Description of verified connection: hollow section steel column is anchored with anchor bolts to
concrete pad; the column is loaded by combination of axial force and bending moment.
Inputs
Compressed column is designed as maximal 3rd class to avoid stability problems.
The study was performed for parameters: size of the column, dimensions of base plate, grade of
concrete, dimensions of concrete pad.
Component method
Three components are examined: column flange and web in compression, concrete in compression
including grout, welds. This study is mainly focused on two components related closely to column
bases: “concrete in compression including grout” and “anchor bolt in tension”. Other components are
designed to have sufficient resistance and do not interfere with the overall resistance of the joint.
All components are designed according to EN 1993-1-8 [1], EN 1992-1-1 [2] and ETAG 001 – Annex C
[3].
Resistances related to concrete failure are designed according to European Technical Approval
Guidelines ETAG 001 Guideline for European Technical Approval of Metal Anchors for Use in Concrete
– Annex C: Design Methods for Anchorages. “Steel failure”, “Pull-out failure”, “Concrete cone failure”
and “Splitting failure” are relevant for anchor bolt in tension.
• Tension resistance:
𝑘2 ∙ 𝑓𝑢𝑏 ∙ 𝐴𝑠
𝐹𝑡,𝑅𝑑 =
𝛾𝑀2
• Punching shear resistance:
0.6 ∙ 𝜋 ∙ 𝑑𝑚 ∙ 𝑡𝑝 ∙ 𝑓𝑢
𝐵𝑝,𝑅𝑑 =
𝛾𝑀2
1
Brno University of Technology Column base: Hollow section column
Steel failure
Eurocodes do not refer sufficiently to concrete resistances in tension loading and interaction of shear
and tension loading. Therefore, the ETAG 001 – Annex C – 5.2.2.2 is used in this study for hand
calculations of steel failure mode for consistency. Software uses the standard Eurocode assessment.
Pull-out failure
Pull-out failure of anchor bolts loaded by tension is designed according to ETAG 001 – Annex C –
5.2.2.3. This failure mode is dependent on anchor type. It is presumed that the resistance will be
provided by manufacturer.
Splitting failure
Splitting failure of anchor bolts loaded by tension is designed according to ETAG 001 – Annex C – 5.2.2.5
Splitting failure due to installation and 5.2.2.6 Splitting failure due to loading. This failure mode is
relevant for anchorage to thin concrete columns or beams and is dependent on anchor type.
The thickness of the weld on the flanges is the same as the thickness of the weld on the web.
