Incident angle of ground motion

[Muntasir]

Hi,

Is it possible to define the incident angle of ground motion in seismostruct?

Thanks.

[huffte]

I believe that this may be only be mimicked by applying the loads in appropriate proportion to represent different angles of incident motion.

For example, to mimic a 45 degree angle of insicence of a particular record, one could apply 0.707 times the record in each of the x and y directions simultaneously to the base nodes.

[Muntasir]

Dear Hufte,

Thanks for the reply. But how accurate this procedure is? have you tried this method?

Thanks.

[huffte]

Muntasir, I have not actually tried the method I mentioned above and upon further thought, I believe there are better options.

If you have a record pair you wish to apply - for example, EW in the x-direction and NS in the y-direction, or fault normal and fault parallel - you can actually rotate the motions themselves (http://peer.berkeley.edu/products/nga_s ... 02006).pdf) as follows:

Let x and y be the as-recorded pair of motions and let t be the angle to rotate to axes u and v:

u = x(cost)+y(sint)
v = -x(sint)+y(cost)

This essentially reduces to my previous suggestion if your load is a single record instead of an orthogonal record pair.

Yet a third option - which may be the best, depending upon the complexity of your model - is to actually rotate the model through a coordinate transformation and still apply the records in the x and y directions.

[Muntasir]

Thanks a lot Huffte for your help.