Structural Post Questions


Q: What corrdinate system is used for the reports in the Structural Post Processing menu ?
REV 7.05

A: The reports in the Structural Post Processing menu are reported in the beam coordinate system. You can ask MOSES to tell you what this system is in the &summary menu with the BEAM command.


Q: Why do the AISC LRFD results reported not match my hand calculations?
REV 7.05

A: The problem is due to your model. The option

       -SECTION
was used to change the area and inertias of the section (in particular, the area was changed by increased by a factor of 85). The options -P_N and -M_P were however, not used to change theses factors. According to the documentation, if a value is not overridden with an option, the computed value will be used. For this case, the Pn value used is probably 1/85 smaller than what you would like.

A good rule is that you should model as close to what you have as possible and override as few items as possible. I can think of no  reason at all to override the area of a "prismatic" beam. My guess is that is it not prismatic at all but a shape that would be better modeled with a tube, box, or W.


Q: Why does MOSES not follow the standard scaling rule for damage?
REV 7.04

A: The standard scaling rule to which you refer says that if the damage for a stress concentration SCF1 is D, then the damage for a stress concentration SCF2 is D*(SCF2/SCF1)**M where M is the slope of the SN curve and the all of the contributions to the damage are within the curve. An important thing to notice here is that almost all SN curves have more than one slope and the above is valid only for a single slope SN curve.

All published SN curves all have limits, and what are we to do
outside these limits? Before Rev 5.06, we actually computed the
damage based on the curve. We found that this can be seriously
wrong. In particular, if one has a beam with really large stresses,
then the larger the stresses, the smaller the CDR! This is because
as the stress becomes large, more of it is outside the range
of the SN curve. We have now changed this so that if the
probable maximum stress in greater than the first point in the
SN curve, damage is accumulated as if it were all at the first
point on the curve. This gives some discontinuities in the CDRs,
but it is better than overly optimistic results.


Q: Are the "Stress Cycles" reported with "counting" peak to trough values or amplitudes?
REV 7.06

A: These results are the exception to the rule that nothing in MOSES is peak to trough. Stress Cycles are peak to trough to be consistent with the definition given in almost all sources.


Q: Why did I not get an increase in the CDR of A**m (m is the slope of the SN curve) when I increased the SCF by A?
REV 7.02

A: Your approximation is only good provided that there are no cycles below the endurance limit or above the left most point in the curve.


Q: Why do I get "Hydrostatic Check" results for a member that is not in the water?
REV 7.01

A: MOSES interprets these checks to be checks on the hoop stress. Thus, cone shape members get hydrostatic collapse checks because they have hoop stress.


Q: How can I get MOSES to resize stiffener spacing?
REV 7.01

A: Everything is done via class. Create classes with different spacings and resize. Click Here for an example.


Q: Why, if I use -USE_MEAN NO on an environment to check only the dynamic load, do I get only positive values.
REV 7.01

A: Without a mean, the stress is cyclical with the same positive peak as the negative valley. If you are going to do a code check. You should make the negative values with a combination; i.e.

  &data env
      env wave -sea ......
  end_data
  cases -spectral wave
  cases -combine c_wave wave -1.
  beam code -load c_wave


Q: It seems like MOSES reports the CDR's for the W curve regardless what SN curve we define with the -SN option on the BEAM_POST command. What is going on here?
REV 7.00

A: In the past (before REV 6.02) the SN curve used for beam fatigue was specified on the BEAM_POST command. It no longer is and the -SN option on BEAM_POST is ignored. Now one has an SN curve associated with both ends and at each change in section. These SN curves are specified when the BEAM or PLATE is defined, so you need to make sure the -SN option with the correct SN curve is specified there. Click here to read the documentation (this link and the following subsections) or Here to read what the REV 6.02 release document has to say on the subject.


Q: How does MOSES compute "Joint Crushing"?
REV 7.00

A: Click here and look in the Miscellaneous Topics section and then the Joint Crushing subsection.


Q: MOSES did not choose the correct chord for a joint, how do I redefine it?
REV 7.00

A: MOSES finds the largest "thru member" and proclaims it the chord. The only control the user has over this is to alter your model so the largest thru member is the one you want.


Q: Where on the section are the eight points at which MOSES computes stresses (and fatigue)?
REV 7.00

A: For tubes they are spaced at equal angles around the circle. For other shapes it is more complicated. The first four are the "bending" points which are at the points at the extremes; .i.e. (max y, max z), (min y, max z), (max y, min z), and (min y, min z). Now, any non tube can be viewed as being composed of several "boxes". Four points will be added such that points 1 and 3 have the z coordinate of the maximum and minimum z coordinate of the centroids of the boxes and points 2 and 4 will have the y coordinate of the maximum ad minimum y of the centroid of the boxes.






Q: How do I know what SCF is used for determining a fatigue CDR?
REV 7.00

A: The SCF used depends on the classification of the joint which depends on the loads in the chord and the braces. What MOSES does is to assign a fraction of each type of joint to the joint in question; i.e. the joint is A K joint, B T joint, and C X joint and A + B + C = 1. So the SCF that MOSES uses is:

     SCF = A * SCF_k + B * SCF_T + C * SCF_x
The Joint SCF Summary provides the SCFs for all 3 joint types and for Punching Shear the factors A, B, and C are printed. For Fatigue, however, this is simply too much information to present. The only way you really know which SCF is being used is if you use -cls_mean yes. Otherwise, you get a new joint classification for each force response operator.


