|
|
re: system sizing
27 mar 2000
george ghio wrote:
>>>>>>>the panels put out 100ah per day and we use 100ah per day.
>>>>>>>the first day we have no sun but we use 100ah
>>>>>>>the next 4 days we get 100ah per day.
>>>>on an average day, you used 100 ah, but you only collected
>>>>6cycles(4daysx100ah+1dayx0ah)/30days = 80 ah on an average day...
>>in the question the panels put out 100amp hours per day. no average in it.
>>on sunny days. however, the average output was 80 ah/day, according
>>to webster. statistical fairies would say the "mode" was 100 ah/day.
>so your saying that 2+2=2 and 1+1+1+1=1 cause thats how the formula for
>averaging works.
i never said 2+2=2 or 1+1+1+1=1. i am finding it difficult to follow your
logic here, george. a solar power consultant like you might correctly say
that the average of 1 and 1 and 1 and 1 is 1, ie (1+1+1+1)/4 = 1.
>>and how, exactly, is "battery lifetime" defined?...
>a battery is considered worn out when it will no longer hold more than
>80% of its original capacity.
aha. thank you. now what does "when it will no longer hold more than
80% of its original capacity" mean? charge it up for a few days and then
discharge it to 0.0 volts at the c/100 rate, or less, and measure the
total amp-hours during the entire discharge? or use something like the
original battery spec? when you say "hold more," that seems to ignore
the less-efficient charging process and the higher self-discharge rate
as a battery ages.
>>>taken from battery energy's data;
>>>10% daily depth of discharge - 12-14 years life...
>>>thanks, george. do they have a lifetime vs temperature curve too?
>yes they do. give me a good reason for desigining a system that will
>cause you to buy batteries more often than you should have to and i
>will chase up the data.
cold batteries last longer, so they need replacing less often.
otoh, they are harder to charge and they have less apparent
discharge capacity (which reappears when they are warmed :-)
>>>or lifetime vs time spent in various degrees of discharge?
>
>>still waiting.
>as above
well, we know that batteries left in a state of deep discharge sulfate
quickly, and one trojan battery guru told me that keeping them on float
charge wears them out too. he said he believed the best way to prolong
the life of an idle t105 was to let it self-discharge to about 70%, then
recharge it. was he right? why 70%? is there a better strategy? can we
predict this from models with measured data?
>>>how do we combine these factors to predict lifetime?
>
>>still waiting.
>as above
pv batteries aren't going to be constantly charged or discharged or
sitting on shelves, but with more models and data and patterns of use,
we might figure out how to use them more economically.
>>...for my hard earned money i would settle for a 20% dod withe the
>>batteries kept as close to 25 degrees c as possable
>>keeping them 25 c when it's 5 c outdoors shortens their lifetime.
>does it? why?
cold preserves things, eg walt disney. most chemical processes like
sulfation and corrosion happen more quickly at higher temperatures.
people have known for over a century now that many things fail faster
at higher temperatures. if you believe in battery "discharge life,"
ie that they fail after a certain number of discharge cycles, consider
that they self-discharge more quickly at higher temperatures, and
these extra cycles shorten the discharge life.
>real example:
>
>s.d. bought second hand telecom batteries which were taken out of
>service after 10 years. s.d. has replaced them this year after 15 years
>use. they were kept at 20 - 25 degrees c for the whole time. we are
>talking about 25 years total use in ideal conditions. how much longer
>would they have lasted if they were kept at 5 degrees.
i don't know. that seems more like a story and a question than an
example. shark repellent works pretty well where i live, 100 miles
from the ocean. telecom batteries used for electronic system backup
are only discharged during grid power failures, ie almost never.
>>and just have to buy new batteries every 20 - 25 years...
>>>10% daily depth of discharge - 13.0 years life
>>why do you believe your batteries would last twice as long
>>as battery energy specifies for half that discharge depth?
>battery energy considers that a battery is considered worn out when it
>will no longer hold more than 80% of its original capacity. i consider
>a battery worn out when it ceases to do the job i chose it for.
>the two are not the same.
seems like a reasonable tradeoff. invest in more pv panel power to decrease
ongoing battery expense. so what's your end-of-life capacity fraction
threshold for a 25 year life at 20% daily dod? got any data to prove it?
the self-discharge rate also goes up. i'm getting ready to replace my
screw gun nicads because they need recharging more often.
>>fig. 5.6 in the 1995 (third) edition of practical photovoltaics by
>>richard j. komp, ph.d. shows ln(#cyc)=8.57-2.31d, approximately,
>>for a motive power battery...
>define ³motive power battery²
dr. komp saith:
forklift trucks and golf carts use large-capacity deep-discharge
batteries which are designed for long life and many discharge cycles.
cycle lives of 1,000 to 2,000 cycles are typical of these _motive
power batteries_, which are expected to last 15 years or more.
these batteries are heavy and expensive since they are built with
thick lead plates and a greater electrolyte capacity that makes them
more reliable under difficult operating conditions. a set of these
batteries, if acquired at a decent price, could give years of
satisfactory service at low total cost in a solar cell application.
>sounds like what is called a traction battery out here and is not the
>best battery to use for home power.
the "best" batteries may be extremely expensive, compared to mass-produced
golf-cart batteries bought from large local discount stores or golf courses
who replace them yearly, or when they can't drive the cart up some hill.
i owned an electric car powered by 14 used golf-cart batteries...
>>the time value of money and the fact that batteries wear out over time
>>(even with no deliberate discharging) seem like good reasons to use
>>fewer batteries with a greater average discharge depth.
>
>best to just design a system correctly and buy batteries that are
>correct for the job.
>
>>otoh, higher currents reduce efficiency and capacity, which bodes for
>>more batteries...
