Euro-SiBRAM’2002 Prague, June 24 to 26, 2002, Czech Republic
Session 8
1. Introduction
The subject of this study is the safety assessment of members of the steel frame (see Fig. 1) by fully probabilistic SBRA method [3]. Study outline possibilities of the SBRA in the assessment of systems with multi component load effects combination. Assessment by the SBRA is compared with another one performed in accordance with the European ČSN P ENV 1993-1 [2] (shortly denominated as EC).
There are differences between both methods:
(a) EC [2], which uses load and resistance factors, looks more likely deterministic and reference value of safety is defined by plastic carrying capacity of the shape. The safety assessment according to EC is published in [4].
(b) SBRA [3] characterize random variables (e.g. loading) by their extreme and bounded histograms. The maximum (extreme) load effects according to SBRA [3] correspond to the factored load effects according to EC. Carrying capacity of the shape is defined by onset of yielding. Safety of the structure is expressed by probability of failure Pf < Pd. Next part of the paper describe assessment using SBRA method.
2. Assignment
Statically determinate plane steel frame (see Fig. 1) is a part of bent of industrial building. Loadings considered in this study are wind, snow, crane girder and dead load (Fig. 2). The second order effects, the buckling and torsion-buckling, are not considered in this study. Safety of the selected cross sections (see Fig. 1) is assessed. The effect of shear forces is neglected.
Fig. 1. Geometric chart of the frame
3. Geometrical and material properties
Proportion and lengths are considered as deterministic. Variation of cross-section area is characterized by appropriate histogram.
|
|
1 |
2 |
3 |
|
|
|
Symbol |
Column |
Column height |
Girder |
Variation |
Histogram |
|
|
HEB 400 |
HEB 400 |
IPE 600 |
|
|
Area [m2] |
A |
1,98E-02 |
1,98E-02 |
1,56E-02 |
±4% |
N1-04 |
Section modulus [m3] |
W |
2,89E-03 |
2,89E-03 |
3,07E-03 |
||
Steel [Mpa] |
fy |
S235 |
S235 |
S235 |
200-435 |
T235Fy01 |
4. Loadings
|
Table 2. Loadings |
|
|
|
|||
|
|
|
|
Nominal |
Factors |
Extreme |
|
Member |
No. |
Symbol |
Loading |
value |
Q |
value |
Histogram |
|
|
|
|
[kN, kN/m] |
|
[kN, kN/m] |
|
Girder |
3 |
gd3 |
Dead load |
5,55 |
1,1 |
6,11 |
Dead1 |
Girder |
4 |
gd4 |
Dead load |
5,55 |
1,1 |
6,11 |
Dead1 |
Column |
1 |
Gd1 |
Dead load |
40,00 |
1,1 |
11,00 |
Dead1 |
Column |
6 |
Gd6 |
Dead load |
40,00 |
1,1 |
11,00 |
Dead1 |
Cantilever |
7 |
Gd7 |
Crane track |
15,08 |
1,1 |
16,59 |
Dead1 |
Cantilever |
8 |
Gd8 |
weight |
15,08 |
1,1 |
16,59 |
Dead1 |
Column |
1, 6 |
wdeD |
Wind |
5,18 |
1,4 |
7,26 |
Wind1 |
Column |
6, 1 |
wdeE |
Wind |
-1,94 |
1,4 |
-2,72 |
Wind1 |
Beam |
3, 4 |
wdeG |
Wind |
-5,18 |
1,4 |
-7,26 |
Wind1 |
Beam |
3, 4 |
wdeH |
Wind |
-1,94 |
1,4 |
-2,72 |
Wind1 |
Beam |
4, 3 |
wdeJ |
Wind |
-6,48 |
1,4 |
-9,07 |
Wind1 |
Beam |
4, 3 |
wdeI |
Wind |
-2,59 |
1,4 |
-3,63 |
Wind1 |
Beam |
3, 4 |
sd |
Snow |
5,40 |
1,4 |
7,56 |
Snow1 |
Cantilever |
7 |
Qdj7 |
Crane girder |
173,25 |
1,4 |
242,55 |
Crane-V |
Cantilever |
8 |
Qdj8 |
Crane girder |
173,25 |
1,4 |
242,55 |
Crane-V |
Cantilever |
7 |
Bd |
Crane girder |
16,50 |
1,4 |
23,10 |
Crane-H |
Cantilever |
8 |
Bd |
Crane girder |
16,50 |
1,4 |
23,10 |
Crane-H |
Planar frame is exposed to dead load, wind load, snow load and crane girder (vertical and horizontal forces) (see Fig. 2). Loadings are considered mutually statistically uncorrelated. Extreme load effects correspond to the factored load effects according to EC (National aplication document of [1])
SBRA allow to anylize wind or crane girder forces to act from the left side or the right side. Direction of forces depend on each simulation and is characterized by appropriate histogram.
