Fishing for good ideas for implementing diversion loads...

[RegGuheert]

Whenever I have thought about implementing diversion loads in the past I have always had lots of ideas but even more questions.  To date, I've never put any diversion loads to work.

But with the recent off-grid system I installed I went with a slightly-larger array and a slightly-smaller battery, partly due to the current economics and reliabilities of PV and batteries.  Here are the basic specifications:

PV: 20-Sharp NU-U240F1 (4800 Wp: 5 strings of 4 modules for 120.4 Vmp @ 39.9 Amp)
CC: MidNite Solar Classic 200
Batteries: 8-SunXtender PVX-405HT (405 Ah @ 48V)
Inverter: Schneider XW6048
Generator: Honda EU6500i converted for propane use
Water is heated with solar thermal year-round
Space heat is using a solar thermal base with propane back-up plus a Finnish fireplace

The location of the batteries, inverter and charge controller has moved back and forth between the house and the array several times, but I THINK we have finally settled on an equipment shed under the array, considering the equipment is now installed! 

The array is in VA and is mounted at a fixed 50 degree elevation which provides optimum wintertime production at the expense of summertime output.  Available electricity generation with a 77% derate factor from PV-Watts shows average daily production with peaks in March and October at 18.1 and 16.1 kWh/day respectively and valleys in December and July at 11.5 and 15.3 kWh/day respectively.  The house has been designed to be very efficient and to have many loads prefer to operate during the daytime, but of course there will still be some loads at nighttime.  Without the home being occupied yet, it is quite difficult to pin down the exact loads, but the nature of non-infinite batteries means that there will ALWAYS be periods of excess PV capacity.  With this system, I'm expecting there will be some periods where there is a significant amount of excess, but of course that is when it is needed the least!

Anyway, here are some ideas that are being considered for diversion loads:

Loads near the Classic controller:
- Equipment shed baseboard heater in wintertime.
- Equipment shed air conditioner in spring, summer and fall.

Loads near the house, 150 feet from the solar array:
- Auxilliary water heater element for 163-gallon tank in wintertime.
- Water pump for decorative fountain in spring, summer and fall

So here are my questions:

1)  It appears that Ryan is having success getting the MNGP to work at distances that will allow it to be in the house.  Is there any way that diversion loads can be controlled from the remote graphics panel?  That does not appear to be the case, as it only has the one plug.  Perhaps there will be another remote panel in the future with additional "stuff" built-in?  Something like another classic without any of that complicated PV charger electronics included?

2)  If the above are not in the works, perhaps there are good-quality diversion controllers available or coming that can communicate with the Classic via ethernet.  Does anyone know of anything along these lines?

3)  What are your favorite solutions for solid-state relays that can be controlled by the Aux outputs on the Classic?  I may need both 120VAC and 240VAC units at different power levels.

4)  It seems obvious to add diversion loads in Float mode, but does the Classic know the available extra power at that time?  I'm thinking it might due to the fast sweep capability.  Can that information be used to control when an Aux output is enabled?

5)  Less obvious is adding diversion loads during Absorb mode, but again, it seems the Classic knows how much additional array power is available, so I'm wondering if Aux outputs can be also controlled based on available additional power when in Absorb.

6)  Is an air conditioner a good diversion load, or is it likely to be tortured to death by the amount of cycling that would go on in the summertime?  Also, will a modern, efficient air conditioner stand up to freezing temperatures in the wintertime?

7)  What are some good ideas for diversion loads in the spring and fall?  As discussed, this is when we have peak potential production but the loads are minimum.  (It sure sounds like a miserable system design when I put it that way, but it is what it is!  )

8 )  Do you have any other "favorite" diversion load ideas that you can recommend?  I'm sure lots of cool things have been done that I have never even imagined, so I would love to learn about them!

9)  Are there things about diversion that "I don't know that I don't know"?  Of course there are!  If there is something obvious that I should be asking about but that I am not, please educate me.  I can (sometimes) be taught!

TIA for any and all responses!  Don't feel like you need to answer all nine queries!  Please just chime in on the ones where you have some thoughts!

Reg

[tallgirl]

Reg,

Pumping water is always a nice use for surplus electricity.  Pumping water uphill into a storage tank is an even better use.

The problem I see with your setup is that "solar PV" and "solar thermal" sort of go together -- when it's too cloudy to make hot water from the sun, it's also too cloudy to make hot water from electricity.

About the only useful think you can do on a bright and sunny day when you have too much electricity, besides pumping water uphill, is charging more batteries.  Which I think might be the best use for the power that you've got -- other than the ones you've mentioned.

