I'm curious if anyone has seen or done real simulations or actual measurements on typical parallel combined and series string solar PV topologies to show what kind of overall impact various per panel/cell impairments have on system power performance.

I have an arrangement with 14 Kyocera 210GX-LP panels per string and 2 of these strings combined at a SMA SB7000US input string combiner

My STC power is 56 x 210 = 11,760 watts. As the light level of any one or group of panels is reduced by anything, it can drag the whole array down as it pulls other panels away from their MPP and it appears it can be dramatic in looking at the VP and IP curves that fall off pretty fast.

I believe someone must have done a basic parallel and series panel simulation with each panel modeled by its VI and VP curve characteristics in the array such that impairments could be applied to the individual panels or partial panels and probability and time of impairment such as:

Shade from sun moving at various latitudes and time of day and real characteristics seen in the world ( e.g. a distant tree versus nearby, antennae, etc..)

Clouds passing of differing height and opaqueness

Dirt, dust bird droppings, on panels

A partially failed panel with different expected failure modes ( short, resistance, altered curves etc..)

Other problems ???? suggest?

From the data, one could try and simulate some real scenarios and see the impact and effect of mean time MTBF ( between failure/occurrence) and MTBR ( time to repair if possible).

Then example = National Solar Magic -- http://www.solarmagic.com/

Then what impact a DC-DC converter that tracks the MPP and matches the VI and MPP of each panel? Then how complex and what is the cost of the optimizer so its cost effectiveness and reliability could be analyzed. It would clearly need to be cheap and reliable. I would need 56 of them.

Then adding a monitor using possibly Zigbee to each optimizer or by itself that monitors the panel VI over time and impairments or failures such that they can be isolated and fixed quickly. Same cheap and reliable ( but safer since its shouldn't be able to fail and impact performance -- just lose visibility)

Then example Enphase microinverter -- http://www.enphaseenergy.com/

They solve both problems by using one 95% efficient DC-AC inverter per panel with a communication link and performance monitoring. Their cost, complexity and reliability is scary but in theory its not bad.

Love to hear any thoughts on real simulation data or other products addressing both problems.

Are their any other approaches?

Tom