Updated mesh analysis

Hi all, 

  and others 

I have found this in the plaxis 2D reference manual.

My questions are 

1. Why when normal analysis followed by the updated mesh analysis, reset displacement zero is needed to be provided?

2. Any old webinar about this updated mesh and updated pore pressure?

Thanks.

Parents
  • Hi,

    Let met try to answer all questions in one go Slight smile

    1) In a normal stress analysis the strain increments in the soil are based on the deformation relative to the undeformed mesh and in a updated mesh analysis the strain increments are calculated based on the deformation relative to the previous step. So those 2 methods will give slightly different strains and with the stiffnesses thus slightly different stress increments. If we would change from normal analysis to updated mesh analysis we would suddenly change the way strains are calculated and thus get a sudden change in stresses as well. If displacements (and thus strains) are reset, the calculation will start from zero strains and this problem doesn't exist.

    To give an example:  let's say we have a 1.0 m long beam and we will in Phase 1 extend it to 1.1 meter, and then in Phase 2 to 1.2 m.

    Now a normal analysis will say the strain increment in Phase 1 is  (1.1 - 1.0)/1.0  = 10% and in Phase 2 it is (1.2 - 1.1)/1.0 = 10% again, so the total strain is 20%.

    An updated mesh analysis will say the strain increment in Phase 1 is  (1.1 - 1.0)/1.0  = 10% and in Phase 2 it is (1.2 - 1.1)/1.1 = 9.1%  so the total strain is 19.1%.

    So both analysis will have different stresses in the beam based on different strain levels. Changing between the two types of analysis may give strange results.

    2)  4) I don't think we've ever done a webinar on updated mesh and updated pore pressures.

    3)  The phase displacement are not necessarily exactly the same. If you would use a model with stress dependent stiffness (for instance Hardening Soil) then the two methods give different strains, thus different stresses and thus different stiffnesses. And because of the different stiffnesses the displacements will be different. In general for the Mohr-Coulomb the displacements will be the same as Mohr-Coulomb has a constant stiffness.

    Kind regards,

    Dennis Waterman

  • Dear Mr. Dennis,

    Thanks for the great explanations. i need some clarifications in your answer please.

    1. How it is said that high deformation needed updated analysis? Which is correct way of analysis normal or updated? updated pore pressures doesn't change the user defined water levels, so when we do excavation problems we defined the water levels at that time how updated water pressure option is functioning?

    2. As per the answer, in MC, both deformed and phase displacement are same, then we can do only normal analysis without updated mesh is ok? and updated mesh analysis can't produce cumulative deformation "or" if we start the analysis with updated mesh and following phases if we don't check the reset displacement zero will it produce accurate and cumulative deformations (because i got to know from the reference manual that updated mesh analysis starts from un-deformed geometry)?

    Thanks.

  • Hi,

    No problem, I will do my best to make it clear.

    1) Both methods are in principle correct, but with updated mesh analysis it's possible to calculate higher levels of deformation and also with higher accuracy. However, updated mesh analysis is generally slower and numerically less stable and therefore it's not the default analysis in PLAXIS.  If strain levels remain small (let's say under 10%) there is almost no difference in results. So Updated Mesh is generally used for dams and embankments on soft soils as strain levels are large. For for instance excavation problems there is no advantage in using Updated Mesh.
    The Updated Pore Pressures (UPP) doesn't update phreatic levels, it updates the pore pressures in the mesh according to the deformed mesh. Let's say we take a point in the soil that is initially 10cm above the water level, which means that in that point there are no pore pressures en in a normal analysis the pore pressures are determined according to the undeformed mesh so there will never be pore pressures unless the phreatic level changes. But if in reality that point gets a settlement of 20cm it will then be 10cm below the phreatic level and should have a 1kPa pore pressure but the normal analysis does not take that into account.  If you're using UM + UPP analysis the calculation will every calculation step check the updated location of the point and calculate the new pore pressures....so the point will indeed change from dry to a 1 kPa pore pressure if it settles from 10 cm above the water level to 10 cm below the water level.

    2) Please note that for MC using UM doesn't make a difference for the soil stresses, but there are more effects that in the end can make a difference if strain levels get quite high. For instance second order effects in plates and walls:  the are initially straight but get curved during the calculation and with UM the curvature is taken into account to calculate for instance bending moments. However, these effect will start becoming noticeable for larger strain levels, let's say starting from 10-30%,
    In short: Updated Mesh really moves the nodes in the mesh every calculation step to the new deformed location and then continues to calculate from there. A normal  analysis will always calculate deformation relative to the undeformed (initial) mesh. In principle the user shouldn't mix those 2: either calculate the whole project with updated mesh, or the whole project with normal calculation phases.  Changing from updated mesh to normal or vice versa will give a mismatch in strains that the calculation either has to solve first (which will give an adjustment of stresses which may in case of an excavation for instance give a spike in structural forces) or the user has to reset the displacements to avoid the mismatch....but then of course you will loose the entire displacement history.

    Best regards,

    Dennis Waterman

    Answer Verified By: Nitharshan 

Reply
  • Hi,

    No problem, I will do my best to make it clear.

    1) Both methods are in principle correct, but with updated mesh analysis it's possible to calculate higher levels of deformation and also with higher accuracy. However, updated mesh analysis is generally slower and numerically less stable and therefore it's not the default analysis in PLAXIS.  If strain levels remain small (let's say under 10%) there is almost no difference in results. So Updated Mesh is generally used for dams and embankments on soft soils as strain levels are large. For for instance excavation problems there is no advantage in using Updated Mesh.
    The Updated Pore Pressures (UPP) doesn't update phreatic levels, it updates the pore pressures in the mesh according to the deformed mesh. Let's say we take a point in the soil that is initially 10cm above the water level, which means that in that point there are no pore pressures en in a normal analysis the pore pressures are determined according to the undeformed mesh so there will never be pore pressures unless the phreatic level changes. But if in reality that point gets a settlement of 20cm it will then be 10cm below the phreatic level and should have a 1kPa pore pressure but the normal analysis does not take that into account.  If you're using UM + UPP analysis the calculation will every calculation step check the updated location of the point and calculate the new pore pressures....so the point will indeed change from dry to a 1 kPa pore pressure if it settles from 10 cm above the water level to 10 cm below the water level.

    2) Please note that for MC using UM doesn't make a difference for the soil stresses, but there are more effects that in the end can make a difference if strain levels get quite high. For instance second order effects in plates and walls:  the are initially straight but get curved during the calculation and with UM the curvature is taken into account to calculate for instance bending moments. However, these effect will start becoming noticeable for larger strain levels, let's say starting from 10-30%,
    In short: Updated Mesh really moves the nodes in the mesh every calculation step to the new deformed location and then continues to calculate from there. A normal  analysis will always calculate deformation relative to the undeformed (initial) mesh. In principle the user shouldn't mix those 2: either calculate the whole project with updated mesh, or the whole project with normal calculation phases.  Changing from updated mesh to normal or vice versa will give a mismatch in strains that the calculation either has to solve first (which will give an adjustment of stresses which may in case of an excavation for instance give a spike in structural forces) or the user has to reset the displacements to avoid the mismatch....but then of course you will loose the entire displacement history.

    Best regards,

    Dennis Waterman

    Answer Verified By: Nitharshan 

Children