As for remote controls, the Classic will supposedly be able to issue Modbus macros to control external devices, such as a Brand "MS" Relay Driver.  If you want fancier than that, I sell gear that will do tricks.

[RegGuheert]

Thanks for the reply, tallgirl!

Pumping water is always a nice use for surplus electricity.  Pumping water uphill into a storage tank is an even better use.

Hmm.  This house is at the top of a large hill that leads down to a stream.  Vertical drop may be 100 feet or so (with a much longer horizontal distance).  I know that utilities around here use pumped storage to consume energy at night and make it available during the daytime.  I'm wondering if it there is any utility in pumped storage for PV in which you pump during the day and generate at night.  Somehow I imagine batteries to be much cheaper, but it sounds like a solution Rube Goldberg would LOVE!  Has anyone done this?

The problem I see with your setup is that "solar PV" and "solar thermal" sort of go together -- when it's too cloudy to make hot water from the sun, it's also too cloudy to make hot water from electricity.

Agreed.  Solar thermal for space heating is a very difficult proposition.  I tried to dissuade the homeowner, but they wanted to do this.  In the end I think it will be a nice solution.  Given that the system heats water as well as the home, I expect we will have PLENTY of hot water available during most of the year.

That said, what is different in this application is that this house may be unoccupied much of the time.  As a result, there will likely be long periods in the wintertime when there is an overload on the solar thermal system but full batteries and plenty of excess PV available on sunny days.  In these situations I would like to be able to heat the house using the excess electricity.

About the only useful think you can do on a bright and sunny day when you have too much electricity, besides pumping water uphill, is charging more batteries.  Which I think might be the best use for the power that you've got -- other than the ones you've mentioned.

There is always a limit to the amount of batteries that you can charge whereas there will definitely be deficits in the heat available for the task from the solar thermal system.

I would also like to have a small heat pump in the under-array equipment room.  In the wintertime, I would like to have it heat the room when there is excess PV and in the summertime I would like to have it cool.  Alternatively we could use an air conditioner and a resistive heater.  This task is a little easier, as the load is right near the charge controller.  Has anyone done something like this?

As for remote controls, the Classic will supposedly be able to issue Modbus macros to control external devices, such as a Brand "MS" Relay Driver.  If you want fancier than that, I sell gear that will do tricks.

I'm interested!  I'll have a look at the MidNite Solar Relay Drivers and your website when time allows.  Thanks!

Reg

[RegGuheert]

4)  It seems obvious to add diversion loads in Float mode, but does the Classic know the available extra power at that time?  I'm thinking it might due to the fast sweep capability.  Can that information be used to control when an Aux output is enabled?

Thinking about what I want to implement makes me feel like this would be an attractive approach to diversion.  To do this properly would require the ability to program in a large amount of hysteresis, even allowing for a NEGATIVE cutout power level.  Also, I think it would be necessary to allow for a delay setting to accommodate startup surge and to reduce the cycle frequency to minimize wear-and-tear on the load.

Any thoughts on possibly implementing this capability, boB?  I imagine you have considered this before.  Maybe this feature is already in there?  Perhaps there are showstoppers here that I have not considered?

TIA,

Reg

[boB]

4)  It seems obvious to add diversion loads in Float mode, but does the Classic know the available extra power at that time?  I'm thinking it might due to the fast sweep capability.  Can that information be used to control when an Aux output is enabled?

Thinking about what I want to implement makes me feel like this would be an attractive approach to diversion.  To do this properly would require the ability to program in a large amount of hysteresis, even allowing for a NEGATIVE cutout power level.  Also, I think it would be necessary to allow for a delay setting to accommodate startup surge and to reduce the cycle frequency to minimize wear-and-tear on the load.

Any thoughts on possibly implementing this capability, boB?  I imagine you have considered this before.  Maybe this feature is already in there?  Perhaps there are showstoppers here that I have not considered?

TIA,

Reg

Well, this is what the diversion mode "Relative to Charge stage voltage"  is supposed to be for.  OutBack calls it "Opportunity" mode....   We call it "Use it or Lose it" mode.  (Youie-Louie, sort of like the song, Louie-Louie).   This mode keeps the Absorb / Bulk timers counting while diverting and is supposed to divert as best as it can while keeping the battery voltage at the programmed Float or Absorb voltage.  If there is enough power available, it will be able to hold that voltage.  If not, then it just goes back to MPPTing best as it can and diversion turns off during that time.

Another mode that has been discussed   is to divert based on a programmed amount of power generated,  all with hysteresis adjustment of course.

