Thursday, May 29, 2014

Some more up to date comparisons...

I'm off to work on Rev 2 of the design but it seems appropriate to layout why it is worth the bother.

Here are some comparisons to two good examples to give context :

Kyocera 325W Polycrystalline Solar Panel KD325GX-LPB
and (the awesome)
SunPower SPR-X20-445-COM (445 watts!)

Fortunately they are both 2.19 square meters each so watts per square meter STC can be derived but a better gauge is PTC (avail here) which mostly accommodates the Temperature Coefficient in a more real world manner:

Kyocera's goes down to 290.4 for the panel and so 132.6 per square meter
Sunpower goes down to 412.7 for the panel and so 188.45 per square meter
Hybrid V1 is 23 watts per tube and 128 per square meter

We could imagine those panels reformed as magic edge-less strips as long and wide as our tubes and get
23.9 watts per strip from the Kyocera
33.9 watts per strip from the Sunpower
23.0 watts from our tube (setting aside the thermal for the moment) 

How to layout the hybrid value? Here is an attempt at a tote board, normalized to per square meter

Top scored cells BOLD

OR if you'd rather:
1 panel of 2.19 square meters (1.66m X 1.32 m = 2.19 m^2) ≈ 12 Hybrid Rev 1 tubes ( .18 m square each X 12 tubes  = 2.16m^2 of area)

More on Hybrid version 1:
The Hybrid collector's first design iteration was analyzed in part by BRO ( This design revealed some shortcomings that I've attacked in the version two but the analysis of version 1 can serve as a "floor" for evaluation. The figures were good enough to keep going on... Some of the shortcomings of the rev one design and their easy resolution I'll cover later. But for now Rev 1 is the best understood as far as throughput. I think 20% of the losses in Version 1 can be clawed back but all of this post is about Version 1.

For a square meter of tubular collector 320 watts of visible light hit the PV cells in clear light. That light arrives on 10X10mm cells at 46,700 W/m^2 (46.7 Suns.) The 320 watts would hit 98 10mmX10mm concentrator cells in that design. and the cells should be able to operate at @66C or below. They could be expected to run at 40% (due to the rich diet and the high concentration/low temp.) that 320W*.40= 128w =12.8% efficient (started vs the 1000 Wm^2 total light) on the PV side in realistic atmospheric conditions. There are troubles with that 12.8% number but they cut both ways. The fraction of incident light captured in the heat components is about the same (340W) but the conversion is much more efficient ~80% at 160C output temp. So 340W*.80=272 or 27% on the thermal side for a net 39.8%

Some of the columns need explication:
Relative Performance on a Heat Load:
Cogenera (in a white paper) sees the _heat value_ as 25% of electric value in a Watt to Watt comparison when that heat is delivered at Domestic Hot Water Temps. Sounds fair. So 128+(272*.25)=196. I weighted them all at the Sunpower as 1 (the conversion losses in electric resistance heat being very low, these correspond to the Gross Efficiency number.)

Cooling Power in Watts:
Cogenra shows the way again in a paper "Cooling with Cogen" they find the value of high temp heat used in chilling service to be higher than plain old thermal DHW service due to the COP (coefficient of power) in double effect chillers. Thus we get the "Cooling power in Watts" comparison. In most cases I imagine those "Cooling Watts" would be Watts not used ("negawatts" generated at peak rates) and potentially freeing up watts from the electrical side to sell at those peak rates or to otherwise address demand charge problems. All subject to the particular installation conditions and metering contracts of course.