feedback control of resonant frequency and optimum pressure

[Perfeito]

Hi all!

I am presently working on this project where i used an ultrasound device to induce vibrations inside a fluid filled beaker. I excite the ultrasound device with a specific resonant frequency and the output is a pressure signal at the pressure sensor. Here the optimumk pressure that is to be reached does not exactly depend on the excitation frequency, It is rather around it. So i need an adaptive control probably (PLL or extremum seeking control or hill climbing methods) I was going through a few research articles and found myself confused on which method better suits my application. 

I have developed a LabVIEW model using DAQ mx interface and got up till resonance tracking and control. 

My question is, should i follow the method of impedance calculation and phase change i.e PLL or should i go with extremum seeking control?

Regards,

Vipul. V

[johnsold]

Which method is appropriate depends on what you want to do and how well you can describe the system behavior.

You indicate that you want to reach an "optimum pressure" but that pressure does not occur exactly at resonance. Do you have a mathematical relationship between the frequency for optimum pressure and the resonant frequency? Is the physics/fluid mechanics of your system well understood? Is there a good model for the behavior? Can you detect the optimum pressure directly or through some indirect method?

A PLL is best when trying to lock onto an existing signal. If you have a specific offset from the reference frequency, PLL circuits can be designed to track that also. From your description (which I do not fully understand) I suspect that a PLL is not the best choice.

Lynn 

[Perfeito]

Dear Lynn!

The existing system is a complex system. 

Yes there is a mathematical relationship between the excitation frequency and optimum pressure. I haved attached a block diagram that could possibly be easy for you to underestand. In the process I excite the system by running white noise, i then choose the excitation frequency from the resulting spectrum. Then i exite the system by running the sin signal with the recorded excitation frequency. Resulting pressure in the system is then obtained by a pressure sensor and then i take the FFT to obtain the peak Sound pressure Level. This whole process is done with the CLIO setup where we need to excite the system with each single frequency to obtain the optimum..

So i was looking for a method where we could, after FFT continuously detect the peak and stabilize it. This actually is more related to an example that i have attached. Please find the attachment and i hope it would be a little easy to guide me.

Regards,

Vipul.V  

[johnsold]

Vipul.V,

From the time of your post it appears that we may be on opposite sides of the world.

I glanced quickly at the documents you posted.  I think I understand what you are trying to do.  I will read through the Kuang paper and see what I can make of it. It will probably be later today or sometime tomorrow before I post back.

Lynn

[johnsold]

Vipul.V,

I finally had some time to read through the paper by Kuang. 

I now have several questions:

1. Do you have equipment to measure the voltage and current applied to the ultrasound transducer, or do you just have pressure measurements?

2. How are you generating or controlling the excitation frequency and amplitude? Do you have a function generator as Kuang used or are you using some other method?

3. Will you control only the frequency or will you also control the amplitude?

4. What is the nomnal resonant frequency of your transducer?

5. How fast does the feedback loop need to operate?

6. What is the duration of a session? Or, how long does it need to keep everything stable?

7. Does the amplitude setpoint change during a session?

8. What are the consequences if the system fails to regulate properly? Does equipment get destroyed? Can people be injured or killed? Does it cost a lot of money?

9. What computer, OS, software, and DAQ hardware are you using?

Lynn

[johnsold]

Do you get a good estimate of the resonance by exciting with noise?  While noise has a flat spectrum, the amount of energy at any one frequency is quite small. A chirp signal or swept sine wave has a similar spectrum when averaged over time. It has the advantage of having most of the energy at one frequency (or in a small band of frequencies) at a particular time. This can produce a much larger response when searching for the resonance.

Lynn

[Perfeito]

Dear Lynn!

I am really sorry for the delay with  my reply. I was busy with my office work so didnt have much time to give you a detailed reply.

Regarding your last reply you asked me a few questions to which the answers are given as following.

1) I have the setup to measure the voltage and current along with the pressure measurements, I measure the voltage but i measure the current interms of voltage using a resistor.

2) I control the excitation frequency using a multifunction generator.

3) I control the frequency and amplitude but then controlling amplitude depends on the power supplied so i measured the impedance and tested with the different input levels to observe the amplitude and phase difference. I was able to solve the amplitude control case as it is directly proportional to the power supplied and also the frequency chosen. The frequency intern depends on two cases one has to be the impedance and other is the phase difference.

4) The piezoelectric transducer that i have in my setup is operated at 20kHz.

5) The feedback loop has to be quite past as it is more like lets say hitting the football continuously to prevent it from falling.

6) In the initial stages the duration of the session coulod be lasted to few samples during which the data can be easily acquired.

7) once the power input to the system is stabilized, then amplitude remains stable. But as we are trying to obtain the amplitude nat a paryticular resonant frequency, amplitude differs at different frequencies

😎 it is na small equipment so distruction is not a problem and i am sure that this setup will not harm anyone if it fails.

9)I am using the DELL LATITUDE and alternatively a PC, both are windows 7 operated and the NI DAQ that i use is DAQmx NI PCI-6024E

I started of with the impedance measurements but i Miss the crucial link between the present state and the further control process. I hope this information is upto your requirement and please do contact me in case if there is still something that need to be specified!!

Vipul

[gottardo.giuseppe]

If you need, you can find here my master thesis. I worked on the same topic some years ago.

Best regards

Giuseppe