Midas Questions

Question:

Good afternoon,

It is vital that we achieve at least 90% mass participation in our model. We currently reach a maximum of 86% with hundreds of mode shapes. Can MIDAS advise us on ways to mobilize more mass for our particular model?

In the Eigenvalue Mass Summary Table, we noticed some mode shapes contained 0% mass participation, yet the structure had experienced a deflection for these mode shapes. For mode shape 100, the maximum structure deflection was .57 inches; although small, this deflection indicates the mass was engaged and should therefore register mass participation in the Mass Summary Table. How can the table contain 0% mass participation for this mode shape? We also noticed on previous versions of our model, where we unintentionally had certain areas with D.O.F. left unconstrained, the summation of the mass participation was over 100% in the Mass Summary Table. Please explain how the summary table for masses sums the mass participation? Does it add mass participation from the mode shapes in a way that could add up to greater than 100% mass?

Our compression model uses rigid links (non-elastic) to model our bearings at the FIXED bents to fix all degrees of freedom, except rotation about the bent caps. Since the model has multiple curves and the rigid links are oriented in the global coordinate system, we chose to release these rigid links in the RX and RY directions. For the EXPANSION bents, we utilized rigid elastic links with a very high stiffness in each direction to model the bearings, and 0 stiffness in the SRx and SRy directions. Can MIDAS comment on this nodal linkage arrangement? At the expansion bents, should we use rigid master-slave links?
To model our pile footings, we used one multilinear spring at each pile to represent the soil stiffness at that pile. We are aware that MIDAS will convert a multilinear spring to a linear spring during a dynamic analysis. How is the elastic stiffness calculated for the multilinear spring when it is automatically converted to a linear spring by Midas for dynamic analysis?

Thank you in advance for your help with these questions,

Answer:
Hi,

1. If there are many small regions which vibrate partially in the model, with Eigen value analysis, you should try larger number of frequencies as program finds every mode which causes vibration. So, please try number of frequencies approximately until 400 to achieve 90% mass participation.
Or you might use Riz Vectors. Using Riz vectors type is easier to achieve 90% mass participation. Define starting load vectors adding ground acceleration for each axis like the figure below.

2. Even you viewed 0% mass participation, it might not be exactly zero. It might have tiny value under 0.0001 for example.
For case over 100% mass participation, we need the relevant model file so that we can look into the problem. Or if there was any error message after performing analysis, you should not believe any results.
3. I recommend to using rigid type of elastic link. You adjust angle of links inputting value of Beta angle at the elastic link table.

4. When midas Civil converts a multilinear spring to a linear spring, the elastic stiffness calculated based on initial stiffness of multilinear spring.

Thanks & Regards,
Jihoon Kang
Creation date: 10/23/2017 12:49 AM      Updated: 10/23/2017 10:20 AM
Files   
image002.png
18 KB
image003.png
56 KB