Designing a 30kW system - details?

[chrisjx]

My first shot at trying to design an off grid system for a new house near Brenham Texas.

In my existing home, I use about 19kWh/day. It is a small house with no AC, only gas heating. Also gas stove and gas clothes dryer. The biggest draw in the house is a home lab server system which draws about 400W continuously. In any case, this will be far a new house (2400 square feet) on 10 acres so low energy components will be favored and because space is not a problem, we plan to set up the array on a ground mount system. I've set the sizing to support 30kWh per day and perhaps that's a first problem but it's what I've calculated so far.

What follows is a very basic overview of the primary requirements and some components I think will be required and some generic prices I've pulled from the internet. No manufacturers defined yet. I know I'm missing some of the smaller items like breakers and bus bars but I need to get this out to some people with experience and see if I'm even on the right track, regardless of how accurate I am in this first pass.

Any tips or advice appreciated.



Location and Requirements
- Average Daily Usage: 30 kWh
- Days of Required Autonomy: 3
- Location and Solar Irradiance: 30.15, -96.46 - open field, no trees
- Soil Analysis: Dark clay soil

Components and Pricing
- Solar Panel Type: Mono-crystalline panels, 400 Watt each = $400
- Charge Controllers: 3200W/48V = $300
- Battery Type: Lithium-phosphate batteries, 10kWh each = $1800
- Inverter: 6000W = $1600
- Ground-mounted Racking: Aluminum or galvanized steel = $8000
- Single strand 2/0 AWG copper wire: $200 / 100 feet


Calculations and Design

# Expected Hours of Full Sun
- Average of 5 hours of full sun per day.

# Solar Panels and Strings
- Total Energy for 3 Days: (30 kWh/day times 3 days = 90 kWh)
- Energy Production Per Day: (90 kWh / 5 hours/day = 18 kW)
- Panels Required: (18,000 W / 400 W/panel = 45 panels)
- Optimal String Configuration: 5 strings of 9 panels
- Total Cost for Panels: (45 times $400 = $18,000)

# Charge Controllers
- Controllers Needed: 5 (one per string)
- Total Cost for Controllers: (5 times $300 = $1,500)

# Inverters
- Inverters Required: 2 (for resilience and 240VAC capability)
- Total Cost for Inverters: (2 times $1600 = $3,200)

# Breakers
- Breakers for Panels: 5
- Breakers for Inverters: 2

# Cabling
- Distance to Inverter: 60' underground
- Wire Required: 2/0 AWG copper wire (60 feet)
- Total Cost for Wiring: ( frac{60}{100} times $200 = $120)

# Batteries
- Batteries Required: 9 (to meet 90 kWh for 3 days)
- Total Cost for Batteries: (9 times $1800 = $16,200)

# Tilt & Orientation
- Fixed: 30 degrees
- Tilt for Winter: 40 degrees
- Adjustment Dates: October 15 to April 14
- Tilt for Summer: 20 degrees
- Adjustment Dates: April 15 to October 14
- Orientation: South facing

# Ground Mount and Foundation
- Material: Aluminum or galvanized steel
- Total Cost: $8000

Summary
- Total Solar Panels Cost: $18,000
- Total Controllers Cost: $1,500
- Total Inverters Cost: $3,200
- Total Batteries Cost: $16,200
- Total Ground Mount and Foundation Cost: $8000
- Total Cabling Cost: $120

# Total System Cost: $47,020

 

[Texican]

panels are more like 50 cents a watt, not 1.00
so cut the panel cost in half

will you have a grid connection avbl ?

 

[chrisjx]

panels are more like 50 cents a watt, not 1.00
so cut the panel cost in half

will you have a grid connection avbl ?

Good to know I'm off on the high side for panels. That adjustment brings my current total estimate down to $38,020.

Grid connection: For better or worse, not now. The power company and unhelpful neighbors have made it quite costly to get tied to the grid. IOW the only pole I can access is about half a mile away and it's route would put poles in front of a great view of a neighboring house. To bury it would come at a high cost (about half of what I expected the system to cost) so for now, I'll be off-grid.

That question does bring up a point that it seems wise to evaluate the increased cost to make the system capable of being grid tied in the future. Or, at least not require a massive re-design. We never know if one of the neighbors changes their position on not wanting to provide an easement for access, a tie-in, guy wires, etc.

Thanks for your help.

