FSTAR
FSTAR
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Analysis
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 Cumulative Damage Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 General Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 Elastic-Plastic Formulation Definitions . . . . . . . . . . . . . . . . . . . . . . . . 2-4 Analysis Procedure for Cyclic Loading . . . . . . . . . . . . . . . . . . . . . . . . . . Normal Procedure (at a Location) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . All-Nodes Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simplified Elastic-Plastic Formulation . . . . . . . . . . . . . . . . . . . . . . . . . 2-5 2-5 2-6 2-7
Description of Elements
. . . . . . . . . . . . . . . . . . . . . . . . 3-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
In de x
Contents
. . . . . . . . . . . . . . . . . . .4-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-1 Commands Likely to be Used for a Given Analysis . . . . . . . . . . . . . . . . .4-1 Analysis Menu (Analysis > FATIGUE) . . . . . . . . . . . . . . . . . . . . . . . .4-2 Analysis Menu (Analysis > FATIGUE > FATIGUE LIST) . . . . . . . . .4-2 Analysis Menu (Analysis > FATIGUE > FATIGUE DELETION) . . .4-3 Results Menu (Results > PLOT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-3
. . . . . . . . . . . . . . . . . .5-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-1 How FSTAR Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-1 Cylinder Under Axial Cyclic Loading Example . . . . . . . . . . . . . . . . . . . .5-4 History Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-4 Starting The Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-4 Specifying the Fatigue Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-5 Defining the Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-5 Defining a Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-6 Specifying the Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-7 Specifying the Fatigue Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-8 Running the Fatigue Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-10 Interpretation of the Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-11 Fatigue Caused by Pressure Loading Example . . . . . . . . . . . . . . . . . . . .5-12 Structural Modeling and Stress Analysis . . . . . . . . . . . . . . . . . . . . . . .5-13 Fatigue Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-14 Fatigue Analysis (All-Nodes Option) . . . . . . . . . . . . . . . . . . . . . . . . . . .5-17 Fatigue Caused by Thermal Loading Example . . . . . . . . . . . . . . . . . . . .5-17 Creating the Model Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-18 Specifying and Running the Thermal Analysis . . . . . . . . . . . . . . . . . .5-19 Specifying and Running the Stress Analysis . . . . . . . . . . . . . . . . . . . .5-20 Specifying and Running the Fatigue Analysis . . . . . . . . . . . . . . . . . . .5-20 Running the Analysis Based on the Elastic-Plastic Formulation . . . . .5-21
In de x
ii
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
In de x
iii
Contents
In de x
iv
Introduction
Introduction
The suitability of a mechanical or structural component for specific service involving cyclic application of loads and thermal conditions is determined by the use of the fatigue module. The basic postulate adopted for fatigue calculations with spectrum loading is the cumulative damage theory based on the Miners rule.
In de x
1-1
Chapter 1 Introduction
base. The solution is based on the Miners rule approach and the ASME Boiler and Pressure Vessel Code. There is also a capability to evaluate the fatigue usage factor using the ASME Code for a simplified elastic-plastic formulation. The powerful graphics capability of the COSMOSM software can be utilized to display color plots of fatigue life and to identify the fatigue critical regions.
In de x
1-2
Analysis
Introduction
This chapter covers background information, definitions and the procedure for conducting a fatigue analysis.
In de x
2-1
Chapter 2 Analysis
number of cycles n1 smaller than N1 will produce a smaller fraction of damage, say D1. Factor D1 is termed damage fraction (usage factor). Operation over a spectrum of different stress levels results in a usage factor Di for each of the different stress levels Si in the spectrum (in the following development each one of the different load level operations, which may consist of a number of cycles, is called an event). When these factors sum to unity, failure is predicted; that is, (2-1) The linear damage rule states that the damage fraction (usage factor), Di, at stress level Si is equal to the cycle ratio ni/Ni. Thus, the damage fraction D due to one cycle of loading is 1/N. In other words, the application of one cycle of loading consumes 1/N of the fatigue life. The failure criterion for variable amplitude loading can now be stated as (2-2) In the above postulate it is assumed that there is no interaction between different events, i.e. each event (consisting of a number of cycles of the same load level) occurs in complete isolation from the other events. However, in practical applications, every load cycle of the spectrum may contain multiple load levels. The fatigue analysis of FSTAR also includes this latter effect as will be described in more details in the section on Analysis Procedure for Cyclic Loading.
Definitions
General Definitions Loading History
A series of load steps which may occur for a number of times.
Event
A portion of loading history which has a frequency of occurrence.
In de x
2-2
Fatigue Loading
A specified point in the loading history such as where the stress level is at the extreme.
Location
A point on the structure for which the fatigue calculation is to be performed.
Stress Condition
Corresponds to the state of stress at a location for a particular (fatigue) loading.
2-3
Chapter 2 Analysis
Processing Phase
Initial COSMOSM finite element structural results (from STAR, HSTAR, ASTAR, or NSTAR modules) applicable as input to the fatigue calculations.
