Accelerat ing t he world's research.

IRJET-PROBABILITY FAILURE OF
COLUMN IN STEEL STRUCTURE
IRJET Journal

Related papers Download a PDF Pack of t he best relat ed papers 

IRJET-A ST UDY ON PROBABILIT Y FAILURE OF A COLUMN IN RC FRAME BY VARYING LIVE LOAD

IRJET Journal

IRJET-RELIABILIT Y ANALYSIS OF STAT IC RESPONSE OF ST RUCT URES WIT H IRREGULARIT IES

IRJET Journal
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072

PROBABILITY FAILURE OF COLUMN IN STEEL STRUCTURE

Manjunath G.S1, Dr.K Manjunath2, Sandeep Kumar D.S3
Student, Civil Engineering, PESCE, Karnataka, India
1

Professor, Civil Engineering, MCE, Karnataka, India

2
3Assistant Professor, Civil Engineering, PESCE, Karnataka, India

---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The evaluation of the safety of structures is a to loadings. Design methods may be classified in the
task of much importance. It has been one of the subjects of following ways
a) Permissible Stress Method
interest for engineers. The safety of a structure depends
b) Ultimate Strength Design and Plastic Design Method
on the resistance R, of the structure and the action S (load c) Limit State Method
or load effect) on the structure. The action is a function of d) Probability-based Limit States Design
loads (live load, wind load, etc.), which are random
variables. Similarly, the resistance or response of the 1.1 Probability-based Limit State Design
structure depends on the physical properties of materials, In this design method, probabilistic methods are used to
and the geometric properties of the structure which were guide the selection of the partial safety factors to loads and
resistances of the structure or structural elements or
also subjected to statistical variations, and are
material of the structures, and they result in the desired
probabilistic. overall safety. The principal advantages of this design
method are:
The design variables being random, it becomes much a) More consistent reliability is attained for different
more important to assess the level of safety in the situations as the different variabilities of the various
probabilistic design situation. Column being the vital resistances and loads are considered explicitly and
most structural element, probability of failure of a column independently.
is linked to the overall safety of a structural system. With b) The reliability level can be chosen to reflect the
this in view, an attempt is made to assess the safety consequences of failure.
c) It is a tool for exercising judgement in non routine
levels ensured by the design methodology of the present
situations.
code of practice IS 800-2007. This requires the d) It provides a tool for updating standards in a
information on the probability distributions of the rational manner.
resistance part and action part in each limit state. The The conceptual framework for the analysis of structural
work is carried out on the following lines: reliability and probability-based design is provided by the
classical reliability theory.
In the present study level 2 reliability of the design of steel
compression member are carried out as per IS 800-2007. 1.2 Variables in engineering problems
The variables in engineering problems can be classified as
The statistics and probability distribution of axial load, a) Deterministic: An approach based on the premise
uniaxial moments and biaxial moments in a particular that a given problem can be stated in the form of a
question or a set of questions to which there is an
column are generated by digital simulation on a selected
explicit and unique answer is a deterministic
steel frame using ETABS 2013. The generated data on approach. Example: material characteristics.
axial load, axial load with uniaxial moments and axial b) Probabilistic: A probabilistic approach is based on
load with biaxial moments is subjected to statistical the concept that several or varied outcomes of a
analysis. situation are possible to this approach. Uncertainty
recognized as yes or no type of answer to a question
Key Words: Reliability Analysis, Probability failure, concerning structural performance is considered to
be simplistic. Probabilistic modeling aims at a study
Monte Carlo simulation, Steel structure. of a range of outcomes to a given input data.
Accordingly the description of a physical situation
or system includes randomness of data and other
1. INTRODUCTION uncertainties. The selected data for a deterministic
General principles for checking safety define a method for approach would not be sufficient for a probabilistic
calculating the behavior and strength of structures subjected

© 2015, IRJET.NET- All Rights Reserved Page 1971

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072

study of the same problem. The raw data would b) Factored moment along Major axis
provide a meaningful probabilistic analysis. Mz(top) = 20.892 kN-m
Mz (bottom) = 8.836 kN-m
A probabilistic approach aims in determining the probability c) Factored moment along Minor axis
p, of an outcome, one of that may occur. Probability would be My(top) = 12.267 kNm
any percentage between p=0% and probability=100% or any My (bottom) = 6.075 kNm
value between p=0 and p=1. In a specific problem the
number of likely outcomes may be limited and it may be The section is found to be safe under various checks in the
possible to consider the probability of each outcome. design as per the provision of IS: 800-2007.
Example: Wind load, ocean-wave height, earthquake etc.
3. GENERATION OF LOAD STATISTICS AND
2. DETERMINISTIC DESIGN OF COLUMN RESISTANCE STATISTICS
For deterministic design a typical model is considered. A For probabilistic design, standard normal deviates are
four storey steel frame having, columns of height 3.5m are obtained by generating two uniform random numbers v1
spaced at 5m centre to centre, beams of length 5m, wall load and v2 (with a uniform density range between 0 and 1).
of 20kN/m, slab of 125mm thick of M25 grade concrete, with Then the standard normal variates are u1 and u2 are
live load on slab of 3kN/m2 are used for modeling. calculated.
Here, in present study the variations is applied for fck
characteristic cube strength of concrete in slab, depth of slab,
live load on slab, length of beam, live load on beam and
height of column.
For 100 values of standard normal variates, the above
parameters are applied and number of times frame analysis
is carried out in ETABS and the results are used to find the
probability of failure of columns of the structure.

