PhD-Level Ultrasonic Engineering • Medical • Manufacturing • Consumer
What FEA program should you use to simulate power ultrasonic devices?
This is the first post in a series describing the common mistakes and misconceptions that engineers make when simulating an ultrasonic transducer in finite element software.
Do you really need piezoelectric physics?
The first mistake they make is that they assume they need to use piezoelectric multiphysics modules in order to simulate those devices. This may at first look obvious - if you want to simulate a piezoelectric device, you should be able to model those physics. This is fundamentally wrong assumption because simple standard elastic material physics (standard in any FEA program, expensive or inexpensive) are sufficient to determine many key performance characteristics of ultrasonic devices.
Resonance Simulation
Firstly, take the resonance frequency and mode shape (displacement profile, nodes, antinodes) - modal analysis solves for the resonant modes and their frequencies. You do not need to perform harmonic simulations in order to determine the resonance frequency and the displacement distribution, nor do you need to include piezoelectric phyiscs in 95% of cases.
Static Simulations
Secondly, to perform static simulations to determine displacement without piezoelectric physics, you can apply forcing on the faces of the piezoelectric material according to the piezo crystal's own blocking force. Yes, you will have to implement a hand calculation in order to determine what force you need to apply, but the process is not complex!
What could go wrong?
What happens when you use piezoelectric physics when you don't need to is that you will (1) over-complicate the problem and (2) overstretch your goals for the simulation.
When the FEA modeling gets more complex, there is additional time spent in perfecting simulations and running optimizations. Understand that your FEA model is in fact a
MODEL not
THE REAL THING. There are many differences which occur in a real prototype that cannot be reliably predicted with FEA. The two largest phenomena that come to mind are non-linearity and component interface effects.
Imagine you are going on a vacation. You have to drive there. The "car ride" is the FEA modeling step and the "destination" is your prototype.
The Simple Principle
So, should you use piezoelectric physics in your modeling approach? If you don't know or cannot explain concepts in piezoelectricity, you should not. In that case, you should focus on simulations focusing on elastic material physics only, and continue your studies on piezoelectricity. A plus side of this approach is that you will have simpler simulations, an emphasis on prototypes, and also the ability to use lower cost FEA programs and modules.