Problem in Newmark method Implementation

For developers writing C++, Fortran, Java, code who have questions or comments to make.

Moderators: silvia, selimgunay, Moderators

Post Reply
debasismpt
Posts: 10
Joined: Mon May 13, 2013 7:48 am
Location: IIT ROORKEE

Problem in Newmark method Implementation

Post by debasismpt »

i am working on a project PML(perfectly matched layer) modelling of unbounded soil domain.For implementing Newmark method in this formulation expressions are generated like given below
tangent stiffness
k=c1(ke+kt)+c2(ce+ct)+c3(me)
effective internal force
p(n+1)=(me) a(n+1)+(ce)v(n+1)+(ke)d(n+1)+Pi(n+1)
where a(n+1),v(n+1),d(n+1) are acceleration,velocity,displacement at n+1th step respectively.Pi(n+1) is internal force at n+1th step.Calculation of all these values are not a problem. value of ke,kt,ce,ct,me also can be calculated.
Main problem here is i have to give different stiffness and damping matrix for tangent stiffness calculation.How can i do this :?: Element Abstract class allows only one getstiffness(),getmass() and getdamp() functions.Is there any way i can solve this problem by adding new element class or new integrator class :?: If u have any suggestion Plz help me.
fmk
Site Admin
Posts: 5884
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Re: Problem in Newmark method Implementation

Post by fmk »

have your element return ke+kt in the getStifneess() call, likewise for the getMass().
Hsucycle
Posts: 1
Joined: Tue Mar 11, 2014 3:44 am

Re: Problem in Newmark method Implementation

Post by Hsucycle »

I am also working on a 2-D SDOF dynamic problem with only one zerolength element subjected to a point load P(t). The material has bilinear stiffness, k_elastic and k_plastic.

The main problem here is that the mass of this system is not constant, which is to say I have to calculate equivalent mass. The “equivalent mass" here is equal to a factor, a1, multiples the mass, when the stiffness in the elastic range; when the stiffness in the plastic range, The “equivalent mass” is equal to a factor, a2, multiples the mass.

I am wondering how to do that? Is my issue the same from debasismpt?
fmk
Site Admin
Posts: 5884
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Re: Problem in Newmark method Implementation

Post by fmk »

as you have only 1 zerolength element and it is a 2 dof problem you can obtain the tangent stiffnesses from the material after every step and then reissue a mass command to set the mass.

i.e.
while ($tCurrent < $tFinal) {
analyze 1 $dt
set currentTangent [eleResponse 1 material $matNum tangent]
if {$tangent ???} {
mass 1 ????
} else {
mass 1 ????
}
set tCurrent [getTime]
}
abbasp
Posts: 6
Joined: Thu Jul 31, 2014 11:21 pm
Location: International Institute of Earthquake Engineering and Seismology (IIEES)

Re: Problem in Newmark method Implementation

Post by abbasp »

debasismpt wrote:
> i am working on a project PML(perfectly matched layer) modelling of
> unbounded soil domain.For implementing Newmark method in this formulation
> expressions are generated like given below
> tangent stiffness
> k=c1(ke+kt)+c2(ce+ct)+c3(me)
> effective internal force
> p(n+1)=(me) a(n+1)+(ce)v(n+1)+(ke)d(n+1)+Pi(n+1)
> where a(n+1),v(n+1),d(n+1) are acceleration,velocity,displacement at n+1th
> step respectively.Pi(n+1) is internal force at n+1th step.Calculation of
> all these values are not a problem. value of ke,kt,ce,ct,me also can be
> calculated.
> Main problem here is i have to give different stiffness and damping
> matrix for tangent stiffness calculation.How can i do this :?: Element
> Abstract class allows only one getstiffness(),getmass() and getdamp()
> functions.Is there any way i can solve this problem by adding new element
> class or new integrator class :?: If u have any suggestion Plz help me.

Dear debasismpt,
Have you finished implementing PML into openSEES successfully? Could you please make me aware of the process?
Post Reply