I'm working with a solar contractor who has hit upon the idea of connecting four 225 Watt PV panels to a GE Geospring heat pump water heater. The GE will be modified to prevent use of its resistance elements. Since the GE compressor uses between 300-700 Watts, the concept appears feasible at first blush.

The GE will serve as a preheat tank to a conventional resistance tank, effectively doubling system first hour capacity most of the time

The idea is to leverage the COP of a HPWH in a warm climate (Jacksonville, FL) to create an alternative to solar thermal with similar cost and energy / solar factors. Additional benefits include far simpler plumbing and useful cooling and dehumidification wherever the HPWH is installed.

Your thoughts please!

I think the recovery time might be an issue. In reasearching this water heater, I believe I read that the normal recovery is about 2 hours.

Don't bother, feed the power to to the grid and keep them separate. Those panels don't put out 225w very often and seldom when you need it. The money spent on thermal is a much better buy and it is storable power (in a tank).

I agree that HPWH do exhibit long recovery times. Heating a completely cold tank using the heat pump alone can be expected to take 3-5 hours in the south, and extra hour up north

The idea here is that the HPWH will run during the peak sun hours (it uses between 300 and 700 Watts, so panels need not be at peak output) and then have the HPWH store the heated water for the rest of the day's usage by the downstream tank. The HPWH would probably be set 10*F above the desired hot water supply temperature to allow for standby losses.

My intent in creating this topic is not to reopen the thermal / PV debate but rather to get input on whether this would work, and if not, why not? One concern I have for the designer is whether the system will be able handle the inrush current as the HPWH starts.

Posted By engineer on 02 Oct 2012 07:27 AM
One concern I have for the designer is whether the system will be able handle the inrush current as the HPWH starts.

That is a genuine concern. Does florida not have Net metering?

I'd also be worried about inrush current, which could easily be twice the running current. Even if it does work, marginal current and low voltage could mean an early end to the compressor motor. You could use a battery bank and inverter and/or some kind of capacitive start, but this could easily make the project more complicated and expensive than it's worth?

We have net metering - solar contractor is trying to avoid going through the utility rigamarole with this "product".

We'll have to cheaply solve the inrush concern or scrub the plan

I don't know enough about the electrical characteristics of the panels to be able to discern a way to sense when the panels are seeing sufficient sun to start the compressor. My understanding is that Voc is quite high for relatively low levels of light; if so Voc is a poor predictor of sufficient power.

I agree with sesmith . The inrush current could be as high as seven time the run load current. The cost of the added components needed far out weighs the potential savings . Solar thermal would be a far better method with less problems.

You have to have the panels charge a battery and then the battery runs an inverter with enough output to handle starting current. Then you also need something to detect battery charge level and not allow the HP to run at all when it is low.

Absent a grid-tied, batteryless solar PV system, the cost economics of having batteries, maintaining them, etc, will blow out any reasonable economic analysis.  Especially if location is a typical city such that there's little rationalization for long power outages (days).

A grid-tied, batteryless solar PV system connects to the circuit breaker box and simply supplies electricity to anything in the home that calls for it, i.e. the HPWH.  Whatever solar output is not used is sold to the grid.

Four 225 W panels in a grid-tied, batteryless system is probably going to supply about 630 W at the peak of the day, on average across an entire year.  It will be zero in the morning and zero at the end of the day.

See my charts at http://www.welserver.com/WEL0043 for illustration of the solar power output curve.  My system is 36 panels at 225 W each, so simply divide the nubers by 9 and you'll have a model for your 4 panel proposed system.  I'm in a hot climate too (Dallas) such that environment should be similar to yours.

I'd be surprised if the cost economics work even for a grid-tied batteryless system.  Only 4 panels is pretty small to make the numbers work, unless there is substantial subsidy available in FL.

Hope this helps.

Best regards,

Bill