 

[MichaelK]

Quite a few of your numbers are way off, some too high, and others too low. Let's start with the panels. Do NOT pay retail for panels. Shop on Craigslist for cash and carry deals. Here's an example of someone within driving distance that is offering quality panels for ~2.5W/$.

New REC 290 watt Solar Panels - electronics - by owner - sale - craigslist

New REC 290 watt solar panels with black frames. $110 each. Few remaining for sale. We also have other brand / wattage panels and related equipment in stock. Click "more ads by this user" button...

houston.craigslist.org

I only spent about 30 seconds shopping, so you are likely to find better deals than that.

You are not going to get a high-quality controllers for just 300$ Midnight200s are ~600$ right now. I think though that you will just need two or maybe three of those.

Again, 1600$ is not going to get you a quality sine-wave inverter. Look at Schneider's XW6848. It's running about 3800$ right now. Without whole-house air-conditioning, or a well-pump, just one 6848 will be fine. For really big loads, larger than what you are talking about, the 6848s can be paralleled. It's native 120/240V, so you don't have to have two of them.

A second area that you'll be able to save money is the ground mounts. Making them myself out of schedule 40 pipe, and unistruts, I finding my rotating arrays cost me about 400$ in steel. Here's a pic of my V4.0 design. Although it is right now handling 4 large residential panels, my up and coming V5.0 design can handle six. I'll have pics of that ready in another week or so. Four of these arrays, spaced about 20' apart, each with six of those 290W panels, will make 30kWh per day. You might get by with just three.

You do not need 2 gauge wire from the panels to the electronics. Wiring three of those panels in series, you can run power at 90+V with little voltage drop. 10 gauge will be fine. Wire each array 3S2P. Then 4 parallel strings go off to two controllers. Midnight and Outback make outstanding charge controllers that work in that range.

Take a good look at this 4V Rolls-Surrette battery. Twelve of these cost me only 7400$ total. https://www.rollsbattery.com/battery/4-cs-17p/

 

[Doggydogworld]

This outfit in Pearland sells pallets of panels at ~50 cents/W. You can find cheaper if you hunt (and travel), but the best deals are usually for odd lots (e.g. 9 panels left over from a job).

$180/kWh sounds cheap for battery. Remember, you need 180 kWh of lead acid for 90 kWh usable. Or 112 kWh of LiFePO4.

Without a grid connection it's better to downsize the battery and put the money toward a decent generator. You don't want to use it often, but you also don't want to pay for batteries you only need a few days a year. It's a balancing act.

Will the new house have A/C? For 2400 sqft in Brenham that typically draws more than 30 kWh on summer days. Not unusual to see ~100 kWh a day, almost all A/C. Aggressive house design can help, as can zone cooling, setting thermostat to 80 degrees, etc.

 

[chrisjx]

Quite a few of your numbers are way off, some too high, and others too low. Let's start with the panels. Do NOT pay retail for panels. Shop on Craigslist for cash and carry deals. Here's an example of someone within driving distance that is offering quality panels for ~2.5W/$.

New REC 290 watt Solar Panels - electronics - by owner - sale - craigslist

New REC 290 watt solar panels with black frames. $110 each. Few remaining for sale. We also have other brand / wattage panels and related equipment in stock. Click "more ads by this user" button...

houston.craigslist.org

I only spent about 30 seconds shopping, so you are likely to find better deals than that.

You are not going to get a high-quality controllers for just 300$ Midnight200s are ~600$ right now. I think though that you will just need two or maybe three of those.

Again, 1600$ is not going to get you a quality sine-wave inverter. Look at Schneider's XW6848. It's running about 3800$ right now. Without whole-house air-conditioning, or a well-pump, just one 6848 will be fine. For really big loads, larger than what you are talking about, the 6848s can be paralleled. It's native 120/240V, so you don't have to have two of them.

A second area that you'll be able to save money is the ground mounts. Making them myself out of schedule 40 pipe, and unistruts, I finding my rotating arrays cost me about 400$ in steel. Here's a pic of my V4.0 design. Although it is right now handling 4 large residential panels, my up and coming V5.0 design can handle six. I'll have pics of that ready in another week or so. Four of these arrays, spaced about 20' apart, each with six of those 290W panels, will make 30kWh per day. You might get by with just three.

You do not need 2 gauge wire from the panels to the electronics. Wiring three of those panels in series, you can run power at 90+V with little voltage drop. 10 gauge will be fine. Wire each array 3S2P. Then 4 parallel strings go off to two controllers. Midnight and Outback make outstanding charge controllers that work in that range.