Is this the  kind of stuff you are thinking about ?   We will definitely entertain other ideas too of course.  Bring 'em on !

boB

[RegGuheert]

Well, this is what the diversion mode "Relative to Charge stage voltage"  is supposed to be for.  OutBack calls it "Opportunity" mode....   We call it "Use it or Lose it" mode.  (Youie-Louie, sort of like the song, Louie-Louie).   This mode keeps the Absorb / Bulk timers counting while diverting and is supposed to divert as best as it can while keeping the battery voltage at the programmed Float or Absorb voltage.  If there is enough power available, it will be able to hold that voltage.  If not, then it just goes back to MPPTing best as it can and diversion turns off during that time.

Another mode that has been discussed is to divert based on a programmed amount of power generated, all with hysteresis adjustment of course.

That's what I'm wanting to do!  The idea is that each time the controller does a sweep it can calculate the difference between the peak power it can produce versus the amount of power that it is producing.  It could then compare the amount of available extra power to some threshold provided by the user and switch on a load if the available power is high enough.  After some timeout period, it could again start comparing the available power with some lower threshold (default 0) at which it will then turn the load back off.  The idea is to only divert if there is sufficient power to run the load without any impact on battery charging.

Is this the  kind of stuff you are thinking about ?   We will definitely entertain other ideas too of course.  Bring 'em on !

boB

Thanks!  Does this sound doable, IYO?  (I'm not asking for a commitment, as I know you are very busy!)

[boB]

Well, this is what the diversion mode "Relative to Charge stage voltage"  is supposed to be for.  OutBack calls it "Opportunity" mode....   We call it "Use it or Lose it" mode.  (Youie-Louie, sort of like the song, Louie-Louie).   This mode keeps the Absorb / Bulk timers counting while diverting and is supposed to divert as best as it can while keeping the battery voltage at the programmed Float or Absorb voltage.  If there is enough power available, it will be able to hold that voltage.  If not, then it just goes back to MPPTing best as it can and diversion turns off during that time.

Another mode that has been discussed is to divert based on a programmed amount of power generated, all with hysteresis adjustment of course.

That's what I'm wanting to do!  The idea is that each time the controller does a sweep it can calculate the difference between the peak power it can produce versus the amount of power that it is producing.  It could then compare the amount of available extra power to some threshold provided by the user and switch on a load if the available power is high enough.  After some timeout period, it could again start comparing the available power with some lower threshold (default 0) at which it will then turn the load back off.  The idea is to only divert if there is sufficient power to run the load without any impact on battery charging.

Is this the  kind of stuff you are thinking about ?   We will definitely entertain other ideas too of course.  Bring 'em on !

boB

Thanks!  Does this sound doable, IYO?  (I'm not asking for a commitment, as I know you are very busy!)

Ahhhh, I see now.  It sounds easy at first but when sweeping, it is usually done because it is in Bulk or Float MPPT and needs
all the output power it can make (and more) just to run the loads already... Your diversion load would have to be enabled just
to do a sweep, unless you want the battery voltage to overshoot, possibly severely, during times of excess power availability.

The sweep will hopefully stop early i the battery voltage reaches the Absorb, Float or EQ set point.

A way around this problem could be to temporarily turn on your extra diversion load. So if might be better to just turn
on that diversion load once in a while during Absorb, Float or EQ.

boB

[RegGuheert]

Thanks, bob!

So it sounds like a full sweep is ONLY done when the load, including the battery, is more than the array can produce.  When the array power is sufficient to reach the current setpoint, it sounds like you sweep from one of the ends of the power curve up to the point where the demand is met and stop there.  Makes sense.  (For some reason, I was under the impression that the sweep was fast enough to cover the entire range without causing the battery to spike.  Of course if it went that fast, then how would you know what setting to use to obtain the proper battery voltage? Duh!)

The load I have in mind is a heat pump for the power shed so it is not something I want to turn on and off too often.

O.K.  Before I give up on this, let me explore one more possibility:  I imagine for any given array a fairly simple equation could be used to estimate the peak available power given Isc and Voc.  Perhaps fitting only a parameter or two is need to fit the shape well as the endpoints move around.  Since Voc will typically move around only slowly with temperature and light level, if you normally perform the sweeps starting at the short-circuit part end of the curve (which I doubt), then it may be possible to accurately estimate the available power just from the partial sweep that is already done.  No additional sweeping would be required.  Otherwise, perhaps a measurement of short-circuit current would be also be needed to make the estimate.

The point here is that only an *estimate* of available power would be required for this purpose.  Nuisance cycling could be prevented by either having the estimator err on the low side or by having the user pad the power requirement on the high side.

Anyway, just thinking out loud.  The bottom line here is that being able to turn on a diversion load based on the available extra power would be a VERY useful capability IMO.  Thanks for any consideration!