Thanks Mohamed. The question is all about variable frequency. +/- 5 Hz variation about a known nominal value, the frequency of variation (sinusoidal) is 2 Hz. These are all the boundary conditions in hand!

Dear Osama,

You may utilize a fast and accurate PLL mechanism to extract instantaneous fundamental frequency, first. Then follow the existing harmonic separation approaches.

Regards

Dear Babak, is there a preferrable or specific implementation of PLL that results in a good compromise between speed and accuracy?

Dear Osama,

The following papers may help.

Article Single-phase PLLs: A review of recent advances

Article Dynamics Assessment of Advanced Single-Phase PLL Structures

Article A Simple Approach to Enhance the Performance of Complex-Coef...

Article An Open-Loop Grid Synchronization Approach for Single-Phase ...

Article A Nonadaptive Window-Based PLL for Single-Phase Applications

Regards

Thanks Babak, appreciating your support

Thanks Mohamed, it is on the point

Dear Osama,

The following paper just came into my mind. Maybe useful.

Article A comparison of control algorithms for DSTATCOM

Practically, you can use PLL for variable frequency and LPF for fundamental component.

In matlab/simulink harmonics GUI there is an aswer you needed...in table model you can see it nominal value of individual harmonics of voltage/current depend on order/frequency harmonics...

To obtain the fundamental voltage component, you can use selective harmonic elimination technique.

I would not trust a Phase locked loop - they are surprisingly temperamental and could lock onto any frequency (or even none of them) including harmonics; so it won't be certain to find a fundamental from a mixture. I would just do a fourier transform and look for the lowest significant component.

Thanks Dr. Maine for the valuable advice. I would like to add your good self as a reviewer of a relevant paper of mine if you don't mind.

Osama, I would be happy to review your paperr. Also, how fast do you think the frequency is varying? Don't forget that a variable frequency can be resolved into infinite numbers of sidebands expressed as Bessel functions and might not even show a fundamental frequency at certain frequency deviations.

Thanks Dr. Maine, in fact I have submitted the paper to some journal a week after my request to your good self to be added as a reviewer. When your answer was not available for that time i assumed that you might be busy so that you can't provide time for reviewing the paper.

The frequency change in my case is not too fast, as I described in my question. It is in the order of a fraction to a couple of Hz per second.

If you don't mind I can send a copy of the submitted paper just to get your valuable advice and notes on it. Best regards

Thank you Arafa, it would be interesting to see your paper and I will happily make some comments on it.

Dr. Maine, the paper text was forwarded to your good self today morning

My email is [email protected]

Thanks for your time

I and other colleagues have made an effort to answer this question in a three phase setting. The method applied here can be simply extended to single phase application by creating an orthogonal signal and some other slight modifications.

We are developing this single phase application in a new research article

Article Frequency adaptive sliding fourier transform for synchronizi...

I extend the previous research effort conducted with my colleagues to the case of single-phase system. The work won the best paper prize in (The 6th International Conference on ADVANCED CONTROL CIRCUITS AND SYSTEMS (IEEE-ACCS’019)

Hilton Hurghada Resort, Hurghada, Egypt 17-20 Nov.2019)

The link to the paper is attached here and the simulink model can be found in the following link:

https://uk.mathworks.com/matlabcentral/fileexchange/73414-high-performance-single-phase-pll-with-sft-phase-detector?s_tid=srchtitle

Also, the work in this subject is collected in this project on research gate:

https://www.researchgate.net/project/High-Performance-VSI-Synchronization-Methods