Scale Factor
Is a stress multiplier. It applies to stresses which are read for a fatigue loading from the database stress file. The database stress file contains the finite element structural solution obtained during the Processing Phase.
Linearized Stress
The equivalent linear stress distribution along a section which has the same net bending moment as the actual distribution.
In de x
2-4
Membrane Stress
The constant portion of stress such that pure moment acts on a section plane after the membrane is subtracted from the equivalent linearized stress.
Bending Stress:
The variable portion of stress equal to the equivalent linear stress minus the membrane stress.
Design Stress
The allowable design stress intensity value (Sm) at a particular temperature as specified by the Sm-T curve. (These curves may be found in tables I-2.1 and I-2.2 of Reference 5).
Sm-T Curve
Design stress versus temperature.
When a design fatigue curve (S-N curve) is not defined, the program evaluates the alternating stress intensities according to the ASME Boiler and Pressure Vessel Code [Reference 1]. Alternating stress intensities are evaluated between all possible pairs of loading combinations. That is, for a defined location, all six stress components of loading A will be subtracted from the corresponding components of loading B to yield component stress range from which the alternating stress intensities are evaluated according to the General Definitions section. These alternating stress intensities are listed for all possible combinations in decreasing order together with their corresponding loading pairs. No usage factor is evaluated in this case. When a design fatigue curve (S-N curve) is defined, the program checks the alternating stress intensity list from the top (highest value) to the bottom (lowest value). It evaluates the partial usage factor Di for the ith alternating stress intensity in that list by evaluating the cycle ratio ni/Ni (as its equivalence). The ith alternating stress intensity is formed by the combination of loadings AE and
2.
In de x
2-5
Chapter 2 Analysis
BF, where E and F are the corresponding events. Here, ni corresponds to the lower number of cycles remaining from either events E and F, and Ni corresponds to the allowable number of cycles interpolated from the design fatigue curve (SN curve). After evaluating the partial usage factor Di, the program updates the alternating stress intensity list by reducing the number of cycles of both events E and F by ni. Consequently, one of the two events E or F will be eliminated (or both if E and F have the same number of cycles) and the other event will have ni cycles less in the later calculations. Elimination of an event results in elimination of the corresponding loadings. Once a loading is eliminated the corresponding stress intensities (formed by combination of that loading with other loadings) will also be eliminated from the list. After updating the list the program checks the next alternating stress intensity in the list and evaluates the corresponding partial usage factor, adds that to the cumulative one and updates the list. This procedure will be repeated for the next alternating stress intensity in the list and continues until all stress intensities are considered.
3.
An S-N curve is defined by log-log interpolation between the points (on the curve) and linear interpolation among all the (S-N) curves with different stress ratios if more than one curve is defined by the user. Also available are two predefined (optional) curves for Carbon or Austenitic Steels which the user may consider instead. For any alternating stress within the stress range S1 and S2 (the first and last points) of an S-N curve the program uses log-log interpolation to find the corresponding cycle. For any stress larger (smaller) than S1 (S2), the program assigns N1 (N2) as the corresponding number of cycles, where N1 and N2 correspond to the first and last points of the curve. Therefore, it is important that the user defines the S-N curve for a wide range of cycles. For the same reason, if an endurance limit is to be implemented into the curve, it is appropriate to have a relatively large number of cycles assigned to the last point (N2) on the curve.
All-Nodes Option
This option is similar to the normal procedure except that the fatigue calculation is performed for a structure at all nodes. The program has the following limitations:
1.
In de x
Since the calculation is done on the nodal basis, no location has to be defined (all defined locations will be ignored).
2-6
All defined concentration factors will be ignored. All stresses are read either from stress files created during the processing phase or considered to be zero if they are so defined (using command FT_LOAD (Analysis > FATIGUE > Fatigue Load). No modification on the nodal stresses are allowed (i.e., any assignment made by the command FT_STREAD (Analysis > FATIGUE > Apply Stress) is ignored). No elastic-plastic formulation will be incorporated into the calculation at any node.
4.
For an axisymmetric structure Figure 2-1. Two Types of Sections Defined such as a pressure vessel and a by Locations at Their Two Ends fatigue critical location (on the inner or outer surface), define a C L thickness-through section with one end at the desired location (Figure 2-1). The section must present a rational plane of 1 2 bending which, in most cases, will be perpendicular to both 4 Y surfaces and the mid-plane. For most of the structures with 3 parallel surfaces, such as pipes, shells and external nozzles, this X criteria can be easily met. However, in the irregular areas such as in the nozzle to shell juncture, rational planes of bending may be approximated so as to be perpendicular to the mid-plane and have the same angle between the section and the surface on both sides (see Chapter 6). Define an Sm-T curve. Typical values of Sm are given in Tables I-1.1 and I-1.2 of Reference 5.
2.
In de x
2-7
Chapter 2 Analysis 3.
If an Sm-T curve is defined, the calculation should proceed by evaluating the equivalent linearized stresses along the defined section (Figure 2-2). The linearized stresses are the sum of the bending stresses and the membrane stresses as outlined in Chapter 1.