Design of the weld is done according to EN 1993-1-8 – 4.5.3.2(6).
Shear loading:
1,0
𝛾𝑀𝑠 = ≥ 1,25 for 𝑓𝑢𝑘 ≤ 800 MPa and 𝑓𝑦𝑘 /𝑓𝑢𝑘 ≤ 0,8
𝑓𝑦𝑘 /𝑓𝑢𝑘
𝛾𝑀𝑠 = 1,5 for 𝑓𝑢𝑘 > 800 MPa or 𝑓𝑦𝑘 /𝑓𝑢𝑘 > 0,8
𝛾𝑀𝑐 = 𝛾𝑐 ∙ 𝛾2
Tension loading:
2
Brno University of Technology Column base: Hollow section column
𝛾2 = 1,4 for systems with low but still acceptable installation safety
Concrete pad dimensions: 𝑎′ = 750 mm , 𝑏′ = 750 mm , ℎ = 800 mm; concrete grade C20/25
Base plate dimensions: 𝑎 = 350 mm; 𝑏 = 350 mm; 𝑡 = 20 mm; steel grade S420
Anchor bolts: 4 x M20, As = 245 mm2, anchor head diameter a = 60 mm, steel grade 8.8
Grout thickness: 30 mm
Due to complexity of the construction of interaction diagram, only basic examples are shown here:
Pure compressive stress
According to EN 1992-1-1 – 6.7 Partially loaded areas:
1st iteration:
3
Brno University of Technology Column base: Hollow section column
𝑏2 ∙ 𝑑2 750 ∙ 750
𝑘𝑗 = √ =√ = 2.143
𝑏1 ∙ 𝑑1 350 ∙ 350
𝛽𝑗 ∙ 𝑘𝑗 ∙ 𝑓𝑐𝑘 1 ∙ 2,143 ∙ 20
𝑓𝑗𝑑 = = = 28.571 MPa
𝛾𝑐 1,5
𝑓𝑦 420
𝑐 = 𝑡 ∙ √3∙𝑓 = 20 ∙ √3∙28.571∙1.0 = 44.3 mm
𝑗𝑑 ∙𝛾𝑀0
2nd iteration:
𝐴 562 500
𝛽𝑗 ∙ √𝐴𝑐1 ∙ 𝑓𝑐𝑘 1.0 ∙ √ ∙ 20 3 ∙ 𝑓𝑐𝑘
𝑐0 56 035
𝑓𝑗𝑑 = = = 40.309 ≤ = 40 MPa
𝛾𝑐 1.5 𝛾𝑐
𝑓𝑦 420
𝑐 =𝑡∙√ = 20 ∙ √ = 37.4 mm
3 ∙ 𝑓𝑗𝑑 ∙ 𝛾𝑀0 3 ∙ 40 ∙ 1.0
Normal tension
According to ETAG 001 – Annex C four proofs are required for tension loading:
1. Steel failure
2. Pull-out failure
3. Concrete cone failure
4. Splitting failure
1. Steel failure
The tension load is uniformly distributed over all anchor bolts, thus force acting on one bolt is
𝑁𝑐,𝐸𝑑
𝑁𝐸𝑑 =
𝑛𝑏
𝑛𝑏 … number of bolts
4
Brno University of Technology Column base: Hollow section column
𝑁𝑅𝑘,𝑠 196
𝑁𝐸𝑑 ≤ = = 130.667 kN
𝛾𝑀𝑠 1.25
Anchor bolts are close to each other, group of anchors have one common concrete cone.
𝑐𝑚𝑎𝑥 320
∙ ℎ𝑒𝑓 = ∙ 500 = 213.3 mm
′ 𝑐𝑐𝑟,𝑁 1.5 ∙ 500
ℎ𝑒𝑓 = max = 213.3 mm
𝑠𝑚𝑎𝑥 250
∙ ℎ𝑒𝑓 = ∙ 500 = 83.3 mm
{ 𝑠𝑐𝑟,𝑁 3 ∙ 500 }
′
′
ℎ𝑒𝑓 213.3
𝑠𝑐𝑟,𝑁 = ∙ 𝑠𝑐𝑟,𝑁 = ∙ 3 ∙ 500 = 640 mm
ℎ𝑒𝑓 500
′ ′
𝑐𝑐𝑟,𝑁 = 0.5 ∙ 𝑠𝑐𝑟,𝑁 = 0.5 ∙ 640 = 320 mm
0 ′ 1.5
𝑁𝑅𝑘,𝑐 = 𝑘1 ∙ √𝑓𝑐𝑘,𝑐𝑢𝑏𝑒 ∙ ℎ𝑒𝑓 = 7.2 ∙ √25 ∙ 213.31.5 = 112.170 kN
0
𝐴𝑐,𝑁 562 500
𝑁𝑅𝑘,𝑐 = 𝑁𝑅𝑘,𝑐 ∙ 0 ∙ 𝛹𝑠,𝑁 ∙ 𝛹𝑟𝑒,𝑁 ∙ 𝛹𝑒𝑐,𝑁 = 139.47 ∙ 409 600 ∙ 0.934 ∙ 1 ∙ 1 = 143.938 kN
𝐴𝑐,𝑁
𝑁𝑅𝑘,𝑐 186.224
𝑁𝑐,𝐸𝑑 ≤ = = 79.965 kN
𝛾𝑀𝑐 1.8
4. Splitting failure
Splitting failure will not occur; the edge distance is in all directions 𝑐 ≤ 𝑐𝑐𝑟,𝑠𝑝 and the concrete pad
height is ℎ ≥ 2 ∙ ℎ𝑒𝑓 .