>
>or just the correct batteries.
"correct," as in "what george says"? :-)
>using averages to design a solar power system will leave you up to your
>average neck in an average truck load of average shit.
i am finding it difficult to follow your logic here, george. it seems
to me that a solar power consultant might correctly say that we need
to collect at least much solar energy as we use on an average day.
>the only place that you can apply an average is to the autonomy of
>batteries and depth of discharge. this average will be over 3 to 7
>days. not 365 days
sure. worst-case days, or maybe a worst-case month.
>system sizing by the numbers
oh goody.
>to determine the load, multiply the power of the appliance by the
>number of hours per day it will be operating. this gives an answer in
>wh for that appliance.
>
>e.g. a 300 watt pump running for 2.5 hours per day uses:
>
>300 x 2.5 = 750 wh
> = .75 kwh
great. (did you write this yourself?)
>add up all the energy used by all lights and appliances to get the
>total energy used for the household.
over a day. that's power, not energy.
>battery voltage
>
>to minimise system losses, as the load goes up so should the system
>voltage.
>
>load less than 1kwh per day - 12 volts
>load between 1kwh & 3kwh per day - 24 volts
>load greater than 3kwh per day - 48 volts
why not use 48v for a tiny system? the inverters and wiring can be
cheaper or more efficient. maybe this raises the battery cost.
>at this point we will convert our load to amp hours.
wh v
^ ^
>2352.9|/24|= 98 ah
>battery selection
>
>days of autonomy required is the number of days you want the system to
>run without/or low sun. 4 - 7 days is common. we will use 5 here.
why 5?
>maximum depth of discharge
>
>most batteries are will have a max dod of 70%. the batteries should not
>exceed this level after the selected days of autonomoy.
sounds cool. is he related to leonard nimoy?
>average daily depth of discharge (nick will love this)
>
>divide the maximum depth of discharge by the number of days of autonomy
>to get the average daily dod
>
>e.g. average daily discharge = 70% / 5 days = 14%
i love it.
>battery bank capacity
>
>daily load ah(...efficiency corrected) x days of autonomy / max dod
>(98 x 5) / .7 = 700 ah
nice.
>the battery capacity varies with the rate at which the batteries are
>discharged.
maybe the inverter wants a few nicads to lower the peak current draw
from the main battery bank...
>the discharge over 5 days of autonomy closely matches the 100 hour rate
>of discharge.
unless you turn on a toaster.
>therfore use the capacity of the batteries at the 100
>hour rate of discharge when determining the batteries to be used.
if you say so.
>allowance for low temperatures of battery enclosure...
>
>here is a rough and ready temperture conversion chart
>
> temp.degrees c conversion factor [motive power]
> -10 .85 0.62
> -5 .88
> 0 .91 0.77
> 5 .94
> 10 .96 0.88
> 15 .97
> 20 .99 0.95
> 25 1.00
>
>the capacity of batteries is influenced by the temp of the inclosure in
>which they are installed.
the apparent capacity. i added a few golf-cart points.
>as the temp goes down the capacity goes down.
just the apparent capacity, as i understand this: if you charge a
battery to 100% at 25 c, then store and discharge it at 0 c, you
only get 77% back, but you can get back the other 23% if you reheat
it to 25 c, as needed. meanwhile, the time it spends at 0 c makes
it last longer and self-discharge more slowly. a smart inverter or
homeowner might control the reheating to keep the batteries as cool
as possible, depending on the capacity needed.
>assume that the lowest daily average temp is 10 degrees c this means
>that we must determine the battery capacity on that day to allow for
>the worst case.
or warm up the battery. this warming energy can "release" more battery
capacity than it consumes, even with a battery powered heater. as an
alternative, we might turn on a small fan to blow some warm house air
through an outdoor insulated battery box. some plastic water jugs inside
the box could keep it from dipping much below 0 c in a cold climate
and allow it to store coolth from summer night air, as in a cool cell.
>placing batteries in parallel is not recommended. uneven charging and
>discharging of cells will lead to cells with different capacities in
>the system. the bank will then take on the characteristics of the
>poorest cell. paralleling also requires fusing of each string.
my electric car had aluminum wire links for series battery fuses.
>assume a efficiency of 90% unless other wise noted by the maker...
>energy from array = 98 / .9 = 109 ah
on an average day, as the maker determines.
>available energy from the sun
>
>this is expressed in peak sun hours(psh). radiation data for your area
>may be available in mj/m2/day or kwh/m2/day. to convert mj/m2/day to
>psh divide by 3.6. the figure forkwh/m2/day is the same as psh.
that works in the sahara desert...
>in our example we will assume that the site of the installation is in
>an area where the psh is 4.5
nrel measured an average min 2.1 kwh/m^2-day of sun in december over
the last 30 years on a lat+15 degree tilted south surface in phila.
the mean was 3.1. they also say the average daily max temp is 6.3 c,
which is nice for pv panel power.
>if we now divide the energy required by the psh figure we will get the
>output current required from the array.
>
>array current required =109 / 4.5 = 24.2 a
sorta.
>the output current of the mdule should be the current at normal
>operating cell temperture (noct) at 14 v.
and 6.3 c? why not add 3:1 reflecting fins to silicone-potted panels
and trickle some water over their faces in summertime, when more sun
is available?
>if we choose a module with inoct@14v = 3.96 a its guaranteed current
>will be 3.6 a - 5%. this is because the manufacturers state the output
>of their modules to be + or - 5%
>
>guarantied current = 3.96 x .95 = 3.76
a random assortment of 16 panels would have a "guarantied current"
that's closer to average, ie 3.96 a x (1-0.05/sqrt(16)) = 3.91 a.
nick
|
|
