syncronisation with grid

Well, your PLL will track the frequency and phase of the grid. Once you know the frequency and phase, you will be able to track the rotating vector which in turn will allow generating synchronized inverter output. Note, the inverter output magnitude and phase need to synchronize with the grid. It goes without saying one will need to incorporate various noise reduction schemes to have a robust PLL immune to line noise, variation etc. The different types of faults line sag, line drop, frequency change etc will need to be taken into account as well. By the way, I am assuming you are referring to a three phase grid.

Hi Brijendra

You have to be more precise what you want to do. Do you want to derive a sine reference fro the controller, which is in phase with grid voltage ? Do you really need your sampling/switching frequency to be integer multiple of grid frequency? In this case does the sampling/switching need to be phase locked or is it enough to be frequency locked?

Cheers, Mitja

Hi Mitja

    I need to derive a sine reference which should be in phase with grid voltage and has the same frequency as that of grid. One thing i missed in the post was that it is a single phase grid. If any body has some application notes on this then please share it with me.

Hi Arefeen

     I forgot to mention in my post that it is a single phase grid not three phase. How do I synchronise with single phase grid? If you have any documents  on this, please share them with me.

Hi,

Brijendra Verma said:

I need to derive a sine reference which should be in phase with grid voltage and has the same frequency as that of grid. One thing i missed in the post was that it is a single phase grid. If any body has some application notes on this then please share it with me.

In this case you can get away without PLL. I prefer doing DFT on the input for the grid frequency bin only, then doing inverse DFT which yields first harmonic of the input variable. Then you only need to normalize it by the magnitude, which can be derived from the A and B parts of the DFT. It worked for me and I like this approach as it does not require any tuning, but there are probably a lot of different approaches

Cheers, Mitja

Hello... Sorry I haven't worked with a single phase system although the concept should be very similar to a three phase grid. If you are looking for a quick simple way to synchronize then can try detecting zero-crossing of your grid. You will of course have to deal with noise to avoid false detection but still detecting zero crossing will be quick first implementation.

Hi, Brijendra,

we will have propoer examples for your application soon. Can you check back in about couple of months?

Hi Arfeen,

Greetings. I have a similar problem. Actually I have two separate applications - one for single phase and another for three phase. The difference from Brijesh's application is that I do not have a hardware PLL, I have to write one in software.

I know I maybe asking for a bit too much ..... but I would really appreciate any help you can give me. I also know you have enormous expertise in this area ... since I attended your lecture at the Tenth anniversary of TIDC in Bangalore, however, I'll be honest I did not understand much at the time. Since then, however, I have done a bit of work on this synchronization problem - without much success and desperately need help and so in a better position to appreciate any inputs from you.

To begin with can you help me to develop the code for rotating vector please.

I hope you will be kind enough to offer me your help.  :)

regards

Anand

Hello Anand ... I am glad to know you still remember TIDC presentations ...that was few years back :-) anyway, I am not sure what kind of information you are looking for .. if it is related to software then TI C2000 group is your best resource and according to Zhen the system code sounds like is under development. We mainly work on three phase systems but implementing a single phase system should not be too difficult. Your first challenge will be to implement a software PLL so that your system can track the utility frequency and if needed can shut down your system safely. Your PLL will also have to robust enough to tolerate system noise and fluctuations. Once you have your PLL implemented, the next challenge is to design your power electronics hardware - mainly the filter, sensing circuits, charging systems, transformers etc. 

You can implement your rotating vector using the park and Clarke transformations. Basically sense your line voltages ... apply Clarke transform to obtain D & Q components from which you can calculate the rotating angle... just remember the angle will be about 30 degrees ahead of the actual value - this will be obvious once you implement the code. One good way to test this is to use real-time debugging to plot the waveform in CCS window. By the way, most of these software blocks are already available from TI.  I hope this helps and good luck with your project. By the way, what is your application and the power ranges?

Hi Arfeen,

Wow, Thanks for the super quick response.

Arefeen said:

Your first challenge will be to implement a software PLL so that your system can track the utility frequency and if needed can shut down your system safely.

Yes I have all the other software components to soft start and safely shut down the system as well as the software PLL implemented.

Arefeen said:

Your PLL will also have to robust enough to tolerate system noise and fluctuations.

That is exactly my problem though! The PLL I have implemented is not accurate enough.

Arefeen said:

Once you have your PLL implemented, the next challenge is to design your power electronics hardware - mainly the filter, sensing circuits, charging systems, transformers etc

 Yes - We have all of that in place. We have a working prototype.

Arefeen said:

You can implement your rotating vector using the park and Clarke transformations. Basically sense your line voltages ... apply Clarke transform to obtain D & Q components from which you can calculate the rotating angle...

 I am not sure how Park and Clarke Transforms can be used in a single phase system?

Arefeen said:

By the way, most of these software blocks are already available from TI.

Arefeen said:

I hope this helps and good luck with your project. By the way, what is your application and the power ranges?

Yes it does! Thanks .... At the moment I can't discuss the application and power ranges on an open forum...... Would it be too much to ask for your email at ti.com? I can discuss our project in detail then.....

Thanks once again

Best regards

Anand

Hello Anand .. In case you haven't noticed.. Currently I am not with TI ... :-) I am currently with Toshiba International Corporation, managing the electrical side of the low voltage ASD R&D group. We are responsible for LV ASD product roadmap, supporting customer base as well as in-house manufacturing. All our products are currently based on three phase AC Induction drives and ranges up 1400HP with a three phase supply ranging from 380V to 690V depending on application and location. Any way, if needed I can be contacted at arefeen@ieee.org

Hi Mitja Nemec,

I am trying to implement DFT in F28335 and I built the codes using matlab/simulink/embedded coder. While the same codes works in matlab environement, I get erroneous and inconsistent results while doing the same in hardware. Let me also tell you that I am trying to measure the RMS value and the Phase angle of the 60Hz mains. You mentioned that you were able to successfully implement that. Could you may be share your codes with me or maybe share what ever relevent documents that you have regarding this?

Let me also mention that I am new to F28335 and infact have limited knowledge regarding microcontroller codings.

Also I am using the pwm modules to generate hardware interrupts at 2880hz. My intention is to sample the 60hz signal at 48samples per cycle, implement DFT and figure out the phase and rms value. Then I also intend to time-tag this value (working towards building a Phasor measurement UNit)