5
Brno University of Technology Column base: Hollow section column
Interaction diagram
Further information about the construction of multilinear interaction diagram can be found in [4].
Points of multilinear interaction diagram of square hollow section are shown in Figure 2.
6
Brno University of Technology Column base: Hollow section column
50.000
40.000
30.000
20.000
0.000, 14.223
10.000
7
Brno University of Technology Column base: Hollow section column
Shell elements, special spring and contact elements are calibrated using characteristics according to
the component method.
Elastic-plastic stress-strain diagram is used for material of shell elements. Assessment is based on the
maximum strain given according to EN 1993-1-5 [5] by value of 5%.
Bolts are modelled using special spring elements and assessment is carried out according to standard
procedures described in EN 1993-1-8.
The concrete pad is modelled using Winkler subsoil model. The stiffness is uniform with the value:
k
Ec
1
4
Abp h
( 1 ) 3
Aref 2 d
The following values for coefficient were used: Aref =10 m2; α1 =1.65; α2 = 0.5; α3 =0.3; α4 =1.0.
The resistance of component concrete in compression including grout is determined using intersection
of two areas. The first area is the effective area for compression only (point 4) calculated according to
the Eurocode [1] using iterations until the difference between additional bearing widths |𝑐𝑖 − 𝑐𝑖−1 | is
less than 1 mm. Second area is where the concrete is in compression from finite element method
results. This second area is especially useful to determine the position of neutral axis. The intersection
of these two areas is the effective area for generally loaded column. This algorithm provides effective
area 𝐴𝑒𝑓𝑓 for any arbitrary loaded column. The design bearing strength of the joint 𝑓𝑗𝑑 is calculated
according to the Eurocodes [1, 4]:
𝐴𝑒𝑓𝑓
𝐹𝑅𝑑𝑢 = 𝐴𝑒𝑓𝑓 ∙ 𝑓𝑐𝑑 √ ≤ 3,0 ∙ 𝑓𝑐𝑑 ∙ 𝐴𝑒𝑓𝑓
𝐴𝑐,1
The average stress σ on the effective area 𝐴𝑒𝑓𝑓 is determined (whole compression force divided by
effective area) and the check of the component is then executed: 𝜎 ≤ 𝑓𝑗𝑑 .
The results from Idea Connection software for the same anchorage example are shown below.
8
Brno University of Technology Column base: Hollow section column
Figure 4: SHS 150x16 column base solid and meshed model in CBFEM
Point -1
9
Brno University of Technology Column base: Hollow section column
In Idea Connection software prying forces occure in pure tension loading (Figure 5) while in component
method no prying forces should be present. This causes the difference of 10 % in CBFEM and CM
resistances.
Point 0
Figure 7: Stress in concrete and effective area Aeff (hatched area) for point 0
Pure bending causes no prying forces anymore (Figure 6) and the difference between CBFEM and CM
bending moment resistance is less than 2 %. Concrete cone breakout is decisive in tension bolts
resistance and stress in concrete is low (Figure 7).
10
Brno University of Technology Column base: Hollow section column
Point 1
Figure 8: CBFEM results for point 1 (compressive force and bending moment)
Figure 9: Stress in concrete and effective area Aeff (hatched area) for point 1
CBFEM resistance for combined compressive force and bending moment is determined by keeping the
compressive force on the same value as in CM and modifying the bending moment to achieve 100 %
utilization in CBFEM. Bending moment resistance from CBFEM (Figure 8) is 12 % lower at −756 kN
compressive force than from CM. The difference may be explained by the different position of neutral
axis (Figure 9).
