Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Immobile piezo
Posted 6 ott 2010, 04:55 GMT-4 2 Replies
Please login with a confirmed email address before reporting spam
Hello everyone,
I´m new to COMSOL (using a trial version right now) and I like it a lot so far. However, after completing a couple of tutorials I tried to set up a model for my research, similar to the "Piezoacoustic transducer" model in the model library (Acoustics-> Tutorials). In short, I try to do a transient analysis of a piezoelectrical device, that excites an acoustic wave in the adjacent medium. I had some trouble with my model, so I boiled it down to a relatively simple model, a 2D piezoelectric rectangle, that I try to excite via a sinusoidal voltage.
When I build it as a 2D axissymmetrical geometry, I can see the deformation of the piezo over time (see Piezo_Symmetrical), although it´s tiny, so it might just be a computational error. But when I do the exact same thing for the whole model (building it as a normal 2D model without using axial symmetry) and run it, there is no deformation at all (see Piezo_Whole). What am I missing?
I asked the comsol support about this, but haven´t heard back from them in almost a week. It would be nice, if someone could have a look at my models and point me in the right direction,
Alex
I´m new to COMSOL (using a trial version right now) and I like it a lot so far. However, after completing a couple of tutorials I tried to set up a model for my research, similar to the "Piezoacoustic transducer" model in the model library (Acoustics-> Tutorials). In short, I try to do a transient analysis of a piezoelectrical device, that excites an acoustic wave in the adjacent medium. I had some trouble with my model, so I boiled it down to a relatively simple model, a 2D piezoelectric rectangle, that I try to excite via a sinusoidal voltage.
When I build it as a 2D axissymmetrical geometry, I can see the deformation of the piezo over time (see Piezo_Symmetrical), although it´s tiny, so it might just be a computational error. But when I do the exact same thing for the whole model (building it as a normal 2D model without using axial symmetry) and run it, there is no deformation at all (see Piezo_Whole). What am I missing?
I asked the comsol support about this, but haven´t heard back from them in almost a week. It would be nice, if someone could have a look at my models and point me in the right direction,
Alex
Attachments:
2 Replies Last Post 6 ott 2010, 10:02 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
one thing , to get consistent units typu in for your Voltage
V0*wv1(t[1/s])
the t[1/s] transforms the time t into seconds and make it without units, currently functions like sinus want unitless varaibles . This turns off the unit warning (orange color). If you leave the mouse pointer some seconds on the orange equation, you will seea popo-up message giving you some clues on what is wrong. Orange = unit error, its just a warning, it will calculate, but the results might be wrong. While red is typing error that must be corrected to solve your model.
Your 2D axi is solved in time domain (I would have used modal harmonic since you have a sinus excitation it's much quicker to solve) but you leave the time stepper free, this means that COMSOl will try to find a gradient and the time values and will decide when it calcualtes making a gross undersampling of your 200 [1/s] oscilaltion (?)
Have you tried to plot your curve, you will see that the sinus is in fact wv1=sin(200*t) which means
f=200/2/pi=31.8 Hz
In the time stepper try to use "strickt" or "intermediate" stepping and not "Free" then you would probably see better your oscillation
Finally for a 1m large block, 100V/m is not really a large field to excite a PZT so do expect very small displacements
====
For the full model, I do not have time to look it up just now, perhaps later today
--
Good luck
Ivar
one thing , to get consistent units typu in for your Voltage
V0*wv1(t[1/s])
the t[1/s] transforms the time t into seconds and make it without units, currently functions like sinus want unitless varaibles . This turns off the unit warning (orange color). If you leave the mouse pointer some seconds on the orange equation, you will seea popo-up message giving you some clues on what is wrong. Orange = unit error, its just a warning, it will calculate, but the results might be wrong. While red is typing error that must be corrected to solve your model.
Your 2D axi is solved in time domain (I would have used modal harmonic since you have a sinus excitation it's much quicker to solve) but you leave the time stepper free, this means that COMSOl will try to find a gradient and the time values and will decide when it calcualtes making a gross undersampling of your 200 [1/s] oscilaltion (?)
Have you tried to plot your curve, you will see that the sinus is in fact wv1=sin(200*t) which means
f=200/2/pi=31.8 Hz
In the time stepper try to use "strickt" or "intermediate" stepping and not "Free" then you would probably see better your oscillation
Finally for a 1m large block, 100V/m is not really a large field to excite a PZT so do expect very small displacements
====
For the full model, I do not have time to look it up just now, perhaps later today
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
first of all thank you for your quick reply. As I´m just starting to use COMSOL little hints like the one with the units are very much appreachiated.
Yes, I plotted the curve...basically, I´m not looking for a specific oscillation frequency, I chose omega=200 1/s because it oscillates like 4 times in the given time intervall, so that lets me see, if the voltage oscillation is as I expect it.
I`m playing around with the time stepper to get some experience with it...but just leaving it at the default values seemed to be the quickest way to do it.
Yeah, I missed the scale for my model, you are right. 1m is a bit oversized hehe...
Again thank you for your input, it would be great if you could look at the whole model,
All the best,
Alex
first of all thank you for your quick reply. As I´m just starting to use COMSOL little hints like the one with the units are very much appreachiated.
Yes, I plotted the curve...basically, I´m not looking for a specific oscillation frequency, I chose omega=200 1/s because it oscillates like 4 times in the given time intervall, so that lets me see, if the voltage oscillation is as I expect it.
I`m playing around with the time stepper to get some experience with it...but just leaving it at the default values seemed to be the quickest way to do it.
Yeah, I missed the scale for my model, you are right. 1m is a bit oversized hehe...
Again thank you for your input, it would be great if you could look at the whole model,
All the best,
Alex