Take a good look at this 4V Rolls-Surrette battery. Twelve of these cost me only 7400$ total. https://www.rollsbattery.com/battery/4-cs-17p/

Another user, @Doggydogworld chimed in about an ebay listing from Pearland. Might be a similar posting.

We are planning to build a SIPs structure and use probably 5 or 6 mini-split w/ heat pumps. We will also have a well. So we're hoping the high R factor of th SIP design helps with cooling/heating. I was also looking at Days of Required Autonomy: 3. I'm not sure if this is how that gets calculated but we're looking at 45 x 400W panels to fill up 90kWh (9 x 10kWh) battery system for 3 days of excess or lower than expected sunny days.

For the controllers and Inverters you seem to be quoting numbers that I don't quite understand. For example, I was under the impression that it is good to parse out components more or less into sub segments and break the system down for redundancy. So one would have x number of panel strings, x number of controllers, and x number of inverters so that if sub segments fail, the whole system doesn't fail.

The numbers will change, but my initial idea was for 30kWh per day... so 45 panels = 5 strings of 9 panels each, so 5 controllers, feeding 9 batteries, 9 batteries feeding 3 inverters... or something along those lines.

You mention I might get away with 2 or 3 controllers and 1 or 2 inverters admittedly at higher capacities. Ach! I read now how you are recommending the controller setup. 10 gauge wire, each array 3S2P, 4 parallel strings to two controllers, Midnight and Outback controllers.

I'll look into your ground mount designs. Wind will be an issue as it is often very windy out on the hill at Brenham. Although a 30 degree tilt doesn't seem to be much of an angle of attack.

I was looking at a 48V system so not sure how the 4V? battery works; how many, etc.

Thanks for your help and input,
Chris.

 

[chrisjx]

This outfit in Pearland sells pallets of panels at ~50 cents/W. You can find cheaper if you hunt (and travel), but the best deals are usually for odd lots (e.g. 9 panels left over from a job).

$180/kWh sounds cheap for battery. Remember, you need 180 kWh of lead acid for 90 kWh usable. Or 112 kWh of LiFePO4.

Without a grid connection it's better to downsize the battery and put the money toward a decent generator. You don't want to use it often, but you also don't want to pay for batteries you only need a few days a year. It's a balancing act.

Will the new house have A/C? For 2400 sqft in Brenham that typically draws more than 30 kWh on summer days. Not unusual to see ~100 kWh a day, almost all A/C. Aggressive house design can help, as can zone cooling, setting thermostat to 80 degrees, etc.

The list from the Pearland person look great. I'll need to get more familiar with the type of models they're offering so I can feel confident I'm not getting less than what I expect.

Does your comment re: equivalent ratings mean 112 kWh for 90kWh LiFePO4 usable, and does that come from efficiency ratings, or hours of usable sun, or how deep we dare not go into depleting the available power in batteries... or all the above. It's an important point you make. I was also trying to build into the formula a value I heard called Days of Autonomy (I set to 3) which I took to mean how many days the system should continue to perform for x number of days with poor radiance. I also calculated usable sunlight but forgot efficiency (which I wonder if it's built into the ratings).

Also, I'm shocked by the "Not unusual to see ~100 kWh a day" but I know it's been brutal in Texas. This is a must to manage the heat. I just looked up Houston's record this summer with 45 days over 100F and 23 consecutive days. I can see getting a generator for a week or two of days, but 45 days or more days is going to burn a lot of fossil fuel in any given summer in Texas. The challenge, as you said, is a balancing act, and is to find the point of solar power that can reasonably be counted on and how much gas generating needs to be used.

We do plan to use 5 or 6 mini-splits with heat pumps throughout the structure and to build the house using SIPs. We hope that helps with the cooling/heating.

Thanks for your advice.
Chris.

 

[Hedges]

48V / 2V = 24 batteries
48V / 4V = 12 batteries
48V / 6V = 8 batteries
48V / 12V = 4 batteries.

Generally all about 120 lbs each. If you go 2V, you get more capacity out of a single string.

The Rolls-Surrette are particularly long life FLA. Need watering and occasional equalization.
I'm using AGM 6V 405 Ah, for a smaller bank, higher cost per Wh of cycle life but low maintenance.

People who use LiFePO4 generally love it, say it is much better than lead-acid. Including, they get back more of the power they put in charging.

There are getting to be some bigger lithium batteries. The server rack type used 100 Ah cells so 5kWh, but Will reviewed a different type with 300 Ah I think.