Q: I have several members that are governed by "Hyd. Tension". How can this be since the structure is out of the water?
REV 7.00

A: Your deck model includes cone elements. According to RP 2A-WSD 21st Edition, 3.4.1.c Subparagraph 2, hoop stress must be checked for conical transitions. This provides the same comments in the code check as for hydrostatic collapse. This only applies to MOSES Rev 6.02 and higher.


Q: Do the API ratios obtained during a hydrostatic collapse check on a beam include the Safety Factors defined by the API for Axial and Hoop stresses (1.67 and 2.0)?
REV 7.00

A: Of course they include the safety factors - without them it would not be a code check.


Q: How can I check a joint between two wide flange sections?
REV 7.00

A: In MOSES, we use joint only for a tubular joint, not a connection of two non tubes. For beam fatigue, MOSES checks both ends and any section change. The ends are what you are calling the "joint". The reason for checking the intermediate points is that you have SCFs for changes in OD and T, for cones, etc.


Q: Why when I check a cylindrical member subjected to combined hydrostatic pressure, axial tension and bending manually do I get a different RP2A unity ratio than MOSES?
REV 6.02

A: According to RP2A, the following equation should be satisfied in this case:

A**2+B**2+2*ν*|A|*B≤1    (3.3.3-1)
Since the equations 3.3.1-2 and 3.3.3-1 check the same thing we take a square root here in order to be consistent and get unity ratios closer to those of 3.3.1-2. when a zero hoop stress is applied. Notice that this change does not effect whether or not a member satisfies the code.


Q: We cannot find certain joints in the fatigue report. Does MOSES print out fatigue results for all joints, or just those joints with damage to report?
REV 6.02

A: If there is no damage on a joint, it does not get reported. Now, we print the Cumulative Damage Ratios to the 5th decimal place, so if this happens, there really is no damage to report.


Q: Why do we obtain 8 CDRs for each end of beam of a prismatic member?
REV 6.01

A: Why should you "get only 4"? In some cases the maximum stress is not at the extremes of the section, so we compute stress at 8 points that will normally govern: 4 extreme points, the intersection of the two neutral axes, and the 3 most extreme points on a neutral axis.


Q: Why when I analyze the same model using the old and new Moses version I get big differences in CDR values?
REV 5.10

A: When you run the new version of the program you get small differences in stress values due to program improvements. One should keep in mind that fatigue damage is approximately proportional to stress raised to the power of the inverse slope of the SN curve. Since CDR is a sum of ratios of the applied number of cycles and the number of cycles for which the given stress range would be allowed, then small changes in stress result in much greater changes in CDR value. For example, just a 5% stress difference gives a 20% to 23% difference in CDR depending on the type of the SN curve.


Q: How can I resize beams and change only the wall thickness?
REV 5.08

A: When beams are resized, the possible sizes are defined by a selector. Thus, define a selector that only changes the thickness. For example:

     &SELECT :LEG -SEL P60@
     ~LEG TUBE 60 1 -RED :LEG
will resize using only the tube shapes in the table which are 60 inches in OD.


Q: As per API joint check criteria, the Fy of the chord should be 2/3 of the tensile strength or the Fy of the member, which ever is lower. Does MOSES take care of this or do I have to manually change the Fy of the chord member to whichever is the lower value?
REV 5.08

A: Of course MOSES treats this correctly. Be sure, however, that you have defined the tensile strength correctly. The default is to use a fraction of the yield. We know that this is not correct, but it prevents bogus joint checks for 50 ksi material when people have not altered the tensile strength from the default.


Q: Why is it when I do my code check, there are 3 member reported to be overstressed, but when I do the pictures for member IR > 1.0, the plots show more than 3?
REV 5.08

A: Actually, your statements are not correct! According to your data you asked for:

beam code -summary 1 1e30 0 1
The manual clearly states that what you get here is the largest code check for each class. Thus, your output says that there are three classes overstressed (not members). Your pictures are telling you that there is more that one member failure for at least some of the overstressed classes.


Q: How can I determine the deflection of the barge in a certain environment?
REV 5.08

A: You can use the JOINT DISPL command from STRPOST menu. However, to get something that makes sense you need to look at deterministic load cases, not spectral ones. To generate deterministic cases, you could use either the LCASE -TIME, LCASE -PROCESS, or CASES -TIME command.


Q: There are supposed to be three unity ratios reported in the Hydrostatic Collapse Check. Why in some cases is one of them missing?
REV 5.07

A: The code has three sections: hoop stress, tension, and compression. For some cases, the can be no tension (or compression). In other words, the axial load in the beam is such that the maximum normal stress (axial stress +- bending stress) is always of the same sign. In this case one of the check cannot be made and we print no results.


Q: Which stresses are considered for the fatigue analysis of structural components?
REV 5.07

A: For beams we use the normal stress and for plates we use Von Mises stress.


Q: How can I get zero for the allowable punching shear?
REV 5.03

A: In some cases with high chord stress, Qf can become negative. When this happens, vpa is set to zero.


Q: What is "Mom AMF" in the beam check summary output?
REV 5.03

A: This is the bending moment modifier, in a slightly more usable format:

Mom AMF = 1/[Cm/(1-fa/F'ey)].