1
FreeVaP – Examples
The following examples are intended to make the user of FreeVaP step-by-step familiar with
the different possibilities offered by FreeVaP.
The respective FreeVaP files may be found in the folder Examples.
Exa_1 – Browsing through different types of variables
Have a look to FreeVaP’s available types of variables, play around and get acquainted to
input and output of VaP:
• Open FreeVaP
• Open New from the File Menu or Open Exa_1 from the Examples folder
• Define the simplest possible G-Funktion, e.g., G = A.
• Define the variable A. The suggestion is to start with Normal(1.0;0.1)
• Change variable types and look to the changing shapes of the pdf. Be aware of the need to
add lower and upper bounds to some of the variable types. Change also the Mode, e.g., to
Moments or to Parameters.
• Choose Method MC and see Monte Carlo live
• Change to 100.000 runs, change from “Automatic” to “Define Values”, increase Classes to,
e.g., 30, and define range. Then, again watch MC live
• Change variable type and use Method MC again
• Press Report LSF and Save … or Save as … to the Examples folder
• Press Print from the File Menu, if you so desire.
Play around with the G-Function:
• Add a definition of the G-Function to G1 = A1+A2. Define both variables A under
Parameters Mode as Rectangular(0;2).
• Run MC and see the result, triangular variable: see the Central Limit Theorem SED 5
Chapter 2.62.
• Press Report LSF and Save … or Save as … to the Examples folder
• Add another G-Function as the product of A: G2 = A3*A4.
• Define both variables A3 and A4 as Normal(1;0.6)
• Run MC and see the result: again the Central Limit Theorem as G tends to look as
lognormal
• Press Report LSF and Save … or Save as … to the Examples folder
• Press Print from the File Menu, if you so desire
Exa_2 – Your way back home from your office
Just as a stupid story behind this example: Assume that you are interested in the probability of
being back home from your office at some given time, or, in other words, that your way back
lasts longer than, e.g., T=40 minutes. The way back is on two independent stretches on metro
trains and 10 minutes walk altogether. Metros run in intervals of Int=12 minutes, but there is
neither a timetable nor a coordination of train schedules. As you will wait 1, or 2 … or 10 mi-
nutes at each of the stations, the pdf of Int is rectangular. The train travels last 8 and 15 minu-
tes, resp.
2
There are a number of answers ready almost without thinking: Your way home lasts at least
33 minutes and maximal 57 minutes, and a mean travel time seems to be 45 minutes. FreeVaP
may give a more complete answer:
• Open New from the File Menu or Open Exa_2 from the Examples folder
• Define the G-Funktion, e.g., G = T– (10+8+15+Int1+Int2).
• Define variable T as Deterministic(40) and both Int1 and Int2 as Rectangular(0;12)
• Run MC and see the pdf of travel time and the probability of travel time being less than T.
• Press Report LSF and Save … or Save as … to the Examples folder
• Change T to adapt p(g<T) to your liking.
An example of the results is that there is a 17% chance that your trip home lasts longer than
50 minutes. So long? It’s a pity.
Exa_3 – Use VaP with two variables
Have a first look to two variables problems as encountered in analysing any reliability
problem.
• Open New from the File Menu or Open Exa_3 from the Examples folder
• Define the Limit State Function as G = R – S.
• A simple example, a running contest. You know your speed quite well and set it to R =
Normal(100;10). You do not know the speed of your competitor. You think he is slower,
but you do not really know. Your guess is S = Normal(80;20).
• Choose FORM from the Menu Methods. The result is the probability pf = 0.186 that he
will win the contest though you are quite sure (0.814) that you will win.
• Think about other examples for G = R – S. Have a look to SED 5 Chapter 3.1.
• Change definitions of R and S accordingly. Examine all of the Methods provided.
• Compare the results.
• Press Report LSF and Save … or Save as … to the Examples folder
• Press Print from the File Menue, if you so desire
Exa_4 – The risk to loose or the chance to win money in offering a lump sum
A contractor in view of erecting a big warehouse is asked to provide a lump sum contract for
the whole undertaking. The cost of different pieces of the work to be done are known from
experience, but – of course – are stochastic, e.g., not known beforehand. He is testing out risk
and chances with VaP.
In order to concentrate on the task of this example we simplify the problem. The estimates of
the contractor are as follows:
• M Material Normal(40;6)
• L Labour Normal(30;3)
• D Design and detailing Normal(15;2)
• A Administration and Management Normal(10;1)
• RP Risk&Profit Normal(5;1)
• LS Lump sum to offer Deterministic(105)
The procedure now is:
• Open New from the File Menu or Open Exa_4 from the Examples folder
• Define G = LS–(M+L+D+A+RP) and the above mentioned variables
• Press FORM from the Menu Methods and learn, that the probability that you loose money
is pf = 0.242. If you do not like this result, change LS to a higher value.