Figure 2-2. Equivalent Linearized Stress Along a Sections
S
Linear Stress
(Membrane + Bending)
Membrane
4.
Each alternating stress intensity (evaluated according to the Normal Procedure section) is increased by a factor Ke. This factor is determined by considering an equivalent alternating stress intensity based on linearized stresses (see Chapter 6).
In de x
2-8
Description of Elements
Introduction
The following table shows the elements that are used in the FSTAR program. For detailed descriptions of each element, you are referred to Chapter 4 of the COSMOSM User Guide manual.
Table 3-1. Elements for FSTAR Element Type 2D Spar/Truss 2D Elastic Beam 3D Elastic Beam 3D Spar/Truss Elastic Straight Pipe Boundary Element General Mass Element Elastic Curved Pipe 2D 4- to 8-Node Plane Stress, Strain, Body of Revolution 3D 3- to 6-Node Plane Stress, Strain, Body of Revolution Triangular Thick Shell Quadrilateral Thick Shell 3D 8- to 20-Node Continuum Brick Element Name TRUSS2D BEAM2D BEAM3D TRUSS3D PIPE BOUND MASS ELBOW PLANE2D TRIANG SHELL3T SHELL4T SOLID
In de x
3-1
Table 3-1. Elements for FSTAR (Concluded) Element Type 3D 4-Node Tetrahedron Solid 3D 10-Node Tetrahedron Solid 2-Node Gap/with Friction Triangular Composite Shell Quadrilateral Composite Shell Triangular Thin Shell Quadrilateral Thin Shell Spring Element 3D 4-Node Tetrahedron Solid with Rotation Axisymmetric Shell General Stiffness 8 or 9-Node Isoparametric Shell Element 8 or 9-Node Isoparametric Composite Shell 8-Node Composite Solid 2-Node Rigid Bar 3D 8- to 20-Node Isoparametric Piezoelectric Solid 6-Node Shell Element Element Name TETRA4 TETRA10 GAP SHELL3L SH3LL4L SHELL3 SHELL4 SPRING TETRA4R SHELLAX GENSTIF SHELL9 SHELL9L SOLIDL RBAR SOLIDPZ SHELL6
In de x
3-2
In de x
4-1
Specifies the number of cycles Defines fatigue loading Defines stress conditions Defines fatigue properties Defines a fatigue location Defines a section for elastic-plastic formulation Specifies the face and layer number of multilayered shell or solids Performs the fatigue analysis
List fatigue events and their specifications Lists stress conditions Lists all defined fatigue properties Lists specifications for a pattern of fatigue locations Lists specifications for a pattern of fatigue sections
In de x
4-2
Deletes a pattern of fatigue events Deletes a pattern of fatigue loadings Deletes stresses associated with a pattern of fatigue locations Deletes fatigue property specifications Deletes a pattern of fatigue locations Deletes a pattern of fatigue sections
Loads the cumulative fatigue usage factors into the plot buffers Plots the previously loaded fatigue cumulative factor Lists the cumulative fatigue usage factors
In de x
4-3
In de x
4-4
Define the required events with the FT_EVENT (Analysis > FATIGUE > Event
Cycle) command.
2.
Define fatigue loadings using the FT_LOAD (Analysis > FATIGUE > Fatigue Load) command and associate them with the stresses stored by the COSMOSM structural solution. [Stresses are available for fatigue analysis from the nonlinear module only for those time steps which are specified by the NL_PLOT (Analysis > NONLINEAR > Plot Options) command and for Advanced Dynamics, for those specified by the PD_PLOT (Analysis > POST_DYNAMIC > PD_OUTPUT > Set Plot Options) command].
In de x
5-1
You may use the FT_STREAD (Analysis > FATIGUE > Apply Stress) command to input stresses directly. (This command may also be used to modify stresses stored as the result of the initial finite element structural solution). Define fatigue locations with the FT_LOC (Analysis > FATIGUE > Fatigue Location) command. (Not required for all-nodes calculation option). Define fatigue design curves (the S-N curves) with the FT_CURDEF (Analysis > FATIGUE > Property Curve) command or use the A_FATIGUE (Analysis > FATIGUE > Analysis Options) command to specify one of the two pre-defined (optional) S-N curves. Modulus of elasticity must be defined in psi (using the MPROP (Propsets > Material Property) command), if an optional design curve is considered. The predefined (optional) curves are defined in psi versus cycle. (If no S-N curve is defined, the fatigue calculation will not produce usage factors). For shell elements specify the top or bottom face and the layer number with the A_FATIGUE (Analysis > FATIGUE > Analysis Options) command. Activate a location for fatigue calculation with the ACTSET, LOC, ... (Control > ACTIVATE > Set Entity, Loc) command (for the all-nodes option activate zero location). Run the FSTAR module by executing the R_FATIGUE (Analysis > FATIGUE > Run Fatigue Analysis) command. Repeat steps (7) and (8) for other locations if the calculations are made on the basis of one location at a time. If you intend to save the results of the previous calculations in the output file, activate the append flag of the PRINT_OPS (Analysis > OUTPUT OPTIONS > Set Print Options) command.
