What about this for water heating

techntrek

Someone in another thread mentioned using PV for water heating vs. solar thermal due to needing a heat exchanger and antifreeze. Yes, solar thermal is much cheaper for water heating, and more efficient, but it does require some maintenance and components can fail especially if antifreeze is in the system design. Not looking for the dozen reasons you wouldn't want to use PV to heat water. This is more of a mental exercise.

So, let's say for one reason or another you need to heat water with PV. You don't need an inverter since the element works fine on DC. One of the advantages of a water heater is it has good thermal carryover, so just because you need a shower at 5 am doesn't mean you need to heat water at 5 am. No inverter and no need to heat when the sun is down means batteries can go which eliminates that inefficiency. Also no charge controller.

That leaves the panels and where the question is. Normally PV by itself is fairly useless, even solar water pumping needs a pump controller for several reasons. But in this case the load is just a resistive element. So... size the array to keep the voltage below the max the element can handle and all is well, or won't work? Maybe stick a pump controller in the equation - although I'm not sure why you would since that would reduce the voltage in sub-optimal conditions.

inetdog

Re: What about this for water heating

techntrek wrote: »

That leaves the panels and where the question is. Normally PV by itself is fairly useless, even solar water pumping needs a pump controller for several reasons. But in this case the load is just a resistive element. So...

size the array to keep the voltage below the max the element can handle and all is well, or won't work?

Maybe stick a pump controller in the equation - although I'm not sure why you would since that would reduce the voltage in sub-optimal conditions.

The problem is that a resistive heating element has a voltage across it that is proportional to the current flowing, while the panel output has a relatively constant maximum power voltage but a current that varies directly with the amount of light. An MPPT CC effectively changes its input resistance characteristic to keep the voltage at or near VMP. So you could use an MPPT CC to deliver a constant output voltage that goes into a heater. But an MPPT CC also needs to be connected to batteries full time, requiring complicated controls to keep the heater from sucking power from your batteries too.

A solar pump controller will drop the voltage to the motor to deliver as much current as it needs to deliver maximum pump output with the amount of power that the panels are producing. But that will still not necessarily work with a resistance heater as the load. Worth more investigation though.

BB.

Re: What about this for water heating

DC from solar panels is just not "right" for efficiently driving a resistive load... Variable current into a fixed resistance means variable voltage from the panels (i.e., you will only be seeing "optimum power" conversion from the panel in the middle few hours of a sunny day). If that works for you--It is OK.

However, I would certainly look at doing a DC element (issue with DC relays vs AC relays), or just using your AC inverter with (perhaps) smaller wattage heating elements.

The "best" way (most hot water per kWH), would be to use a heat pump water heater (like GE Geosprings) or some of the retro fit heat pump units (with lots of thermal insulation on the hot water tank).

-Bill

techntrek

Re: What about this for water heating

I agree a heat pump would be the most efficient overall, but that also introduces a bunch of complexity and failure modes, again. Inverter, batteries, charge controller, plus the heat pump.

inetdog mentioned a MPPT charge controller to improve the PV output. Must they be attached to a battery? A battery is a resistive load, so is the heating element (as long as the thermostat doesn't turn it off, but that problem can be fixed by killing the PV output when the water is hot enough). Maybe it would be ok directly driving the element?

BB.

Re: What about this for water heating

Technically, a battery is not a resistive load... It is a Voltage Source/Sink. More or less, the battery chemistry tries to keep the battery voltage between 12.0 to 14.8 volts or so (normal operation).

You can place a Resistive load on the Battery Bank and it will see ~12.0 to 14.8 volts. A much narrower range than you will see from a solar panel (still will give you, roughly, a 1.5:1.0 range of power variation with DC voltage).

If you use an AC inverter, then the AC voltage variation will be much less.

The MPPT solar charge controller will optimize for the battery voltage for variable solar panel output current.

I wonder if somebody could use a linear current booster between the solar electric panel to 12 or 24 volt heating load:

http://www.solar-electric.com/7amplincurbo.html

Could avoid the "Battery Bank" -- But LCB's are not cheap either.

-Bill

inetdog

Re: What about this for water heating

techntrek wrote: »

inetdog mentioned a MPPT charge controller to improve the PV output. Must they be attached to a battery?

Many MPPT CCs require a constantly connected battery as a source of control power when the panels are not producing. In particular if you connect the panel power first before connecting the charging leads to a voltage source (the battery) they will not reset properly and will not work. The exact bad behavior you see will vary from model to model.

They also expect a non-linear relationship between the CC output currrent and the CC output voltage, and may not behave well when connected to a resistive load only.

stephendv

Re: What about this for water heating

BB. wrote: »

DC from solar panels is just not "right" for efficiently driving a resistive load... Variable current into a fixed resistance means variable voltage from the panels (i.e., you will only be seeing "optimum power" conversion from the panel in the middle few hours of a sunny day). If that works for you--It is OK.