3
• Press Report LSF and Save … or Save as … to the Examples folder
Exa_5 – Reliability assessment of a column
The following is drawn from an article published in SEI in 1992 (see Schneider, J., 1992).
• Open New from the File Menu or Open Exa_5 from the Examples folder
• Define G = M*0.25*π*D^2*C*T-N*(DL+LL)
and set the parameters of the variables as follows:
• M Model uncertainty variable of resistance Normal(0.85;0.06)
• D Diameter of column drawn from measurements Normal(0.96;0.005)
• C relevant concrete strength drawn from concrete cores Lognormal(46;6)
• T Transfer variable lognormal(1.0;0.05)
• N Model uncertainty variable of loads Normal(1.0;0.07)
• DL normal force from dead load on column Normal(9.45;0.66)
• LL normal force from live load on column Gumbel(1.6;0.48)
• Press Report LSF and Save … or Save as …
• Run Method FORM
• Look at the α-values indicating the importance of the variables on the result
• Try all other methods
• Compare results
• Press Report LSF and Save … or Save as …
Exa_6 – Failure probability of a frame
The following example is drawn from Vrouwenvelder/Schneider, 2017, Chapter 4.75.
• Open New from the File Menu or Open Exa_6 from the Examples folder
• Define
• G1 = 2*MC+2*MB–4*F
• G2 = 4*MC–4*H
• G3 = 4*MC+2*MB–5*H–5*F
• Set the parameters as follows:
• MC Plastic moment of the columns LogNormal(135;13.5)
• MB Plastic moment of beam LogNormal(270;27)
• F Vertical force LogNormal(50;15)
• H Horizontal force LogNormal(40;12)
• Press Report LSF and Save … or Save as …
• Run Method FORM for all three G-functions using the Menu “Activate”
• Look at and note the pf values
• Define the bounds of the failure probability of the frame
Exa_7– An educational RC beam
The following can be seen as an explanation to what is treated from Chapter 3.21 of
Schneider and Vrouwenvelder, 2017. It looks complicated at first sight:
• Open New from the File Menu or Open Exa_7 from the Examples folder
• Define the following quite long G-Function: G = X1*X2*X3*10^-6*(X4-X5-
0.5*X2*X3/(X7*X6)) – (X9*(X10+X11)*X8^2*0.125+X12*X8*0.25)
save it, because typing the thing again would be cumbersome.
4
Now, set the variables as follows:
• X1 N(1;0.03)
• X2 N(2'500;50)
• X3 NL(500;15)
• X4 N(800;8)
• X5 NL(50;10)
• X6 NL(40;4)
• X7 N(1200;12)
• X8 D(9;0)
• X9 N(1;0.1)
• X10 NL(4;0.2)
• X11 GuL(25;5)
• X12 N(100;10)
• Run Method FORM and have a look to βHL= 3.68, and pf = 0.115*10-3 and the α-values.
You will see from these that only X1, X3, X5, X9, X11 and X10 really matter. Save the
thing in order to not loose the variables definitions.
• Keep the definition of these six variables and set all the other variables to Deterministic.
Run FORM again and see the result: βHL= 3.71, and pf = 0.104*10-3, hence practically the
same. When defining the G-function, in many cases it is good enough to look to the most
important variables only.
• Have a look to the α-values again. You will see from these that only X3, X9, and X11
really matter. Again, set all the other variables to Deterministic and run FORM again and
see the result: βHL= 3.78, and pf = 0.072*10-3, hence practically the same.
• Define a new function in skipping the second part of the G-function and name it R. Run
Monte Carlo and have a look to the shape of the histogram. Be aware: you need to set the
options to Define values and set the boundaries.
Run Method Moments and note the mean and the standard deviation of R.
• Do the same with the the second part of the G-function and name it S. Again, note the
parameters of S, e.g. mean and the standard deviation.
Literature
Petschacher, M.: ”Zuverlässigkeit technischer Systeme - Computerunterstützte Verarbeitung
von stochastischen Grössen mit dem Programm VaP”; IBK Bericht Nr. 199, ETH-Zürich,
1993.
Petschacher, M.: ”VaP a tool for practicing engineers”; Proceedings ICOSSAR ‘93: 1817-
1823, 1994.
Schneider, J.: “Some thoughts on the reliability assessment of existing structures”; IABSE,
SEI, 1992
Vrouwenvelder, T. and Schneider, J.: “Introduction to Safety and Reliability in Structural
Engineering”, SED 5, IABSE, 2017.