4.
5.
6.
7.
8.
9.
You may review all your inputs at any time during the editing session or afterward by using commands: FT_EVENTLIS (Analysis > FATIGUE > FATIGUE LIST > Events), FT_LOCLIST (Analysis > FATIGUE > FATIGUE LIST > Locations), FT_CURLIST (Analysis > FATIGUE > FATIGUE LIST > Property Curves), FT_STLIST (Analysis > FATIGUE > FATIGUE LIST > Stress Conditions) or review the result of fatigue calculations after each run with the command: FTGLIST (Results > LIST > Fatigue Usage Factor)
In de x
5-2
You can also check both the inputs and outputs by reviewing the file with the extension .FTG. You may delete/ modify your inputs at anytime during the editing session by using the commands: FT_EVENTDEL (Analysis > FATIGUE > FATIGUE LIST > Events), FT_LOADDEL (Analysis > FATIGUE > FATIGUE DELETION > Loads), FT_LOCDEL (Analysis > FATIGUE > FATIGUE DELETION > Locations), FT_STDEL (Analysis > FATIGUE > FATIGUE DELETION > Stresses), FT_CURDEL (Analysis > FATIGUE > FATIGUE DELETION > Property Curves). You may have color plots of the results (for all-nodes option) by issuing ACTFTG and FTGPLOT (Results > PLOT > Fatigue) commands. For the simplified elastic-plastic calculation (for axisymmetric models) add the following steps:
10.
Define a section through the wall thickness with the FT_SEC (Analysis > FATIGUE > Fatigue Section) command (first define locations at the two ends of the section by using the FT_LOC (Analysis > FATIGUE > Fatigue Location) command) Define the Sm-T curve and material parameters M and N with the FT_CURDEF (Analysis > FATIGUE > Property Curve) command. Activate one of the two end locations and run fatigue (steps g and h).
11.
12.
You may use the FT_SECLIST (Analysis > FATIGUE > FATIGUE LIST > Sections), FT_CURLIST (Analysis > FATIGUE > FATIGUE LIST > Property Curves), FT_SECDEL (Analysis > FATIGUE > FATIGUE DELETION > Sections), and FT_CURDEL (Analysis > FATIGUE > FATIGUE DELETION > Property Curves), commands for the listing and deleting of the inputs in steps j and k. The following three examples are designed to clarify the concept of cumulative damage theory and to show how it is implemented in the COSMOSM Fatigue module.
In de x
5-3
History Loading
The load history is defined according to Figure 5-1. The number of cycles for each event is specified by n1, n2 and n3.
Figure 5-1. Loading History
x (psi) x
50,000 P
20,000
-30,000
Event 1 n1 = 800 Event 2 Event 3 n2 = 2300 n3 = 1200
In de x
5-4
You may modify/correct any event specification by repeating command FT_EVENT (Analysis > FATIGUE > Event Cycle) for that event. To delete an event use command FT_EVENTDEL (AnalysIs > FATIGUE > FATIGUE DELETION > Events).
x (psi)
2 Loading
6 1 3
In de x
Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number> 2 Associated event > 1 Associated load case > -1
5-5
Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number> 3 Associated event > 2 Associated load case > 0 Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number > 4 Associated event > 2 Associated load case > -1 Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number > 5 Associated event > 3 Associated load case > 0 Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number > 6 Associated event > 3 Associated load case > -1
You may review all your input so far by using list command FT_EVENTLIS (Analysis > FATIGUE > FATIGUE LIST > Events). You may modify/correct any loading specification by repeating command FT_LOAD (Analysis > FATIGUE > Fatigue Load), or may delete a loading by using command FT_LOADDEL (Analysis > FATIGUE > FATIGUE DELETION > Loads).
Caution: Deleting an event erases all loadings which were associated to that
event, i.e., all parameters which are otherwise listed for that event [using command FT_EVENTLIS (AnalysIs > FATIGUE > FATIGUE LIST > Events)] will be deleted.
Defining a Location
Define a location for fatigue calculation.
Geo Panel: Analysis > FATIGUE > Fatigue Location (FT_LOC) Reference number for fatigue location > 1 Associated node label > 1
For this example Node label is irrelevant. It is useful for problems for which the stress conditions are available in the database from initial COSMOSM processing phase calculations.
In de x
Stress concentration in X dir > 1.0 Stress concentration in Y dir > 1.0
5-6
Stress concentration factors are considered to be unity in all directions. You may review your inputs by using list command FT_LOCLIST (Analysis > FATIGUE > FATIGUE LIST > Locations) (use default for all prompts). Dashed lines under headings X, Y, Z, and CS are due to the irrelevance of nodal coordinates in this example.
(For location 1, loading 2, define SX = 50,000. All other stress components are zero).
