1

I. INTRODUCTION

Abstract of the Journal

The paper presents an experimental program investigating the behavior of high strength

steel connections consisting of one bolt in double shear. A total of 24 bolted connections fabricated

from three grades of high strength steel with the nominal yield strengths of 550 MPa, 690 MPa

and 890 MPa were tested. The effects of end distance, edge distance and steel grade on the bolt

bearing behavior were evaluated. The bolt hole elongation due to bolt bearing on high strength

steel plate was measured and its implication on the plate bearing resistance was discussed. The test

results were compared with Eurocode 3 and AISC 360-10 predictions and it was found that

Eurocode 3 could be used conservatively to predict the bolt bearing resistance on high strength

steel with nominal yield strength up to 890 MPa whereas AISC 360-10 method tends to

overestimate the bearing resistance of the bolted connection.

A regression analysis was performed based on the test data and those from the literature so

that a more general method was proposed to predict the bolt bearing resistance on normal strength

and high strength steel plate. Splitting failure was observed as a transitional failure mode between

tear-out failure and net cross-section failure. Splitting failure showed a lower resistance than the

bolted connection with tear-out failure, therefore, a reduction factor was proposed to improve the

prediction. The upper and lower boundaries of end distance to edge distance ratio for splitting

failure were theoretically derived and experimentally verified.

Regression

In statistical modeling, regression analysis is a set of statistical processes for estimating the

relationships among variables. It includes many techniques for modeling and analyzing several
 2

variables, when the focus is on the relationship between a dependent variable and one or more

independent variables. More specifically, regression analysis helps one understand how the typical

value of the dependent variable changes when any one of the independent variables is varied, while

the other independent variables are held fixed.

Most commonly, regression analysis estimates the conditional expectation of the

dependent variable given the independent variables – that is, the average value of the dependent

variable when the independent variables are fixed. Less commonly, the focus is on a quantile, or

another location parameter of the conditional distribution of the dependent variable given the

independent variables. In all cases, a function of the independent variables called the regression

function is to be estimated. In regression analysis, it is also of interest to characterize the variation

of the dependent variable around the prediction of the regression function using a probability

distribution.

Many techniques for carrying out regression analysis have been developed. Familiar

methods such as linear regression and ordinary least squares regression are parametric, in that the

regression function is defined in terms of a finite number of unknown parameters that are estimated

from the data. Nonparametric regression refers to techniques that allow the regression function to

lie in a specified set of functions, which may be infinite-dimensional.

R-squared is a statistical measure of how close the data are to the fitted regression line. It

is also known as the coefficient of determination, or the coefficient of multiple determination for

multiple regression. The definition of R-squared is fairly straight-forward; it is the percentage of

the response variable variation that is explained by a linear model. R-squared is always between 0

and 100%. 0% indicates that the model explains none of the variability of the response data around

its mean. 100% indicates that the model explains all the variability of the response data around its
 3

mean. In general, the higher the R-squared, the better the model fits your data. However, there are

important conditions for this guideline that I’ll talk about both in this post and my next post.

II. DATA GATHERING

Test Set up

The tests were carried out on a hydraulic servo-controlled machine with the loading

capacity of 1000kN. Grade 12.9 M24 bolts (without washers) were used to connect the test

specimens to the support with enough shear resistance to avoid bolt shear failure, as shown in Fig.

1. No pretension of the bolt was applied so that load was transferred primarily by bearing not by

friction. A 10 kN load was applied and unloaded before actual loading to make bolt shank bear on

hole wall. Then, the specimens are loaded at a prescribed displacement rate of 1.5 mm/min until

failure of specimens. Two linearly variable displacement transducers (LVDTs) were positioned

along both edges of the specimen to measure the bolt hole elongation and plate deformations in

the force direction, as shown in Fig. 1. The applied load was recorded by the built-in sensor of the

1000 kN machine. A strain gauge was attached at the end surface of each specimen to measure the

tensile strain perpendicular to load direction, as shown in Fig. 1.

Fig 1. Test set up and arrangement of measurements.

 4

Geometric parameters

Single-bolt connections in double shear were investigated in the tests. Eight specimens for

each HSS and in a total of 24 specimens of HSS with various end distance and edge distance were

prepared. Wire electrode cutting was adopted to reduce the effect of heat on material properties.

The specimens are named in terms of SD-𝑒1 /𝑑0 − 𝑒2/𝑑0 -steel grade, where SD represents single

bolt in double shear and 𝑒1 , 𝑒2 and 𝑑2 are defined in Fig. 2. The end distance 𝑒1 is varied from

1.0𝑑0 to 2.5𝑑0 . The edge distance 𝑒2 of 0.8𝑑0 , 1.0𝑑0 , 1.5𝑑0 and 3.0𝑑0 are evaluated in the test.

Nominal dimensions and measured dimensions are summarized in Table 1, where 𝑡 is the plate

thickness and 𝐴𝑛𝑒𝑡 is the area of the net section.

Fig 2. Dimension Test of the Specimen

Table 1
Geometrical dimensions of test specimens
 5

III. CRITIQUE ON THE DATA ANALYSIS

Results

Fig. 3 presents the comparison between the test results with tearout failure and the design

strength predicted by Eurocode 3 and AISC 360-10. The normalized experimental bearing

resistance was found proportional to the end distance 𝑒1, regardless of steel grades.

Fig 3. Comparison between test results and design codes

To find a more accurate expression to describe the relationship between bearing resistance

and end distance, experimental results from different researchers and the background document

of Eurocode 3 were summarized in addition to the test results presented in this paper. A total of

191 samples were taken into consideration in order to form a linear relationship between the

dependent and independent variables. The linear relationship was based on normalized data of

the variables used in the experiment. By linear fitting based on least square method, ultimate

bearing resistance can be expressed in forms of Eq. (1) with R2 =0.99.

𝑒
𝐹𝑏,𝑢 = (1.039 𝑑1 + 0.183) 𝑓𝑢 𝑑𝑡 (1)
𝑜
 6

Fig 4. Comparison between 191 test results and prediction formulae.

Table 2
Sample of the data used for the prediction formulae

Critique of the data analysis

The researchers took into consideration over 191 samples. These 191 samples produced a

linear relationship between the two variables needed to predict the bearing strength of the plate.

The general form of the linear regression is known to be 𝒇(𝒚) = 𝜷𝟎 + 𝜷𝟏 𝒙, where 𝜷𝟎 is the y –

intercept and 𝜷𝟏 is the slope of the line. The value of 𝜷𝟏 in the relationship of the two variables is

equivalent to 1.039. A positive value of 𝜷𝟏 means the value 𝒇(𝒚) increases as the value of x
 7

increases, which reflects on Fig. 4. The value of 𝜷𝟎 is close to zero that is why line started near

the origin of the graph, this is because the y-intercept of the equation of the line is equivalent to

0183. The value of R2 = 0.99 which indicates that the model explains all the variability of the

response data around its mean.

IV. RECOMMENDATIONS

There are no recommendation for this paper, because all the necessary data were correctly

used and were interpreted in way that the readers of this paper would understand all the work done

by the authors.