Hi,

I agree with you. The inverter does not change its MPP because it is not a smart meter to adapt itself to the new condition (partial shading). I recommend you to read some paper in the field of partial shading of PV strings connected to DC links. They explain this matter in complete details. 

Assuming the strings are wired appropriately, there will be very small change in the overall array IV curve and so the inverter MPP won't shift significantly. However, if many strings are shaded then the resulting array IV curve can be distorted substantially and may show more than one maximum power point, of different power outputs. The inverter could lock onto the wrong one and operate the system less efficiently than possible. However, almost all real arrays should be positioned and wired so that this effect can't occur.

I suggest you to read this paper :

https://www.researchgate.net/publication/303318836_Shading_Effect_to_Energy_Withdrawn_from_the_Photovoltaic_Panel_and_Implementation_of_DMPPT_Using_C_Language

Article Shading Effect to Energy Withdrawn from the Photovoltaic Pan...

Answer pretty much depends on the MPPT algorithm onboard inverter.

Traditionnal Hill climbing MPPT might indeed trap on local maximum. In case of 2 (or more) maxima, they would typically stay locked on local maximum with highest voltage, i.e. closest to MPP point without shading.

In your case, only one string is shaded. If shading is uniform over modules of that string, the string will show only one optimum point, likely located at same voltage as the optimum of other strings. In that case total I-V curve of array will also exhibit only one maximum.

Therefore, In your example, Hill climbing MPPT will not be fooled by the shading. You can expect inverter to operate at maximum power of the array. And this result remains true also if shading is not uniform, because 15 other strings are not shaded, they carry the highest power level. Should a second optimum exist, it would be clearly away from the MPP voltage without shade. The loss in voltage (1/16, if shade concentrated on one module) is stronger than the loss in current resulting from the total loss of the string (1/16).

Yet, there exists other stronger shading examples where more than one maximum occurs, and where highest maximum is not at the right side of the IV curve. These cases are rare, as they should be prevented by the geometry of the array compared with potential surrounding shadowers. But these cases exist. For these cases, some inverter makers have developped shadow-proof MPPT, which e.g. intermittently scans the IV curve to detect presence of numerous MMPs.

Quan,

There are several types of MPPT algos, so it becomes to provide you with an answer that applies across the board !

To summarise much of what has been discussed in several of the replies, one of the three possibilitoies exist:

#1 and #3 are the two major issues commonly investigated as cloud transient effects on performance of MPPT's. You can get plenty of literature across the web by doing a simple keyword search.

-Sanjay

Dear Quan,

If your inverter has a functioning MPPT algorithm it will operate at a MPP of the array. This MPP may be a local or absolute maximum, just like a math function has local and absolute maximums. Wether the inverter gets stack at a local MPP or at the absolute MPP depends on the algorithm, as it was explained by others...

You write that the "inverter is not sensitive to such a string shading", but the inverter does not need to sense the conditions of the individual PV panels! The P(V) curve of any array is affected by the condictions of its panels. The MPPs of different panels are affected by irradiation, temperature, and shading, and it is natural that they differ more or less in real life.

What you need to understand is that the current through a string is same for all of its panels and that the voltage at the input of the inverter is the same for all strings connected at that input. This means that in a PV array all panels CANNOT operate at their MPP, but ofcourse, the P(V) curve of the array has a MPP and the task of the MPPT algorithm is to find the input voltage that corresponds to that MPP.

There is no way for all panels of the array to operate at their MPPs unless you use inverter for each panel. But again, this does not mean that the inverter cannot search and operate and the MPP of the PV array.

Regards,

Konstantin Kostov