Geo Panel: Analysis > FATIGUE > Apply Stress (FT_STREAD) Location label > 1 Fatigue loading label > 4 Item number > Actual stresses Normal stress in X direction > -30000 Normal stress in Y direction > 0.0 Normal stress in Z direction > 0.0 Shear stress TAU_XY > 0.0 Shear stress TAU_XZ > 0.0 Shear stress TAU_YZ > 0.0 Geo Panel: Analysis > FATIGUE > Apply Stress (FT_STREAD) Location label > 1 Fatigue loading label > 6
In de x
5-7
Item number > Actual stresses Normal stress in X direction > -20000 Normal stress in Y direction > 0.0 Normal stress in Z direction > 0.0 Shear stress TAU_XY > 0.0 Shear stress TAU_XZ > 0.0 Shear stress TAU_YZ > 0.0
To review the stress conditions specified by command FT_STREAD (Analysis > FATIGUE > Apply Stress) use list command FT_STLIST (Analysis > FATIGUE > FATIGUE LIST > Stress Conditions).
Geo Panel: Analysis > FATIGUE > FATIGUE LIST > Stress Conditions
(FT_STLIST)
First location > 1 Last location > 1 Increment > 1 Loading label > All loadings Item number > Actual stresses
You may use command FT_STREAD (Analysis > FATIGUE > Apply Stress) to modify any stored stress condition (stress may have been stored by using command FT_STREAD (Analysis > FATIGUE > Apply Stress) or pre-stored in the data base in the processing phase) for any combination of location and loading. Command FT_STDEL (Analysis > FATIGUE > FATIGUE DELETION > Stresses) may be used to delete stress conditions stored using command FT_STREAD (Analysis > FATIGUE > Apply Stress).
Default implies that the user is going to specify points on the S-N curve
Stress ratio R > -1
Min. over Max. Stress. Default -1 implies a fully reversible stress cycling
In de x
5-8
Cycles at point 1 > 2000 Stress at point 1 > 80E3 Cycles at point 2 > 5000 Stress at point 2 > 50E3 Cycles at point 3 > 40000 Stress at point 3 > 30E3 Cycles at point 4 > 500000 Stress at point 4 > 20E3
To review your inputs for the fatigue design curve you may use the commands ACTXYPRE (Display > XY PLOTS > Activate Pre-Proc) and XYPLOT (Display > XY PLOTS > Plot Curves). The above five points define a fatigue design curve according to Figure 5-3.
60
a lt
x x
20
(ksi) 40
x
4 5 6 7
x
10
0 3 10
10
10
10
10 8
Cycle
Geo Panel: Display > XY PLOTS > Activate Pre-Proc (ACTXYPRE) Graph number > 1 Curve type > Fatigue Curve type > SN Curve number > 1 Graph color > 12 Graph line style > Solid Graph symbol style > Circle Geo Panel: Display > XY PLOTS > Plot Curves (XYPLOT)
In de x
5-9
In de x
5-10
Geo Panel: Control > ACTIVATE > Set Entity (ACTSET) Set label > Loc Location for fatigue calculation > 1
(This is necessary because the program is designed in such a way that the user may define more than one location). To run the fatigue problem, use command R_FATIGUE (Analysis > FATIGUE > Run Fatigue Analysis). When the analysis is complete, the program will return to the GEOSTAR menu. The resulting outputs are stored in the output file with extension .FTG. The results may be reviewed either with the editor, using the command EDIT (File > Edit...), or using the list command FTGLIST (Results > LIST > Fatigue Usage Factor).
Geo Panel: Results > LIST > Fatigue Usage Factor (FTGLIST) Location label > 1
In de x
5-11
Chapter 5 Detailed Description of Examples Table 5-2. Results [Reference 1] Theoretical Fatigue usage factor at location 1 7.4% COSMOSM 7.4%
For more complicated problems, often an initial finite element structural analysis is necessary in order to find the stress conditions. The following example illustrates the use of COSMOSM. Modules such as STAR, NSTAR, or HSTAR, to evaluate the stress conditions needed for fatigue analysis.
P (psi) 3
20,000
x
P
2"
x
2
10,000
x
x Loading
1"
4"
x
1
0.02 Event 1 0.1 0.12
4 t
3" 5"
2"
0.2 0.22
(10,000 cyc)
In de x
5-12
To evaluate the usage factor (percentage of life used up) at different locations, first the stress conditions at extreme points within a cycle will be evaluated. Therefore, we need only to evaluate the stress field for the two stress levels 10,000 and 20,000 psi.
Given
E = 30 x 106 psi = 0.3
In de x
5-13
At this stage you have the option of analyzing the fatigue problem at one location (node) at a time or at all nodes at once. The first option provides you with more details of the calculation. However, the all nodes option relieves you from the painstaking task of studying each location one by one and provides you with a graphical display of the fatigue life. These options are outlined in the following:
It is recommended to first identify the areas of stress concentration. This may be done, quantitatively, by studying the stress distribution inside the structure using commands in the Results menu. From the displayed contours it is apparent that the tip of the nozzle, at node 105, experiences the highest level of stress.
