COMSOL simulation of frequency-impedance analysis of a piezoelectric element

[@ultrasonicadvisors]

Sign up for my free 5-day course on inputting correct PIEZOELECTRIC material properties into your FEA simulation program: ultrasonicadvisors.com/fea-piezoprop

[@aritradas9493]

Did you change the model, because the mesh model and simulated model are not the same as can be seen from 10:09 and 10:16. If do please share the same.

[@ultrasonicadvisors]

I think they are the same.

[@thecrazykartbyAJ]

Nice video! At 4:10, I dont get why you use the QF of the PZT-5 (QF=500) instead of the QF of 400, which belongs to the material you are using (PZT-4). Thank you!

[@thecrazykartbyAJ]

@@ultrasonicadvisors thank you Husain!

[@ultrasonicadvisors]

The properties were listed for demonstration.

[@sarrasalhi2360]

Thank you Sir ,but what shall we do if we use the physics layered shells?

[@ultrasonicadvisors]

I've not done that

[@abdm.4426]

Thank you very much for the video. Could you please explain why you introduced damping for the Linear elastic materials not for the piezo itself, another question if I have coupling factor of k=0.1 how I can introduce to Comsol , should be a tensor or not please ?

[@ultrasonicadvisors]

Yes. Piezoelectric materials actually have anisotropic loss factors. However, for simpliciation purposes (and for most cases), we can just use a single loss factor. I used a loss factor so that the resonance impedance would not go to 0 and the antiresonance impedance would not go to infinity. You cannot put coupling factor direclty into comsol. You need to convert it into d, epsilon, and s

[@alkistis-mariafotopoulou622]

Why do you multiply by 0.005 [m] for the boundary probe expression?

[@ultrasonicadvisors]

This accounts for the "in-the-screen" or width dimension of the piezo. Questions?

[@parthisl4457]

Sir , if we applied ultrasound wave( 1MHz) to PZT disc , how much voltage generates accross the PZT disc. Can you share COMSOL simulation lecture video for above condition. Normally we applied force in Newton. But I need force in frequency component. In Medical equipment detect ultrasound wave ( for an example ‘ Blood pressure measurement using Piezoelectric effect by an ultrasonic probe.

[@ultrasonicadvisors]

You need to do acoustic simulations. The voltage will be in the order of mV

[@zhangyaoyao1902]

Thanks for the video. It is very helpful for my piezo simulation. Just one part I couldn't understand. In the impedance definition, |Z| = abs(Z) and Theta = arctan(Z"/Z'). According to complex calculation, because V is always real, abs(V/I) = V/abs(I), the magnitude is correct. However, in the phase angle calculation, Z' = real(Z) = real(V/I), which is not equal to V/real(I), no matter whether V is real or complex. In other words, Z"/Z' != I'/I". Confusing part is, according to your method, the curve looks correct and well matching with that I got from analyzer, although it seems not matching with impedance definition. And if I tried to pull out real(V/I) and imag(V/I), it looks reverse too. Not sure if you are still looking at this post, but if you see my reply and want to do some further discussion, please email me or just reply here. I must be something wrong somewhere, maybe my understanding on relations between impedance and current.

[@zhangyaoyao1902]

@@ultrasonicadvisors Hi. Thanks for the reply. I think I found the right solution on this and maybe you could give a try. -es.JZ In you boundary probe you could replace es.nD*es.omega into -es.JZ, which is the Z-orientation current density integral in XY plane. Minus sign is to correct the current direction into upward. Of course when plate direction changes, the component JZ may change to combination of (JX, JY, JZ). es.nD is electric displacement, which is derived from dielectric constant ε, or capacitance, so it naturally has a 90 deg phase difference from current I, when both using voltage's phase as reference. This is why in your video you have to use a wired formula to achieve a plausible result. You could try point 1D plot on atan(imag(-V/es.JZ)/real(-V/es.JZ)) and it will give the right phase angle curve (same with yours, but this fully matches with impedance definition). Since this video is the top search to this topic, I would like to add this comment so in the future people won't bother with the same question as I did.

[@alkistis-mariafotopoulou622]

Is the voltage you are applying AC or DC?

[@ultrasonicadvisors]

AC

[@mdasimalam893]

Very helpful video sir

[@ultrasonicadvisors]

Thanks for the comment!

[@amirkalafi1096]

Thanks sir. Please model piezoelectric of humidifier in 2.4 MH

[@yangjiechen6355]

Thanks for the nice video, I followed the video but the bnd1 I got is real numbers rather than imaginary numbers as you mentioned. I think someone else got the same result as I get.

[@davidegalletta6321]

I think the problem is in the damping. Right now, the damping is applied on the linear material, but then overwritten by the piezoelectric. Insert a mechanical damping on the piezoelectric material and you will see real and imaginary part. Also, instead of defining the current as es.nD*es.omega, I would suggest to use -es.JZ (as someone else suggested in the comments).

[@hyunjungmin9295]

Thank you for sharing the video. Would you mind showing 3D shaped one?

[@ultrasonicadvisors]

Sorry, don't have access to comsol anymore

[@nicolas3360]

Hi, thanks for your video. Could you please let a link to your model including the final curves of electrical impedance. Best regards, Nicolas

[@nicolas3360]

I did not get the same impedance curves by reproducing your model. Il fact, the current that I calculate is purely real, with no imaginary part. It might be because I did not put any damping in the piezoelectric material. Also, I got the resonance but not the anti-resonance.

[@elangovanyedeshkumary242]

@OhRum Cha Me too... I dont get the same impedance cuve

[@swatantrasingh9280]

@@nicolas3360 have you got the solution of this

[@davidegalletta6321]

excuse me, did someone of you managed to solve this issue? Thanks in advance!

[@apiyopao4586]

How to model the same in 3D instead of 2D?

[@apiyopao4586]

@@ultrasonicadvisors Thankyou😊 for the quick response Actually iam having problem in plotting the frequency domain analysis coz when I plot the voltage Vs frequency graph all I get is an empty plot with a dot or straight line in the centre Iam really struggling through it I have seen other mph. files and followed the steps exactly the same still Iam getting empty plots Please help🙏🙏🙏

[@apiyopao4586]

@@ultrasonicadvisors Sir,iam doing the frequency domain analysis again and this time it's taking like hours, literally forever for the study computation The simulation is still in progress I mean is it okay for a computation to take that long in COMSOL How would I know if there's a problem in my design or BC settings cos not even error is popping up, it is running endlessly

[@apiyopao4586]

@@ultrasonicadvisors Transient analysis It's been 2 days now and the simulation is still running how long is it going to take?

[@apiyopao4586]

@@ultrasonicadvisors But why sir?

[@apiyopao4586]

@@ultrasonicadvisors Sir I have got another query While plotting the Voltage iam writing the expression es.V0_1 (as iam using the electrostatics module) with the unit V for the y-axis data But when I click 'plot' i get "failed to evaluate the variable" Could you tell me what's wrong in my expression sir?

[@alak3073]

thank you for the great videos. I watched the video many times trying to find the current value from 3d module(the same in lecture 13 part B) I was able to find voltage like what you did in the previous video but i could not calculate the current in 3d module. could I send you my mph file or if you could not help me with the file could you tell me how to calculate the current from equation or any help

[@zhangyaoyao1902]

I'm doing a lot of 3D simulations on Piezos. You could contact me too if you need any private help. [e-mail]

[@swatantrasingh9280]

@@zhangyaoyao1902 Can you tell me why I am not getting imaginary part of current in comsol