[MLB-WIRELESS] Solar power modelling

[Ben Anderson]

5v 2.4A of the dlink 900ap+ is a lot more than the 1.5watts you accounted
for on the spreadsheet.
Also remember you shouldn't discharge lead acid more than 30% or you'll kill
them quickly.

Bank on about 4-5 hours of sun a day, which means you need 6 times the power
out of your cell than the ap draws assuming 100% efficient batteries.  So
about 10 times the power to account for battery loss...  So to power a 10W
device 24/7, we'd need a 100W panel (around a square meter).  Add in 50%
reserve to account for overcast days. 150W of panels (probably about 1500
bucks worth).


240Wh/day, 7 days = 1680Wh reserve needed.  30% discharge - 5040Wh of
batteries.  Four enormous deep cycle batteries (probably 3-500 bucks each).

So 10W continuous you're talking somewhere around the 3-5 grand mark for a
solar solution.

Of course, presuming dlink overstated their power consumption, ya can scale
those costs back...  Anyone hung a multimeter off an AP to see what it
actually uses?

Ben.


> Hi,
>
> A week or so ago I asked a question about solar powering nodes and where
to
> get good data and simulation information. I got little response and
several
> I want to know as well messages. Here is what I found over the last week:
>
> Solar data.
> I was looking for good sources of solar radiation data.
> http://www.squ1.com
> A consultancy firm that appears to be part of Cardiff University in the
UK.
> They do tools and consultancy on building design. In there shareware
section
> they have a program called the weather tool. They also have available for
> free download hourly data files of weather for lots of locations around
the
> globe. The unregistered version of the weather tool does not allow you to
> export weather data to a text file. it is however a very useful and
> instructive program.
>
> Meteonorm
> http://www.meteotest.ch/en/mn_home?w=ber
> Meteonorm is just the ticket. You can download an evaluation version from
> their site. They have an extensive database of weather for the whole globe
> on an hourly basis. The download version only has 1 years data in it will
> only allow you to export data to a text file for a limited number of
cities,
> Frankfurt, Tokyo, Accra (Ghana), Brazillia and somehwere else I forget.
What
> makes this program ideal for our purpose is that you can define an
arbitary
> inclined plane and get it to export the incident solar radiation levels
hour
> by hour for a whole year. Most excellent.
>
> The model
> From the meteonorm package I exported a data file for Frankfurt and Accra
> (Ghana) for an inclined plane facing north at and angle of 45 deg (not
> otimum angle).
> The output looks like this:
>
> 1 1 7 7 51 46 91 67 74 27.3
> 1 1 8 8 248 122 365 151 424 27.9
> 1 1 9 9 393 215 470 235 352 29.4
> 1 1 10 10 606 236 712 275 545 31.3
> 1 1 11 11 768 257 879 307 637 32.7
> 1 1 12 12 764 359 844 401 466 35.1
> 1 1 13 13 801 339 887 383 528 37.4
> 1 1 14 14 677 372 733 394 372 38.6
> 1 1 15 15 533 342 584 361 272 39.3
> 1 1 16 16 287 247 287 237 73 39.0
>
>
>
> We are interested in column 7 which is the radiation incident on the
> inclined plane. The help documentation for meteonorm tells you what all
the
> columns are.
>
> I imported this into excel and added a few formula.
>
> The model is very simplistic. It models the charge in a battery that
> receives input from the solar cell and supplies output to the access
point.
> For each hour the battery stores a net input of solar radiation*area of
> cell*cell efficiency*batt efficiency
> and has an output of access point power. Units in Wh
>
> Solar radiation in W/m2
> Cell area in m2
> Cell efficiency is proportion of incident ration converted to electicity,
> number in range 0-1
> Batt efficiency is proportion of input power stored in battery, number
range
> 0-1
>
> battery charge is in units of Wh
>
> The equations for calculating the charge in the battery are clamped at 0
and
> the battery capacity in units of Wh
>
> The spreadsheets for Frankfurt and Accra are at:
> http://www.sentinet.demon.co.uk/wireless/frankfurtmodel.xls
>
> http://www.sentinet.demon.co.uk/wireless/accramodel.xls
>
> They use the following assumptions:
> Solar cell efficiency 20%
> Battery efficiency 66%
>
> The solar cell efficiency of 20% may be way off and perhaps a value of 0.1
> might be more realistic.
>
> Conclusion:
> It is surprising how big a battery you need and how small a solar pannel.
If
> we take the model for Accra in Ghana which is a lot sunnier than Frankfurt
> then you need a battery reserve of about 5 days worth of power but a solar
> pannel of only 0.1 square m.
> Frankfurt needs a pannel of 0.2 square m and a battery capacity of 7 days.
>
> A longer run of solar data is needed for a good analysis but this requires
> buying a copy of meteonorm which is $390US. It comes with 10 years of
> weather data. A more complex model would be good.
>
> Anyone know anyone with a registered copy of this program ?
>
> Lyndon
>
>
> To unsubscribe: send mail to majordomo at wireless.org.au
> with "unsubscribe melbwireless" in the body of the message
>



To unsubscribe: send mail to majordomo at wireless.org.au
with "unsubscribe melbwireless" in the body of the message