Geo Panel: Analysis > FATIGUE > Event Cycle (FT_EVENT) Reference number for the event > 1 Number of required cycles > 10000 Geo Panel: Analysis > FATIGUE > Event Cycle (FT_EVENT) Reference number for the event > 2 Number of required cycles > 2000 Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number > 1 Associated event > 1 Associated load case > 0
5-14
Geo Panel: Analysis > FATIGUE > Fatigue Load (FT_LOAD) Reference number > 4 Associated event > 2 Associated load case > 0 Geo Panel: Analysis > FATIGUE > Fatigue Location (FT_LOC) Reference number for fatigue location > 1 Associated node label > 105 Stress concentration in X dir > 1.0 Stress concentration in Y dir > 1.0 Stress concentration in Z dir > 1.0 Geo Panel: Analysis > FATIGUE > Property Curve (FT_CURDEF) Curve/prop item number > 1 Stress ratio R > -1 Cycles at point 1 > 2000 Stress at point 1 > 80E3 Cycles at point 2 > 4000 Stress at point 2 > 60E3 Cycles at point 3 > 10000 Stress at point 3 > 30E3 Cycles at point 4 > 100000 Stress at point 4 > 5E3 Geo Panel: Analysis > FATIGUE > Property Curve (FT_CURDEF) Curve/prop item number > 1 Stress ratio R > -1 Cycles at point 5 > 10000000 Stress at point 5 > 1E3 Geo Panel: Analysis > OUTPUT OPTIONS > Set Print Options (PRINT_OPS) Displacement print flag > Yes ... Output flag > Append Geo Panel: Control > ACTIVATE > Set Entity (ACTSET) Set label > Loc Location for fatigue calculation > 1
In de x
5-15
Geo Panel: Analysis > FATIGUE > Run Fatigue Analysis (R_FATIGUE)
You may review the output file by using the EDIT (File > Edit...) command.
Geo Panel: File > Edit a File (EDIT)
There are other alternative ways of solving a linear static problem such as the
one considered above which may significantly reduce the computational time. One alternative to the above is running the static problem for only one load case (e.g., load case 1) and then associating fatigue loadings 2 and 3 with that load case with a scale factor of 1 and 2.0, respectively.
4.
This results in the modification of structural modeling part of the problem as follows (type the cryptic input in the command window):
GEO> ACTSET,LC,1 GEO> PEL,33,10000.0,1,36,1, GEO> PEL,81,10000.0,1,86,1, GEO> PEL,49,10000.0,1,52,1, GEO> PEL,65,10000.0,1,68,1, GEO> A_STATIC,N,0,0,1e-06,1e+10,0,0,0, GEO> R_STATIC
In de x
The above modification is applicable only to the linearly elastic materials for which the stress (response) is linearly proportional to the applied load.
5-16
You may review the output file by using the EDIT (File > Edit...) command, or issuing the SYSTEM command and using your favorite editor.
Geo Panel: File > Edit... (EDIT)
Review results
2.
At this step, you may utilize the graphic capability of GEOSTAR to display color plots of the fatigue life and identify fatigue critical regions. Fatigue curves can be plotted using the XYPLOT (Display > XY PLOTS > Plot Curves) command.
Geo Panel: Results > PLOT > Fatigue (ACTFTG)
5-17
Assume that the nozzle of example 2 is exposed to a fluid heat up condition which is expected to occur 5000 times during its service life. One complete cycle of this heat up condition is shown in Figure 5-6.
Figure 5-5.
