In my MPPT algorithm,
I am using P&O technique with voltage perturbation, for motor control applications.
For e.g. motor is operating at particular frequency e.g. 30Hz during noon time when sufficient irradiation is there, but if give some perturbation like pressing stop button and in fraction of second again giving start command, the motor frequency increases and now settles at increased frequency and that is MPPT point i verified.
So my doubt is sometimes the system settles at lower MPPT point until i give some large perturbation for it to settle to actual MPPT point.
So i was wondering how to implement this large perturbation and on what basis so that operating point dont get locked in lower MPPT point some times.
Currently my step size is decided by slope(dp/dv), how to give the large perturb in case the MPPT is locked at incorrect point lower that actual mppt sometimes.
Regards,
Mukesh
The question asked is completely dependent on the converter used. However, I would give a generalized opinion. One reason might be that the current from the input might not be constant even with an ideal DC source, i.e. the current is varying continuously. In such cases, the intervals at which you sense the current plays an important role.
For example, consider a current waveform 5+0.5sin(wt) and you are sensing the current at wt = pi/2. Then the current measured is 5+0.5 = 5. Now assume that current waveform is 5.2+0.5sin(wt) and the measurement is done at wt = 3pi/2. Then the current measured is 5.2-0.5 = 4.7 which is less than the previous value. In reality, the second case takes more current. However, the microcontrollers thinks that the second value is lower than the first and may result in computational errors.
The above example may not be seen in PV. However, cases similar to that might happen to result in the misbehaviour. When you give a large perturbation, it might reach a point where the measurements taken might fall in the real differential range. In all such cases, it is very important to adjust the step size or do necessary calculations like calculation mean of multiple readings and then providing it to the algorithm.
I hope this helps to solve your problem.
The key issues described are connected with the transition operation of the motor and the entire system, so the losses might have an effect. The key issues might not be on the MPPT side, I would suggest to evaluate operation of MPPT using not only motor, but some alternative loads to see the performance of the solution.
In case the motor is the required load, I would suggest to construct transition mathematical model of the system, from which you should be able to see, how transition processes in the motor affect your converter.
Thanks all, I will consider all the points suggested and see if i can try out with changes in current algorithm in order to solve the issue.
Mohammed Bilal Danoune Step size is already adaptive but sometimes the algorithm get stuck at other local MPPT point and require more step, which my current algorithm denies and thinks thats the MPP point.
Kharan Shiluveru The current has some noise, but if they are filtered then the algorithm is too slow. Will try to optimise this.
dIlya Kavalchuk Have simulated the same in MATLAB with motor model, couldn't find much related to this issue.
Regards,
Mukesh
Mukesh Chaudhary Either the filter is creating the lag or the other parameters are not optimized like the step size or may be inductors are too big that they don't allow sudden change or the variation of d i.e. delta d is too small.
- Badges
- Science topic