3.1 Random variables

The random variable is a numerical variable whose specific
values cannot be predicted with certainty before an
experiment. The value assumed by a random variable
Fig -1: Plan view of steel structure. associated with an experiment depends on the outcome of
the experiment. This value is associated with every simple
event defined on the sample space, but different simple
event defined on the sample space, but different simple
events may have the same associated value of the random
variables, e.g. the strength of concrete, the wind speed
observed at a location etc.

3.2 Monte Carlo simulation

Monte Carlo simulation is a type of simulation that relies on
repeated random sampling and statistical analysis to
compute the results. This method of simulation is very
closely related to random experiments, experiments for
which the specific result is not known in advance. In this
Fig -2: 3D view of steel structure. context, Monte Carlo simulation can be considered as a
methodical way of doing so-called what-if analysis.
Frame analysis is carried out and results are used for In Monte Carlo simulation, we identify a statistical
deterministic design. Axial load and moments along minor distribution which we can use as the source for each of the
and major axis are used to check the safety of column. input parameters. Then, we draw random samples from each
In the present study standard Rolled I-sections ISHB 450 distribution, which then represent the values of the input
(87.2 kg/m) is taken as trial sections. The axial load and variables. For each set of input parameters, we get a set of
moments for the design are taken from the frame analysis output parameters. The value of each output parameter is
carried out in ETABS. one particular outcome scenario in the simulation run. We
collect such output values from a number of simulation runs.
Data obtained from analysis: Finally, we perform statistical analysis on the values of the
a) Factored axial load = 536.718 kN output parameters, to make decisions about the course of
action (whatever it may be). We can use the sampling
© 2015, IRJET.NET- All Rights Reserved Page 1972
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072

statistics of the output parameters to characterize the output

variation. In the present study due the variations applied the samples
of the column is found to be unsafe which is decided by the
3.3 Variations in resistance parameters interaction value obtained due to combined effect of axial
The fundamental requirement in the reliability study is the load and moments as per the provision of IS: 800-2007.
collection of data on the strength and other physical
properties of the materials of the structures, and the 5. PROBABILITY MODELING
geometric parameters of the sections and the statistical Histogram provides an immediate impression of the range of
analysis of the same. In present study the following the data, its most frequently occurring values and the degree
variations are applied for various parameters of resistance to which it is scattered. It is the presentation of data in useful
which include form. The observations are made and noted down as they
a) Geometrical properties occur and hence the collected data will be in an unorganized
b) Material properties form. This unorganized data is arranged properly. The values
c) Loads are marked in an increasing order. These ordered values are
then divided into intervals and the number of observations
4. PROBABILISTIC DESIGN OF COLUMN in each interval is plotted as a bar.
Limit state of combined axial load and biaxial moments
a) For random variation in different geometrical Histogram and probability distribution curve of the
properties, length and load, corresponding generated sample of Q and maximum moment capacity Mu3
interaction values are calculated with respect to and maximum moment capacity Mu2 about major and minor
both major and minor axis. axis for column subjected to combined axial load and bi-axial
b) Member subjected to combined axial compression moments are plotted.
and biaxial bending shall satisfy the following
relationship The suitability of a probabilistic model to fit the data is
P/Pdy+Ky (Cmy My)/Mdy+ KLT Mz/Mdz ≤ . arrived after applying any one of the following goodness-of-
P/Pdz+0.6Ky (Cmy My)/Mdy+ Kz (Cmz Mz)/Mdz ≤ . fit test in MATLAB.
c) Let iy be the interaction value corresponding to a) Chi-square Test.
Minor axis. b) Kolmogorove-Smirnov (K S) Test.
d) Let iz be the interaction value corresponding to It is found that Normal distribution fits the generated data
Major axis. well, Based on the Chi-square Test. During the reliability
e) If iy > 1.0 or if iz > 1.0, then it is failure. analysis of the present design, the resistance model is
f) Compute the probability of failure by calculated by performing Monte-Carlo simulations.The
g) pf = (Number of such failure)/(Number of variables included in the study are dimensions, material
simulations). properties and load. Loads were obtained from ETABS
model. The current study assumes a normal distribution for
Table -1:Safety check of the column samples all variables.