Given
E Kx h h Tref = 30*10 psi = 0.3 = 8.0E-6 in/in/F = 0.1 BTU/in hr F = 1.0 BTU/in2 hr F (outside surface) = 5.0 BTU/in hr F (inside surface)
2
3"
2"
3 x 1"
4"
2" 5"
= 60 F
Tmax = 300 F
T e m p e r a t u r e
300
150
GEO> TITLE, FT3A: FATIGUE OF A NOZZEL WITH CYCLIC INTERNAL FLUID GEO> PLANE,Z,0,1, GEO> VIEW,0,0,1,0, GEO> CRSPOLY,1,8,0,0, L,8,2,0, L,6,2,0, A,5,3,0,
In de x
5-18
L,5,6,0, L,4,8,0, L,4,10,0, L,3,10,0 L,3,8,0, L,3,6,0, L,3,3,0, L,3,0,0, L,6,0,0, L,8,0,0, GEO> SCALE,0, GEO> CRBRK,3,3,1,2,0, GEO> SF2CR,1,13,2,0, GEO> SF2CR,2,12,3,0, GEO> SF2CR,3,11,14,0, GEO> SF2CR,4,10,4,0, GEO> SF2CR,5,9,5,0, GEO> SF2CR,6,8,6,0, GEO> EGROUP,1,PLANE2D,0,1,1,0,0,0,0, GEO> M_SF,1,3,1,4,4,4,1,1, GEO> M_SF,5,6,1,4,4,4,1,1, GEO> M_SF,4,4,1,4,6,4,1,1, GEO> DCR,12,UY,0,13,1,,
In de x
5-19
GEO> CEL,4,1.,60.,2,16,4,0, GEO> CEL,13,1.,60.,3,16,1,0, GEO> CEL,29,1.,60.,3,32,1,0, GEO> CEL,45,1.,60.,3,48,1,0, GEO> CEL,99,1.,60.,3,104,1,0, GEO> CEL,61,1.,60.,3,64,1,0, GEO> CEL,77,1.,60.,3,80,1,0, GEO> CEL,65,5.,1.,1,68,1,1, GEO> CEL,49,5.,1.,1,52,1,1, GEO> CEL,81,5.,1.,1,86,1,1, GEO> CEL,33,5.,1.,1,36,1,1, GEO> NMERGE,1,160,1,0.0001,0,1,0, GEO> CRMERGE,1,20,1,0.0001,1,1,0, GEO> TIMES,0.000000E+00,0.200000,0.100000E-01, GEO> TUNIF,60.0000, GEO> TOFFSET,273.000, GEO> INITIAL,TEMP,1,159,1,60., GEO> A_THERMAL,T,0.100000E-02,5,1,20,0, GEO> R_THERMAL
Stresses are available for postprocessing (fatigue analysis or stress plot) from
the nonlinear module only for those time steps which are specified by command NL_PLOT (Analysis > NONLINEAR > Plot Options).
Fatigue analysis. Type the following cryptic input in the command window.
5-20
GEO> FT_EVENT,1,5000, GEO> FT_LOAD,1,1,2,1, GEO> FT_LOAD,2,1,4,1, GEO> FT_LOAD,3,1,6,1, GEO> FT_LOAD,4,1,8,1, GEO> FT_LOAD,5,1,10,1, GEO> FT_LOAD,6,1,12,1, GEO> FT_LOAD,7,1,14,1, GEO> FT_LOAD,8,1,16,1, GEO> FT_LOAD,9,1,18,1, GEO> FT_LOAD,10,1,20,1, GEO> FT_LOC,1,130,1,1,1, GEO> FT_CURDEF,1,-1,2000,80E3,4000,40E3,10000,10E3,100000,3.E3, GEO> ACTSET,LOC,1, GEO> R_FATIGUE
If you wish to perform fatigue analysis based on a simplified elastic-plastic formulation then continue with:
You may review all the results by using either EDIT (File > Edit...) or FTGLIST (Results > LIST > Fatigue Usage Factor) commands.
In de x
5-21
In de x
5-22
C L
In de x
6-1
Stress Linearization
For plane strain, plane stress and axisymmetric structures, stress linearization is performed along a section path. A section is defined by nodes N1 and Nn according to Figure 6-2. The program interpolates n-2 equally spaced (integration) points along the path between nodes N1 and Nn. For each point, structural elements are searched to identify an element which contains that point. Once the element is identified, the stresses at that point are interpolated linearly from the element corner nodes. These stresses are denoted as actual stresses.
Figure 6-2. A Section with n Equally Spaced Integration Points Along its Path
C L
Nn N1 N4 N2 N3 N n-1
The equivalent linearized stresses are found along a section by evaluating the membrane and bending stresses according to one of the following two methods.
In de x
6-2
(6-1)
where, superscripts m and a correspond to membrane and actual stresses, respectively, l is the length of the section, x and y are the local coordinates along and perpendicular to the section path (Figure 6-3) with an origin at the mid-wall, and X, Y and Z are the global Cartesian coordinates. Membrane stresses are considered to be constant along the section. The bending component of stresses at position x along the section path (Figure 6-3) is calculated according to:
C L
y x,
Nn /2
N1 /2 X
(6-2)
where superscript b corresponds to the bending stress. The linearized stress at any point along the section is the sum of membrane and bending stresses.
In de x
6-3
Axisymmetric Formulation
In this case the axisymmetric features of the model is more properly implemented in the formulation, for axisymmetric structures, than the Cartesian case. This option is considered in fatigue calculation if a non-zero value is assigned to the input quantity curvature_radius in command FT_SEC (Analysis > FATIGUE > Fatigue Section). Curvature_radius corresponds to the radius of curvature of the average mid-wall centerline in the xy plane as represented by in Figure 6-4. A large value (or -1) for curvature_radius corresponds to straight walls (e.g., cylinder or cone). In order to find the linearized stresses, it is desired to obtain applied forces and moments along the section. Figure 6-5 represents a freebody diagram of the section. A righthanded local coordinate system, x, y, and z is established on the section with the origin at the mid-wall (same as in Figures 6-3 or 6-4). FN and FT correspond to the inplane normal and shear forces on the section in y and x directions and MZ is the bending moment. The three inplane forces and moment on the section over a small sector in the hoop direction are defined as:
Figure 6-4. Curvature-Radius of the Mid-Wall
C L
x Nm
x
Y N1 X
C L
FT MZ Y
FN
MZ FT
x,
In de x
6-4
(6-3)
where, ya and xya are the actual stresses along the section in local coordinates, X is the global coordinate (or equivalently the radius) of point being integrated along the section and xf in local coordinates is the offset of the sector neutral axis from the center line. From FN in the above equation, the average normal membrane stress in the y direction is computed once it is divided by the sector area (Xc is the global coordinate of the mid-section).
