# Solar PV Refrigerator UPS

Let’s do a little design project.  Let’s say that we want to build a solar photovoltaic uninterrupted power supply system(UPS) to be sure we always have refrigeration.  In order to put reasonable boundaries on the problem let’s say that we want a system that can capture enough energy in a day to power a refrigerator for four following days of cloud cover.  What would the system look like?

First we need to figure out how much power a typical refrigerator consumes.  Let’s start with this pretty generic refrigerator (link).  The energy guides says it uses 382 kWhr’s per year.  That means the average power input neglecting startup surges is:

(382 kWh/yearr)/((24hr/day)*(365days/year))= .044 kW or 44 Watts

So for the five days in our operating cycle we need:

(.044 kW)*(5 days)*(24 hrs/day)=5.2 kWhr’s

A kWhr is just a measured amount of energy.  Unfortunately it is a horrible unit.  The world should be using the proper engineering unit for chunks of energy called the Joule.  A Watt, the same unit used to described light bulb power consumption, is simply one Joule used up every second.  Therefore our 5.2 kWhr’s is better stated as:

(5.2 kWhr’s)*(3600seconds/hr)=18838 kJoules = 18,838,356 Joules

Now let’s assume there are six good solar charging hours per day and we use a Sharp 250W PV panel.  I have used this panel and liked its quality construction.

So for six hours at 250W the panel will gather:

(250 Joules/second)*(6 hours)*(3600 sec/hr)=5,400,000Joules

Uh! Oh!  We need more power than one panel can put out so we will need multiple panels.  We will need:

(18,838,356 Joules/5,400,000Joules) = 3.5 panels

Let’s just round up to four panels.  My Sawzall hates cutting through silicon solar cells.

Great!  Now we have the power source figured out so let’s figure out how to store the energy we have collected.  Obviously there will need to be some batteries in the system.  Battery capacity is measured units called “amp-hours”….another horrible unit.  The best description I have read regarding battery capacity is below:

Home power (deep cycle) batteries are generally measured in “amp-hour” capacity. One amp-hour is equal to one amp of current drawn for one hour of time. Amp-hour capacity is generally given as the “20 hour rate” of the battery. Therefore, the number given as the amp-hour capacity for a deep cycle battery will be the number of amp-hours the battery can deliver over a 20 hour period at a constant draw. A 105 amp-hour battery can deliver 5.25 amps constantly over a 20 hour period before it’s voltage drops below 10.5 volts, at which point the battery is discharged.

Let’s consider this Deka 12-Volt 810-Amp Marine Battery since it is readily available through Lowe’s Hardware store.  It is an 80 amp-hour battery so it will deliver four amps for 20 hours until the battery hits 10.5 volts.

In that time it will have expended approximately

(11.5volts)*(4amps)*(20hrs)*(3600sec/hr) = 3,312,000 Joules

Obviously one battery will not get there.  We need:

(18,838,356 Joules/3,312,000 Joules) = 5.7 batteries

Let’s round that up to six batteries.

OK.  Things are starting to come into focus, but there is one more thing to consider, heating during charging.  We don’t want to boil the acid out of them.  The maximum charging rate is considered to be the batteries amp-hour rate or “C” divided by five. Each of our 80 amp-hour batteries is therefore able to accept 16 amps as the maximum charging current.  Our four 250W, 24V panels are able to make:

(4 panels *250W/panel) = 1000 Watts

1000W/24V = 41.6 amps of current coming out of the panels

So if we wire the batteries as two batteries in series to make 24V and three groups of two in parallel, we will have (41.6/3) = 13.9 amps going through each parallel set.  Not too bad!

Hopefully I will get to build this later this year and let everyone know how it goes.  The charge controller and inverter sections will come next.

Hi Folks,

Glad to see everyone is enjoying my solar PV refrigeration writeup. Please, let me be clear. This write up was NEVER meant to be an optimal system. It was meant to frame the “dead minimum” for the stated requirements. There are a bunch of fudge factors that could and should be thrown in ie. whether the panels are tracked or stationary, battery charging inefficiency, inverter inefficiency….. at least you know that one 50W panel will not give you reliable refrigeration and 10kW is not necessary for the Plain Jane fridge you and I probably already have in our kitchens. More panels and batteries will always be better.

I will refine that write up as the summer progresses.

Stephen