Discussion Forum

Hi

there are a few discussion on the forum on applying acceleration fileds = gravity load (try the search). But basically you apply a volumic force F[N/m^3]=rho_smsld[kg/m^3]*G (for _smsld application mode, pls correct if you use another) and G is defined as a "constant", I use G = 1[lbf/lb] a shortuct to write out 9.8065...[m/s^2]

Have fun Comsoling
Ivar

Thank you, Ivar.

I know how to apply a load. I also see that this works fine in time-domain settings. But how about evaluating the frequency response? I thought that any load is then understood as harmonic, as amplitude of a harmonic function. I don't see the pendulum resonsance in the response. I have tried the load already.

Hi

you are correct: in harmonic, the loads are the harmonic amplitude at the frequency you are "parametrically" scanning, pls check that the application mode AND the solver are set to parametric and that you have defined the correct frequency variable name as parameter name

normally it should work, perhaps check again w.r.t. a doc example, as its easy to miss a simple step

Have fun Comsoling
Ivar

Ok, but now back to my problem. I need a harmonic lateral load, and a static gravity load to simulate the frequency response of a pendulum. Is that possible? So what is the method to define static loads in frequency-response problems?

I have checked my code. I see all kinds of resonances (the pendulum has torsion, the suspended mass tilts, etc), but no sign of the pendulum resonance. So gravity is not active if you define it as load, but this is clear somehow since loads are harmonic in frequency-response problems, and gravity acts statically on the system. So I am still looking for a way to study the frequency-response of a pendulum. Since I cannot define gravity as load, is there a trick, another type of constraint to add constant acceleration?

Hi

you are right that in "harmonic" mode all predefined gui forces are set to harmonic (phase amplitude only is defined), a static acceleration field must be applied by hand (probably easiest as a weak form), I need to find back my exercice from a few years ago, indeed a long time since I last applied a seismic load response, because for that, just as for space "launch loads", you cannot really "turn off" gravity, this is neither not "physical" something that COMSOL people had forgotten, but we are all so used to gravity, that normaly we do not ask ourselves why we sit on a chair, and are not flaoting around, as in the ISS ;)

Have fun Comsoling
Ivar

Thank you. Ok, so you say that there should be a way to add gravity "by hand". That would be great. I am not yet familiar with the "weak form", but I will see what I can find.

Ok, I got my pendulum resonance by adding the weak term -rho_smsld*g/L*u*u_test to the subdomain (g is surface acceleration, and L is length of pendulum). But this is just the main effect of gravity, i.e. the restoring force of the pendulum mode for small displacements. Now there is one more effect that I need to add by hand. I need a static tension on the suspenion of the pendulum since the tension alters the speed of violin modes of the suspension string. Any idea how to add static stress to get this right as well?

Hi

just to say your case is not that trivial, it made me rediscover a few new effects, indeed in 2D I can get logical results for a pendulum (but I must define the two forces Fx-rho*G*sin(th) and Fy=-rho*G*cos(th) with th=0.5*(uy-vx) for 2D if not already defined by your application mode, you can extrapolate to 3D) for all cases from static to quasi-static transient, but except for frequency response, and I still do not understand why. I have myself two models that are (have been for some time) stuck on this, I'm heating them up again, and pushing support, together we should get there, no ?

I have attached my pendule case as an Euler beam with an extra mass at its end, but Frequency response is failing

Have fun Comsoling
Ivar

Thanks Ivar for your help with this problem. It seems to me that this shouldn't be a very exoctic problem. I can think of many phyical systems where frequency response depends on a static force like gravity. I will keep working on this. Maybe the right way to go is to start a mathematical treatment to find out in general how static forces can change frequency response. Maybe there is a systematic way to incorporate static forces. My current method to add specific effects of gravity by hand (pendulum restoring force, string tension,...) is not very elegant, and in the end I am not sure if I took all relevant effects into account.