My Interaction
My Mz Mz
P (kN) (Top) values
(Bot) (Top) (Bot)
Y-Y Z-Z
1091.1 27.03 13.40 45.43 19.63 1.18 1.01
1050.4 25.95 12.86 43.61 18.82 1.12 0.96
1106.8 27.47 13.62 46.16 19.96 1.20 1.02
956.52 23.35 11.58 39.31 16.91 1.00 0.85
1052.2 25.99 12.89 43.68 18.85 1.13 0.96
1089.3 26.99 13.38 45.36 19.60 1.17 1.00
979.17 23.94 11.87 40.29 17.35 1.03 0.88 Fig -4: Histogram for Axial Load P of the column

1106.8 27.47 13.62 46.20 20.00 1.19 1.01

920.36 22.33 11.07 37.62 16.16 1.09 0.95
1058.6 26.15 12.96 43.94 18.96 1.13 0.96

© 2015, IRJET.NET- All Rights Reserved Page 1973

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072

variables are supplied as input to the program.

Similarly, the statistics of resistance are generated
using the same technique.
d) The safety margin is defined as the difference
between resistance and action. When the random
values of resistance (R) and action (S) are
generated, check is made if R < S. If so, it is counted
as failure. The probability of failure is the ratio of
(Number of such failure)/(Number of simulations).
e) The probability of failure is found to be 10%,
reliability is 90%.
f) It is possible to evaluate the explicit level of safety
of steel column design by using the specification of
IS 800:2007. The overall factor of safety is implicitly
Fig -5: Normal distribution curve for Axial Load P of the built into deterministic design but it is not explicitly
column known to the designer. The reliability analysis helps
to obtain the explicit level of safety by using a given
design philosophy.
g) The present work attempts to demonstrate the
procedure for evaluating safety level in terms of
probability failure by using the methods of
reliability analysis.
h) The advantage of reliability analysis is that it helps
to formulate the basis for reliability based design
and this will help the designer to design the
structural elements for a known probability of
failure.

REFERENCES

Fig -6: Log Normal distribution curve for Axial Load P of the [1] R.Ranganathan, Reliability Analysis and Design of
column Structures, Tata Mc Grawhill Publishing Company
limited.
Similarly, Histrogram, Normal and Log normal distribution
curve is plotted for moments at top and bottom of the [2] S.Ramamrutham, Design of Steel Structures, Dhanpat
column with respect to both major and minor axis using Rai Publishing Company limited.
MATLABS.
[3] Ravi G, , Reliability studies of Steel Structures in
Flexure, Tension and Compression , M.Tech Thesis,
3. CONCLUSIONS
Indian Institute of Science, Bangalore.
a) The deterministic design methodology suggested by
IS 800:2007 by adopting limit state method with
[4] Yadunandan C.N, Sundararju Iyengar K.T,(2000),
appropriate partial safety factors is assessed for
Parial Load Safety Factor for Strength Design of Steel
explicit level of safety by using methods of
Structures , )E ) Journal-CV, Vol.81, pp 33-36.
Reliability. It is done for three different limit of
states namely
[5] Andrzej s. Nowak, Kevin R. Collins, , Reliability
Limit state of Flexure
of Striuctures , McGraw (ill Publishers, United States of
Limit state of Shear
America.
Limit state of Deflection.
b) Monte Carlo digital simulation is used to generate
[6] CIRIA – Construction Industry Research and
random variables by using the statistics of design
Information Association (1977), Report 63,
variables. The statistics of design variables are
Rationalisation of Safety and Serviceability factors in
taken from literature. MATLAB is used to conduct
structural codes , London.
the digital simulation.
c) The statistics of action are generated by repeatedly
[7] )drish A., Edache P, , Reliability Analysis of
analysing a multi storeyed steel frame using ETABS
Simply Supported Steel Beams , Australian Journal of
software. The randomly generated values of design

© 2015, IRJET.NET- All Rights Reserved Page 1974

 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 02 Issue: 03 | June-2015 www.irjet.net p-ISSN: 2395-0072

Basic and Applied Science 1(1), INSInet publications pp BIOGRAPHIES

20-29.

[8] N.Subramanian, Design of Steel Structures, Oxford Manjunath G.S

University press. Pursuing M.Tech, Civil Engineering,
PESCE Mandya
[9] Samik Raychaudhuri, INTRODUCTION TO MONTE
CARLO SIMULATION, Indian Institute of Technology,
Kharagpur, India.

[10] Dr. Vasant Matsagar, Design of Steel Structures, Indian

Institute of Technology Delhi Dr. K. Manjunath
Professor and Head, Dept. of Civil
[11] V.N.Vazirani, M.M.Ratwani, Vineet kumar, Design and Engineering, MCE, Hassan
Analysis of Steel Structures, Khanna Publishers.

[12] IS: 800- , General construction in steel - Code of

practise , Bureau of )ndian Standards, New Delhi.

[13] SP: 6(1) – 4, (andbook for Structural Engineers ,

Bureau of Indian Standards, New Delhi.

© 2015, IRJET.NET- All Rights Reserved Page 1975