5. Internal forces and load effects combination
Transformation model analyze inner forces of the frame (Axial forces and Bending moment). Model allow to investigate inner forces in the critical cross-sections (see Fig. 1) and allow to create inner forces scatter (see Fig. 4).
The single component load effects combination SSBRA (see /1/) investigate axial forces Ni [kN] or bending moment Mi [kNm] using bounded histograms (represented by GVAR and QVAR) and extreme value of each specific loading.
Fig. 3. Two component load effects combination (Mi - horizontal axis, Ni – vertical axi)s
Fig. 4. Scatter of bending moments M [kNm] (left) and axial forces N [kN]
SBRA method [3] allows making multicomponent load effects combination. Formula /2/ investigate stress si [kPa] in the critical cross-sections, where Ai [m2] is area, Wel,y,i [m3] is section modulus. Variables with subscript var are represented by appropriate histogram (see Table 1). Load effects combination, expressed by stress si, for axial forces Ni and bending moment Mi is done in one simulation step (see Fig. 3).
[kPa]
/2/
6. Reliability assessment
Reliabilty assessment by SBRA method is based on the limit state philosofy. Stochastic character of variables is represented by histograms.
SBRA analyze safety function SFi [kPa] /4/, where RSBRa,i /3/ [kPa] is resistence (defined by onset of yielding) and si [kPa] /2/ is stress due to load effect combination.
RSBRA,i = fy,i /gM = fy,i/ 1.15 [kPa] /3/
SFi = RSBRA,i - ABS (si) [kPa] /4/
Probability of failure Pf,i /5/ is function of safety function SFi [kPa] /4/ in the investigated fibre.
Pf = P(RSBRA,i - SSBRA,i<0) =P(SF<0) […] /5/
Safety condition Pf < Pd comply for all cross-sections
Table 3. Reliability assessment |
||||
Cross-section |
Shape |
Fibre |
Probability of failure Pf |
Reliability level |
1 |
HEB 400 |
p |
Pf = 0,000 002 < Pd = 0,000 070 |
Excessive safety |
l |
Pf < 0,000 001 < Pd = 0,000 070 |
|||
2 |
HEB 400 |
p |
Pf = 0,000 018 < Pd = 0,000 070 |
Usual safety |
l |
Pf = 0,000 006 < Pd = 0,000 070 |
|||
3 |
IPE 600 |
d |
Pf = 0,000 004 < Pd = 0,000 070 |
Excessive safety |
h |
Pf < 0,000 001 < Pd = 0,000 070 |
Table 4. |
||
|
Designed rolled shapes |
|
|
SBRA |
EC |
1-1 |
HEB 400 |
HEB 500 |
2-2 |
HEB 400 |
HEB 450 |
3-3 |
IPE 600 |
HEB 450 |
7. Conclusion
Assessment of the planar frame according to EC [2] has 144 cases of the load effects combinations. Combination of the load effects for a three-dimensional frame could has many more load effects combinations.
Study denotes possibilities of the SBRA [3] in analyzing of the multi component load effects combination and determining of the steel frame reliability.
In this example are shapes designed according to EC [1], [2] more conservative than designs according to SBRA providing considered conditions (see Table 4).
References
[1] ČSN P ENV 1991-1, Basis of design and actions on structures, Part 1: Basis of design Eurocode 1, ČNI, Prague, 1994.
[2] ČSN P ENV 1993-1-1, Design of steel structures, Part 1-1: General rules and rules for building, Eurocode 3, ČNI, Prague, 1994.
[3] Marek, P., Brozetti, J., GUŠTAR, M. (editors), Probabilistic Assessment of Structures using Monte Carlo Simulation, ITAM CAS, Prague, 2001.
[4] Konečný, P., Master thesis: Posudek spolehlivosti vybraných konstrukčních dílců podle norem eurocode a metodou SBRA, FAST VŠB-TUO. Ostrava 2002.