11
Brno University of Technology Column base: Hollow section column
Point 2
Figure 10: CBFEM results for point 2 (compressive force and bending moment)
Figure 11: Stress in concrete and effective area Aeff (hatched area) for point 2
Bending moment resistance from CBFEM (Figure 10) is 3 % lower at −913 kN compressive force than
from CM. The stress distribution under the base plate can be seen in Figure 11.
12
Brno University of Technology Column base: Hollow section column
Point 3
Figure 12: CBFEM results for point 3 (compressive force and bending moment)
Figure 13: Stress in concrete and effective area Aeff (hatched area) for point 3
Bending moment resistance from CBFEM (Figure 12) is 10 % higher at −1070 kN compressive force
than from CM. The stress distribution under the base plate can be seen in Figure 13. The tensile bolts
still govern the joint resistance, but concrete compressive resistance is almost used up.
13
Brno University of Technology Column base: Hollow section column
Point 4
Figure 15: Stress in concrete and effective area Aeff (hatched area) for point 4
The compressive resistance from CBFEM (Figure 14) is 9 % lower than from CM. The difference is
attributed to the smaller effective area Aeff caused mainly by the fillet of this area which was neglected
in case of CM.
14
Brno University of Technology Column base: Hollow section column
Sensitivity study
Results of CBFEM Idea RS software were compared with the results of the component method. The
comparison was focused on capacity and determination of the critical component.
The study was performed for parameters: size of the column, dimensions of the base plate, concrete
grade, dimensions of the concrete pad. The selected columns were SHS 150x16, SHS 160x12.5 and
SHS 200x16. The base plate width and length was chosen 100 mm, 150 mm and 200 mm larger than
the column section, the base plate thickness was 10 mm, 20 mm and 30 mm. The concrete pad was
from grade C20/25, C25/30, C30/37 and C35/45. The concrete pad height was for all cases 800 mm
and width and length was 100 mm, 200 mm, 300 mm and 500 mm larger than the dimensions of the
base plate. One parameter was changed and the others were held constant. The varying parameters
are summarized in tab. 1. Welds were the same around the whole column section. The fillet welds had
the throat thickness a = 8 mm. The coefficient lowering the compressive strength of concrete due to
grout βj = 1.0 was used, meaning high quality grout was used. All steel plates were grade S420 to avoid
yielding and bolts were M20 grade 8.8 in all cases.
Table 1: Selected parameters
100
80
An_SHS 150x16
MRd [kNm]
An_SHS 160x12,5
60
An_SHS 200x16
IC_SHS 150x16
40 IC_SHS 160x12,5
IC_SHS 200x16
20
0
-3000 -2500 -2000 -1500 -1000 -500 0 500
NRd [kN]
15
Brno University of Technology Column base: Hollow section column
160
140
120
An_100
100
MRd [kNm]
An_150
80 An_200
IC_100
60
IC_150
40 IC_200
20
0
-3500 -3000 -2500 -2000 -1500 -1000 -500 0 500
NRd [kN]
Figure 17: Interaction diagrams for varying parameter of base plate offset
180
160
140
An_10
120
MRd [kNm]
An_20
100 An_30
80 IC_10
IC_20
60
IC_30
40
20
0
-4500 -4000 -3500 -3000 -2500 -2000 -1500 -1000 -500 0 500
NRd [kN]
Figure 18: Interaction diagrams for varying parameter of base plate thickness
16
Brno University of Technology Column base: Hollow section column
Concrete grade
The column section SHS 150x16, base plate thickness 20 mm, base plate offset 100 mm and concrete
pad offset 200 mm was used for varying parameter of base plate offset. Concrete grade C20/25,
C30/37 and C35/45 were used in this study. The comparison of interaction diagrams is in Figure 19.
