 4 - Random input variables and their representative parameters

Random variable E (kN / cm 2 ) q (kN / cm) α = qL P
Law of probability Norm Norm Norm
µE µq µq CVE µα CVα
Representative parameters
2,1× 10 4
10 1,30 0,15 0, 01 0,15

Figure 4 shows the convergence of the safe probability of the steel frame in the Monte Carlo simulation to the value of
0,8315 or 83,15% after about 998 sampling in 4,83 minutes. The used convergence criteria of 2,5% justifies the confidence
of the estimated reliability. This result also shows that although we have taken the safety factor is n = 1,1 in the analysis but
because of the randomness of some input parameters, the reliability of the structure is only of 83,15%. The assessment of
the reliability of the structure thus is neccessary.

Fig. 4 – Convergence of the safe probability in the Monte Carlo simulation

6.5 Effect of the coefficient of variation and safety factor

We know clearly that the variation of input random variables and the safety factor influence directly but inversely on
the safe probability of the structure. Thus in order to clear the effect of these parameters, one reconsiders the above column
with different coefficients of variation of the compression load CV = 0, 05; 0,1; 0,15; 0,2; 0,25 and different safety factors
n = 1,1; 1,15; 1,2; 1,25 . The results of the safe probability are listed in Table 5 and presented in figure 5.

Table 5 - Effect of the coefficient of variation and effect of the safety factor on the safe probability
CV
n
0,05 0,10 0,15 0,20 0,25
1,10 0,9449 0,7755 0,7316 0,6398 0,6020
1,15 0,9916 0,8898 0,8082 0,7102 0,6735
1,20 1,0000 0,9337 0,8684 0,7786 0,7459
1,25 1,0000 0,9755 0,9041 0,8276 0,7755

We can easily observe the effect inverse of the coefficient of variation and the safety factor in the Figure 5. The safe
probability of the column decreases when the coefficient of variation increases, whereas it increases when the safety factors
increases. This result seems to be obvious, but it has a very important significance. It shows that if there are many input
random parameters or furthermore with the high randomness in the structural design or in the optimization problem, the use
of the local coefficient such as the overload coefficient seems to be not sufficient. The structure can be in the dangerous
state. In this case, it is necessary to determine a global safety factor, as is done in this study, to assure the absolute safety of
 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126 125

the structure. For example, in this test, if the coefficient of variation is 0,05, the global safety factor needs to be only 1,25 to
obtain the safe probability 100%.

Fig. 5 –Effect of the coefficient of variation (left) and effect of the safety factor (right) on the safe probability of the
column

6.6 Effect of the ratio hb hc

In the section, the researchers investigated the impact of hardness ratio of beam and column to the reliability structure
through the ratio of beam cross-section and column cross-section hb hc . In [13] the value limit when applying the
hc
expression (1), (2) must satisfy the condition= (140 ÷ 400 ) m and=hb (160 ÷ 400 ) mm. Therefore ratio limits hb hc in
the research variable intervals hb =
hc ( 0,3 ÷ 2, 0 ) . Deterministic parameters and random variable with a coefficient of
variation ν = 0,1 ; safety coefficient n = 1,1 .

Figure 6 and Table 6 shown Effect of hb hc to the safe probability PS . Figure 6 shown that when hb =
hc 0,3 ÷ 0, 7 then
safe probability PS are decreases. when hb =
hc 0, 7 ÷ 2, 0 then safe probability PS is an increase. Structures had Safe
hc 0,5 ÷ 0, 7 . From this result, we see that for this frame type and the loading
probability is the smallest value when hb =
hc 0,5 ÷ 0, 7 and reasonable when hb hc ≈ 2, 0 .
should not choose a domain hb =

