I have been considering something similar for a while now, so please report back on your progress.
Longi is apparently the world leader in bifacial panels, and are (or becoming) partners with ZA's Artsolar.
I predict if I ever do this, it will not be in the very near future. The reason I say this is I had a paid calculation done which relied on a high bifaciality factor.
The bifaciality factor I chose was higher than mainstream values, with a view that when I got around to doing this, it would approximate a realistic value not a historical one.
That said, let me try and explain what and why I did what I did.
I am a believer that an MPPT should within its threshold be optimized. Because of the bifaciality factor of theses panels one side is more productive than the other.
This means if your N-S vertical panels have their best side facing E then the morning peak will determine the MPPT size necessary. The afternoon peak will essentially underutilize that same MPPT capability.
So I thought, what if I didn't line up the panels exactly N-S, but offset them so that the sun's arc favoured the afternoon peak (at the expense of the morning peak), so that both peak's were the same size. That way I could use a smaller (read cheaper) MPPT that was fully utilized at both the am and pm peaks. My thinking was that at noon production would be low, so I could possibly also add a north-facing string and achieve full utilization throughout the day.
OK, so far so good. Except that knowing that this offset must exist, is quite different from trying to estimate what it is.
So I contracted Charles Landau at solarpaneltilt.com and posed him with this question ( for a fee).
This is an extract from his reply and I hope it helps in your own quest:
Based on your information:
Latitude 28.746 degrees South
Panels mounted vertically (fence)
Bifaciality factor 83.5%
Calculation for the summer solstice
The two sides of the fence will produce the same amount of power if the good side is oriented 315° from North, which happens to be exactly Northwest.
Your situation is unlike any I've encountered previously, so unfortunately I don't have any intuition of what the result would be at a different latitude or different day of the year. To determine that would require a new calculation, and a new fee.
This latitude approximates Kimberley or Ladysmith (KZN), the bifaciality factor I assumed was at the pinnacle of what could be achieved at the time, with the idea that this value would be commonplace by the time I got around to doing this. I chose the summer solstice as I figured this would be the day of the year that the peak disparity would be at its greatest.
If this information is helpful good luck to you.