 

[chrisjx]

Another question from above... RE: Days of Autonomy?

If I am looking for 30 kWh / day system usage with a 3 day autonomy (in case of, for example, 3 super cloudy days, ), I think I'd need a 90 kWh battery capacity to go for 3 days. To what degree do the other systems like solar panels, controllers, and inverters need to be capacity increased to keep the batteries topped for 3 days of usage. And what about recovery?

Is this even a thing?

Thanks for any insights.
Chris.

 

[Hedges]

Depends on how long until the next 3 days without sun

 

[Hedges]

But for lead-acid, try to hit the optimal charge rate. Might be 0.13C (in addition to what loads draw.)

 

[Balthazar-B]

We do plan to use 5 or 6 mini-splits with heat pumps throughout the structure and to build the house using SIPs. We hope that helps with the cooling/heating.

It's good that you're designing your new residence to be as temperature-stable as possible, and given it's new construction, it should be possible to make it very tight, well-insulated, and efficient.

Unless I missed it, you didn't specify how your new home would be heated, i.e., whether you'll have wood or propane. Either way, you did mention you wanted energy use to be as low as possible.

5 years ago it might have seemed an outlandish suggestion for Texas, but with the possibility, if not likelihood that you'll be experiencing more extreme climate conditions going forward, in both summer and winter, have you explored the idea of using hydronic/geothermal HVAC for both heating and cooling? This would be extremely efficient, albeit with higher upfront costs, and is not uncommon in extreme climates such as Australia. Since you'll be sinking wells and have a lot of ground area to work with, which is going to be excavated anyway, it could be worth adding, and be most useful when your temperatures hit their extremes, especially in winter during periods of zero energy generation. It would provide an additional margin of safety, and perhaps change the amount of battery storage you'll require from the get-go.

 

[Doggydogworld]

Does your comment re: equivalent ratings mean 112 kWh for 90kWh LiFePO4 usable

Rule of thumb is you only deplete lead acid 50% to preserve battery life, and 80% for LiFePO4. I don't think it hurts to go near zero once in a while with LiFePO4, but opinions vary.

Also, I'm shocked by the "Not unusual to see ~100 kWh a day" but I know it's been brutal in Texas. This is a must to manage the heat. I just looked up Houston's record this summer with 45 days over 100F and 23 consecutive days. I can see getting a generator for a week or two of days, but 45 days or more days is going to burn a lot of fossil fuel in any given summer in Texas.

Yeah, I'm near San Antonio which had 74 days of 100+ this year (so far). Last year was 57 days. All time record was 59 and only a few years in history were above 30 or so. There's been a definite change.

Solar gain is often a bigger problem than temperature. Insulation doesn't really help with solar gain. My A/C units don't run nearly as often on the rare overcast 100 degree day as on a sunny one. And that's I have radiant barriers in the attic plus white shutters that reflect sunlight plus brick walls. I'd heavily investigate solar rejection techniques if building off-grid. A partial list:

Build in the shade - problem solved!
Avoid unshaded east-facing and especially west-facing windows
Use overhangs to shade south facing windows in summer only (not hard, summer sun is almost directly overhead at these latitudes)
Buy windows with low-E coatings
Install shutters or seasonal solar screens
Radiant barrier in roof
LIght colored roof (or maybe metal, quite popular in TX)
Reflective exterior walls on east, west and maybe south side

Some of these things hurt wintertime solar gain, when you want it, but things like removable solar screens don't.

The good news is solar production correlates terrifically with solar gain. A stretch of cloudy days means low production but also lower A/C demand. Use a calculator like PVWatts to estimate seasonal output variation. For Brenham it shows ~5 hourJuly average daily net production at 20 degree tilt vs. ~3.3 hours in December at 40 degree tilt.

If you start smaller but design for expansion you can just run the generator extra hours the first year while collecting data on daily variation. Then use that data to decide what makes sense in terms of more panels, more battery, etc.

Ceiling fans are pretty standard in TX. Highly recommended.

We do plan to use 5 or 6 mini-splits with heat pumps throughout the structure and to build the house using SIPs. We hope that helps with the cooling/heating.

The big win with mini splits is you can dial them way back in rooms you don't use. Works even better if you insulate interior walls.

Also look into geothermal heat pumps. They aren't cheap, but can slash HVAC energy usage. It was too expensive for me due rocky soil, but Brenham's geology should be more favorable.

Do you have a backup heat source? Really cold, cloudy snaps are rare, but cause heat pumps to suck power.