100
90
80
70
An_C20/25
60
MRd [kNm]
An_C30/37
50 An_C35/45
40 IC_C20/25
IC_C30/37
30
IC_C35/45
20
10
0
-3000 -2500 -2000 -1500 -1000 -500 0 500
NRd [kN]
100
80
An_100
MRd [kNm]
An_300
60 An_500
IC_100
40 IC_300
IC_500
20
0
-2500 -2000 -1500 -1000 -500 0 500
NRd [kN]
Figure 20: Interaction diagrams for varying parameter of concrete pad offset
17
Brno University of Technology Column base: Hollow section column
Reliability
Reliability of CBFEM software is provided in accordance with the strategy of EC considering partial
safety factors.
Material safety factors according to EN 1993-1-8 [1] and ETAG 001 [3] were used for design resistance
of the connection. For concrete γc = 1.5, for installation safety γ2 = 1.2, for welds M2 = 1.25 and for
plates M0 = 1.0.
In Figure 21, Figure 22 and Figure 23 the reliability graphs are plotted. Only the pure tension, pure
bending and pure compression was selected for reliability graphs because the comparison is simple
and definite. The red and blue dashed lines correspond to the 110 % and 90 % value of resistance from
component method, respectively. The black dashed line corresponds to the complete compliance of
the results from analytical solution using component method and from Idea Connection software. The
points below the blue line have difference higher than 10 % but are on the safe side. The worst result
on the unsafe side was by 4 % higher than according to the component method.
The comparison of combined compressive force and bending moment is more problematic and can be
seen in Figures 16 – 20. Interaction diagram comparison of CBFEM and CM is in Table 2. Values are
determined by tensile resistance from CBFEM divided by tensile resistance from CM in case of point
−1, bending resistance from CBFEM divided by bending resistance from CM in case of points 0, 1, 2 and
3 and compressive resistance from CBFEM divided by compressive resistance from CM in case of point
4. Small differences were caused by the more accurate evaluation of the design bearing strength of
the joint 𝑓𝑗𝑑 and effective area 𝐴𝑒𝑓𝑓 in the software, which uses more iterations and does not neglect
the fillet. In some cases significant differences caused occurrence of prying forces in the CBFEM results
which increased the tension loading of anchor bolts.
120
100
80
CBFEM [kN]
60
40
20
0
0 20 40 60 80 100 120
Component Method [kN]
18
Brno University of Technology Column base: Hollow section column
50
40
30
CBFEM [kNm]
20
10
0
0 10 20 30 40 50
Component Method [kNm]
-4000
-3500
-3000
-2500
-2000
CBFEM [kN]
-1500
-1000
-500
0
0 -500 -1000 -1500 -2000 -2500 -3000 -3500 -4000
Component Method [kN]
19
Brno University of Technology Column base: Hollow section column
-1 0 1 2 3 4
maximum 99,3% 101,6% 101,8% 109,5% 120,9% 104,4%
minimum 70,2% 69,4% 79,8% 85,6% 90,9% 89,3%
average 89,3% 93,6% 87,1% 97,0% 108,1% 94,7%
Recapitulation
Verification studies confirmed the accuracy of the CBFEM IDEA RS software. Results of this software
were compared with the results of the component method recommended in EN 1993-1-8. Most of the
results differences were below 10 %, which is a generally acceptable value. Differences were caused
by the more accurate evaluation of the design bearing strength of the joint 𝑓𝑗𝑑 and effective area 𝐴𝑒𝑓𝑓
in the software, which uses more iterations and does not neglect the fillet. In some cases significant
differences caused occurrence of prying forces in the CBFEM results which increased the tension
loading of anchor bolts.
References
[1] EN 1993-1-8. Eurocode 3: Design of steel structures – Part 1-8: Design of joints.
[2] EN 1992-1-1. Eurocode 2: Design of concrete structures – Part 1-1: General rules and rules for
buildings.
[3] ETAG 001: Guideline for European Technical Approval of Metal Anchors for Use in Concrete –
Annex C: Design Methods for Anchorages. Brussels: EOTA, 2010.
[4] Wald, F. Column Bases. Prague: ČVUT, 1995.
[5] EN 1993-1-5. Eurocode 3: Design of steel structures – Part 1-5: Plated structural elements.
20