Photovoltaics and Inverter

That's redfred. What's the story about an aquarium cooler, I've only experienced aquarium heaters?

Anyway, whether heating or cooling, a sensible goal is to achieve the desired amount of heat transfer during the day, with an integrated total, rather than all at once during the peak noontime sun. And rely on the tank water's thermal mass to get you through the day and night.

Below are 5-minute data points from my solar roof, over 9 successive days. Some are good solar days, others are disasters, like the cloudy 'yellow-squares' day.

Is yours a mission-critical task with no alternate power source? You have to be prepared for those low or nearly-zero-output solar days.

The above plot is at the annual optimum time, in mid-spring. Below, here's a sunny day in mid winter:

Note how the solar power fails to get above 6kW, and drops precipitously after 1pm, killing the normal symmetrical daytime curve. That's because mid-Dec has a low winter sun, which goes behind pine trees some 400-feet away. Only 24kWh produced in this full-sun winter day, compared to 62kWh in the spring time peak.

Some have suggested adding battery capability to your system, these curves help to show why you may need this.

100 watts at 120 volts is a resistor of value 14400/100 or 144 ohms. To find the amount of power delivered to a resistive load, you need to consider the load and source using load lines and the voltage and current characteristics of the source.

A photovoltaic device produces almost constant voltage but the current produced is proportional to illumination. To see the relationship, see this calculator.

http://www.pveducation.org/pvcdrom/effect-of-light-intensity

An example load line vs photovoltaic output is shown below:

https://www.mouser.in/applications/sophisticated-management-solar-energy/

So with your resistive load, the output voltage will vary directly proportional to illumination.

Solar illumination varies through the day, mainly because of absorption in the atmosphere (see Beer's law).

https://en.wikipedia.org/wiki/Beer%E2%80%93Lambert_law

https://www.cat.com/en_GB/articles/white-paper-hybrid-microgrids-the-time-is-now.html

Assuming your "appliance" can accept up to 120 volts, the voltage you can expect during any other time of the day if you connect it directly to the solar panel would follow the plot above, scaled for a maximum of 120 volts.

So if you want it to work all day and not just at noon, you need some kind of voltage regulator and a large enough array to produce enough power in as early in the morning and as late in the afternoon as you wish it to operate and limit the power at midday. The sun furnishes about 1000 W per square meter (at midday), but with the efficiency of solar cells you can expect about 150 W per square meter.

The daily variation is one reason to use batteries. A smaller solar panel could be used to provide average power for the day, and the solar panel output could still be used to switch on your appliance when the sun is out.

GA to Rixter for a very comprehensive answer.

My thoughts for the OP

To make this work effectively you really need to consider a battery both as a buffer for the incoming solar energy and for times of low insolation.

A 100 watt resistive load is going to draw around 8 amps at 12.5v or around 10 amps if driven via an inverter.

A 200 watt inverter would be quite adequate for the job, and a minimum 100 Ah battery would give you reliability of supply on all but the worst of overcast days without dipping below 50% SOC.

You may be able to get away with a cheaper modified wave inverter rather than a pure sine wave type, but be aware that some power packs (as I assume from your comment in #3 this device uses) do not work well with modified wave supplies.

Taking an industry standard 5 peak sun hours per day, and also allowing that you may need less cooling on colder days and outside of PSH times, you could expect to use roughly 1000Wh per 8 hr day. To replace this power you would require a minimum 300 watts of photovoltaics in an unshaded location facing the Equator and preferably pitched at latitude angle.

A regulator would be essential and I would suggest that you use an MPPT type in preference to a series or PWM type as it will provide a higher charging current for the same wattage PV array and it will allow use of a higher voltage array to both limit voltage drop and provide better performance at shoulder sun times when a series regulator would be limiting input due to impedance mismatch pulling the instantaneous I/V point right off the knee.

While an MPPT regulator would run the device without the battery, you will find that operation on all but clear days will be sporadic, and possibly non existent on highly overcast days - this may suit your purposes as you may need no cooling during those periods.

If using a 12v battery setup then PVs should be 24v minimum feeding an MPPT regulator with output voltage sized for the battery being used. The regulator should be positioned close to the battery to limit line losses.

If not using a battery, then you would be better to stick to a PV array of the same nominal voltage as the inverter input. Likewise if using a PWM regulator, the PV nominal voltage must be the same as the battery or inverter.

You would need a relay to turn the system on and off, and this could be achieved with the relay coil connected directly across the PV array with a delayed off feature to cater for clouds etc.