Could you quantify this non-optimum power? I.e. does the difference between optimum power using an inverter vs. non-optimum power using only panels make economic sense. Or would it be better (economically) to just install more PV to make up for the non-optimum bit, instead of installing an inverter?

BB.

Re: What about this for water heating

Hmmm... Lets see if I can make this clear/understandable (it is not easy to see)...

More or less, an MPPT type charge controller (or even a PWM controller charging a battery bank) is a "good match" for a solar charge controller. Remember that one of the equations for power is:

• Power=Voltage*Current

Vmp is (relatively) stable (wanders a bit with temperature), but with voltage "fixed" in the power equation, only the panel current varies with the amount (and angle) of sun. So, the relationship between power and voltage is "Linear". If you get 100% sun, you get 100% current, and 100% output power (roughly). If you have 50% sun, you get 50% current, and 50% output power.

However, when you connect the solar panels directly to a resistive load. If you have set up the panels/array such that the Vmp and Imp equal the voltage rating and current rating across the resistive load--The power equations would be:

• Power = Voltage * Current
• Power = Current² * Resistance

Because voltage across a resistor is a function of the amount of current passing through the resistor (aka heater), when current falls, so the voltage. The second equation is actually easier to see/quantify what will happen.

At 100% current, you have 100% power. However at 50% current, you have 0.50² = 0.25 --- Or 1/2 the sun gives you 1/4 the power (heat). Even 80% sun gives you only gives you 64% of the power...

So, you can see that your resistive heater on a solar panel is only "optimum" very near rated Voltage*Current. And with solar panels that have variable output current during the day, the power collected will fall with the square in the drop of current. And give you optimum power for only a couple of hours in the middle of the day (with good weather, a tracking array, etc.).

You could possibly "hack" this system behavour by using, for example, parallel resistive heaters and switch them in/out depending on voltage across the resistors, or possibly a capacitor bank ("micro battery bank") and a pwm controller to hold Vmp from the array, a linear current booster, etc....

Even with battery banks and AC inverters, we have substantial losses (the overall ~52% end to end efficiency of an off grid AC power system)--So, it is possible you may, over all, get "good enough" performance with a lot of solar panels (and no battery/AC inverter, etc.)--But I would want to look at the question more.

You could try using the Hourly Ouput numbers from PV Watts and a spread sheet and the P=I²R relationship over 1 year with fixed and tracking array and see what the actual amount of power collected would be from a "standard" off grid system and a Panel+Heater system to see which would work better. It may be better (or worst) than I think.

Does this make sense?

-Bill

NorthGuy

Re: What about this for water heating

I would also add that there are certain factors that make the effect less dramatic.

1. The MPPT controller is going to consume some power (5% or so). With direct connection, there's no loss.

2. When voltage decreases, current increases a little bit. So, at 80%, when voltage drops roughly 20% below Vmp, the current will increase from Imp to close to Isc (which is 6% for my panels).

3. You can adjust your heater to NTC rather than STC, so you get a good production most of the time. You only loose dramatically in cloudy weather.

techntrek

Re: What about this for water heating

The linear current booster (solar pump controller) was part of my original question, but my worry about using one is you want to keep the voltage high on the heating element. I've seen discussions here many times saying that there is something like a logarithmic drop in heat output as voltage drops (I know I'm quoting that wrong, please correct me). The solar pump controllers drop the voltage while boosting current to keep the water pumps moving where they would otherwise stall.

I like the idea of a simple voltage-based controller that swaps an extra element in/out of the system. Most water heaters have room for two. Maybe design the PV output for two but drop one out as PV output drops.

techntrek

Re: What about this for water heating

NorthGuy wrote: »

...
3. You can adjust your heater to NTC rather than STC, so you get a good production most of the time. You only loose dramatically in cloudy weather.

What do you mean by NTC and STC? I'm guessing an abbreviation for terms I've seen here a thousand times, but making sure...

NorthGuy

Re: What about this for water heating

techntrek wrote: »

What do you mean by NTC and STC? I'm guessing an abbreviation for terms I've seen here a thousand times, but making sure...

STC is "Standard Test Conditions", which is how all panels are rated.
NTC is "Normal Test Conditions", i.e. 800W/m² etc., wchich gives you lower Imp than STC.

stephendv

Re: What about this for water heating

Thanks Bill, makes perfect sense. So this would be a good example where a split SE+SW array would be better than a direct S facing system, so that there's a longer constant Imp rather than a midday peak.

techntrek

Re: What about this for water heating

Interesting. Couple a dual array with the dual heating elements on a voltage-triggered controller and you could keep things simple while keeping the voltage up.

This begs the question of why you would actually do this. One situation I can think of would be a house well inside the tree line where solar thermal is impossible, and a very long run to an area which does have a good southern exposure. Of course then it would make more sense to just install an oversized PV system.