(6-4)
The MZ contribution to the normal bending stress in the y direction is computed by applying the familiar relationship s = Mz (x-xf)/I with I as moment of inertia of the sector.
(6-5)
An average membrane shear stress (xy component) is computed by dividing FT [in Equation (6-3)] by the sector area, assuming that xy bending shear stress has insignificant contribution (since the shear stress distribution is assumed to be parabolic and equal to zero at the two free surface ends).
In de x
6-5
(6-6) Average membrane stress in the x direction is computed by averaging the actual stresses along the section according to:
(6-7) The bending stress in the x direction (thickness direction) is ignored if in command FT_SEC (Analysis > FATIGUE > Fatigue Section) the corresponding flag is set to one, otherwise it is approximated at the two ends as the difference of the actual and membrane stresses. The hoop membrane and bending stresses are calculated by considering a small sector () in the XY plane. By integrating the total normal force on the sector and averaging it over the corresponding area, the average membrane stress (in hoop direction) is computed according to:
(6-8)
For straight walls where , Equation (6-8) reduces to the familiar form. The hoop bending stress is calculated by evaluating the applied bending moment on the sector. Once the bending moment is evaluated, the hoop bending stress is found analogous to that of the y direction bending stress.
(6-9)
In de x
6-6
As mentioned before, the linearized stress at any point along the section is the sum of membrane and bending stresses.
(6-11)
(6-12) where, M and N are the elastic-plastic material parameters (input quantities on FT_CURDEF (Analysis > FATIGUE > Property Curve) command). References
1.
ASME Boiler and Pressure Vessel Code, Edition 1983, Section III, Division 1, Subsection NB. Kroenke, W. C., Addicott, G. W. and Hinton, B. M., Interpretation of Finite Element Stresses According to ASME Section III, Paper 75-PVP-63, ASME Second National Congress on Pressure Vessels and Piping, June 1975. Kroenke, W. C., Classification of Finite Element Stresses According to ASME Section III Stress Categories, Pressure Vessels and Piping, Analysis and Computer, ASME, June 1974.
2.
3.
In de x
6-7
Gordon, J. L., Outcur: An Automated Evaluation of Two-Dimensional Finite Element Stresses According to ASME Section III Stress Requirements, Paper 76-WA/PVP-16, December 1976. ASME Boiler and Pressure Vessel Code, Edition 1989, Section III, Division 1, Appendices.
5.
In de x
6-8
Index
A Analysis Options 4-2, 5-2 Apply Stress 4-2, 5-2, 5-7, 5-8 ASME Boiler and Pressure Vessel Code 1-2, 6-7, 6-8 B Bending Stress 6-3, 6-5, 6-6 D Design Stress 6-7 E elastic-plastic formulation 1-2,
4-2, 5-21, 6-1 Event Cycle 4-2, 5-1, 5-5, 5-14 Events 1-1, 4-2, 4-3, 5-1, 5-2, 5-3, 5-5, 5-6, 5-11
fatigue design curve 5-8, 5-9 fatigue life 1-2, 5-14, 5-17 fatigue loading 4-2, 5-7 Fatigue Location 4-2, 5-2, 5-3, 5fatigue properties 4-2 Fatigue Section 4-2, 5-3, 6-2, 6-4, fatigue usage factor 1-1, 1-2, 5-2,
5-11, 5-12, 5-16, 5-21 6-6 6, 5-15
S Sections 4-2, 4-3, 5-3, 6-1 Set Entity 5-2, 5-11, 5-15 Sm-T Curve 5-3, 6-7 S-N curve 5-2, 5-8, 5-9 Stress Conditions 1-1, 4-2, 5-2, 5Stresses 1-1, 4-3, 5-1, 5-2, 5-3, 55, 5-7, 5-8, 5-14, 5-20, 6-2, 6-3, 64, 6-5, 6-6, 6-7, 6-8 6, 5-7, 5-8, 5-11, 5-12, 5-13
L Linearized Stress 6-3, 6-7 load history 5-4 Loads 1-1, 4-3, 5-3, 5-6 Locations 4-2, 4-3, 5-2, 5-3, 5-7,
5-13, 5-14, 5-16
U usage factor 1-1, 1-2, 5-2, 5-11, 512, 5-13, 5-16, 5-21
M Material Property 5-2 Membrane Stress 6-5, 6-6 P Plot Curves 5-9, 5-17 Plot Options 5-1, 5-20 Print Options 5-2, 5-15 Property Curves 4-2, 4-3, 5-2, 5-3
In de x
I-1