Fig. 6 – Effect of the Effect of hb hc to the safe probability PS

126 JOURNAL OF MATERIALS AND ENGINEERING STRUCTURES 6 (2019) 119–126

Table 6 - Safe probability PS when ratio hb hc

hb hc 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1

Ps 0,5474 0,3632 0,2947 0,2960 0,2974 0,3658 0,4524 0,4653 0,5158
hb hc 1,2 1,3 1,4 1,5 1,60 1,70 1,80 1,90 2,0
Ps 0,6105 0,6218 0,7132 0,7395 0,7921 0,8395 0,8507 0,8658 0,9158

7 Conclusion

This paper studies the reliability assessment of the steel frame with semi-rigid beam-column joint. The authors have
successfully established with the Monte Carlo simulation method to construct the steel frame’s reliability assessment
program. The parametric tests are also performed to study the effect of the input parameters on the reliability of the
structure. The obtained results showed the important significance of this research.
The results of this initial positive study are the premise for the authors to continue to expand the problem for other
types of structures and to consider various random variable.

REFERENCES

[1]- C. Batho, H.C. Rowan, Investigations on Beam and Stanchion Connections. Second Report of the Steel Structures
Research Committee. London, 1934.
[2]- J.C. Rathburn, Elastic Properties of Riveted Connections. T. Am. Soc. Civ. Eng. 101(1) (1936) 524-563.
[3]- B. Sourochnikoff, Wind Stresses in Semi-Rigid Connections of Steel Framework. T. Am. Soc. Civ. Eng.115(1)
(1950) 382–393.
[4]- G. Max, H. Benno, F. Markus, Reliability Analysis on Capacity Design Rules for Steel Frames. In: Proceedings of
International Conference on Seismic Design of Industrial Facilities, RWTH Aachen University, 2013 pp. 337-339.
doi:10.1007/978-3-658-02810-7_28
[5]- R. Monforton, T.S. Wu, Matrix Analysis of Semi-Rigidly Connected Frames. J. Struct. Div-ASCE. 89(6) (1963)
13-24.
[6]- S.-E. Kim, S.-H. Choi, Practical advanced analysis for semi-rigid space frames. Int. J. Solids Struct. 38(50-51)
(2001) 9111-9131. doi:org/10.1016/S0020-7683(01)00141-X
[7]- M.A. Hadianfard, R. Razani, Effects of semi-rigid behavior of connections in the reliability of steel frames. Struct.
Saf. 25(2) (2003), 123-138. doi:10.1016/S0167-4730(02)00046-2
[8]- D. Kiureghian, Structural reliability methods for seismic safety assessment: A review. Eng. Struct. 18(6) (1996),
412-424. doi:10.1016/0141-0296(95)00005-4
[9]- P. Hong S. Wang, Reliability of Steel frame systems with Semi-Rigid connections. ICLR Research Paper, Series
No.35, 2003.
[10]- R.E. Melchers, The effect of corrosion on the structural reliability of steel offshore structures. Corros. Sci. 47(10)
(2005) 2391–2410. doi:10.1016/[Link].2005.04.004
[11]- H. Zhang, K.J.R. Rasmussen, B.R. Ellingwood, Reliability assessment of steel scaffold shoring structures for
concrete formwork. Eng. Struct. 36 (2012) 81-89. doi:10.1016/[Link].2011.11.027
[12]- N. Kishi, W.F. Chen, Y. Goto, K.G. Matsuoka, Design aid of semi-rigid connections for frame analysis. Eng. J.
AISC. 30(3) (1993) 90-107.
[13]- K. Aleksander, K. Ryszard, G. Marian, Estimation of the Initial Stiffness and Moment Resistance of Steel and
Composite Joints. In: Proceedings of the CTBHU 8th world congrress, Dubai 2008, pp. 1-7.
[14]- T.H. Nguyen, Sensitivity analysis of input parameters in structural design problem and apply to analyze the
reliability of a design of steel industrial buildings, PhD thesis, National University of Civil Engineering, 2017.
[15]- M. Lemaire, A. Chateauneuf, J.-C. Mitteau, Reliability of Structures - Static Mechano-Reliability Coupling (in
French). Ed. Hermès, Coll. Civil Engineering, 2005.