35,000 SF Commercial Office Building

[mciholas]

My company building is 35,000 SF single story located in Newburgh, IN as shown here:



After looking into residential solar for my parents in Florida, I've started to think what I can do to my building, so I am perhaps going slightly mad thinking about solar.

The building is 3 phase 208 VAC 800 amp service (288 KW) but we use a tiny fraction of that. All the lights are LED. The HVAC system is 9 ground source heat pumps. The building is well insulated such that we have mostly cooling loads so my usage is higher in the summer than in the winter. I heat and cool the entire building presently for about $250 in electricity per month (our electric bill is about $1200 month due to the other loads, I have a private meter on the HVAC panel just for curiosity). There are 4 miles of geothermal well pipe under the building (32 wells, each 300 ft deep).

Most electricity is used during the day time and during working hours (most employees have one laptop as their main computer and take it home, fewer have desktop PCs they leave running).

The basic idea would be to install one or more Sol-Ark 30K units which are 3 phase 208 VAC 30 KW (the 60K is 480 VAC so wouldn't work). The first unit would go on the subpanel that feeds the server room and thus serve as a big UPS for them. With ~35 KW of panels and perhaps 50 KWH of batteries, that might actually come close to meeting our local needs. A second unit could be on a different subpanel. I don't think I want to invest in the switch gear to intercept the 800 amp service. I don't really need backup power (thought the server room would be nice), it is mostly about grid up local generation.

We pay a "demand charge" which is a significant part of our bill. It is a charge per KW for the highest demand we've had in a 5 minute period in the last 3 months. So if by chance all the HVAC units line up and we're running a big machine (the test temperature chamber is the biggest load we got at about 5 KW), we can peak up and that costs us money over a long time.

I've got the land to do ground mount, but that big clear roof is really where the panels should go. They would be flat, unfortunately, but the cost for the structure to tilt them up would outweigh the cost of just having more panels. The cost of a ground mount would also favor the roof. I do hate having to put RSD devices on each panel, though. No need for optimizers, nothing to shade the panels. The roof is 1:12 slope with ridge going north/south, so splitting the pane string to each side of the ridge would be reasonable.

The local utility has instantaneous net metering and pays a fraction of retail on grid export, so the goal is to make use the power locally and only export when battery if full and local loads are satisfied. The load curve is favorable for solar, most usage is during daylight, so I hoping the battery size doesn't get too big. I'll have to model that.

Are there other options besides the Sol-Ark 30K for this job? I like that it is 3 phase and not having to parallel 3 single phase units, and I like that it is high voltage for the battery (48 volt current would be extreme). What else exists that can do the job? I figure this is about a $100K job to do right, so quality stuff is important over low cost. Payback maybe in 8 to 10 years.

I realize I am asking other inmates of the asylum if I am insane, but what do you think about making my building solar powered?

Mike C.

 

[wpns]

Sounds good, the nice thing about solar is how modular it is. Put up 10KW and 10KWHR every year (*) until you’ve covered your own consumption.

(*) representative numbers, will depend on chosen equipment. Just start small and plan ahead for expansion.

 

[Brucey]

May want to check what SMA have in that sizing as well, their systems can scale up pretty well from what I've seen

 

[mciholas]

Sounds good, the nice thing about solar is how modular it is. Put up 10KW and 10KWHR every year (*) until you’ve covered your own consumption.

I'd like to avoid a hodge podge system of various mixed inverters, panels, etc. Also, going through the permit, install, inspection cycle multiple times will be a pain, so let's design it the right size and get it all done neatly.

Mike C.

 

[wpns]

I'd like to avoid a hodge podge system of various mixed inverters, panels, etc. Also, going through the permit, install, inspection cycle multiple times will be a pain, so let's design it the right size and get it all done neatly.

Mike C.

So don’t do mixed. Do one integrated system and then add on idential components when you expand. Will make permitting easier as well.

 

[mciholas]

May want to check what SMA have in that sizing as well, their systems can scale up pretty well from what I've seen

The 3 phase SMA inverters I found had no battery capability. Given my net metering rules and demand charges, that's a non starter.

I can find single phase units which I can triple up, usually using 48 VDC battery systems, say 3 X 18KPV, but the extra wiring is not for nothing and you are better off with the purpose made 3 phase unit.

This won't be a "tinker" system, it needs to to go in and just work without a lot of management or tweaking.

Mike C.

 

[mciholas]

So don’t do mixed. Do one integrated system and then add on idential components when you expand. Will make permitting easier as well.

The solar equipment market is moving pretty fast and future availability of any of the components in 2 or 3 years is unknown.

I don't need to do this incrementally for cost reasons, so let's put in the right system from day one. If it happens to be a bit too big or a bit too small, that's okay, we got close enough.

The only thing I think I might adjust in the future is the size of the battery bank. You can never have too much of that and those are consumable items regardless.

One thing I am not sure of is whether the Sol-Ark 30K is adaptable to 3rd party batteries. Sol-Ark sells a battery rack that looks neat, L3-30K-HV, but it is 40 KWH for about $20K, $500 per KWH which is expensive. I'd rather go with more economical batteries and have more of them, if they are compatible. Sol-Ark 30 requires battery comms to work, and it is a high voltage system (batteries in series) which presents some issues that need to be taken care of (such as no exposed battery terminals, good electrical isolation of battery comms, and dealing with leveling between rack units).

Mike C.

 

[Hedges]

What is your cost per kWh? That will show what payback you get from PV.

SMA makes 7.7kW battery + PV inverters (available about now), 11.4kW expected end of the year. HV batteries requited are more expensive than 48V ones.
SMA makes 2.2 MW battery inverters. Transformers used to step up (or down), but likely too large for your needs.

Discover the SMA Battery Inverters! | SMA America

Discover the SMA battery inverters ► Self-sufficient power supply ► Optimum integration of renewable energies ►Stabilizing of utility grids.

www.sma-america.com

What is your peak demand, which you would like to shave?
Can you knock some AC offline for a while to shave peaks?

"With ~35 KW of panels and perhaps 50 KWH of batteries, that might actually come close to meeting our local needs."

Producing kWh needed (with net metering) is one question.
Producing kW peak to avoid demand charges is a different one. How many kWh are needed to shave peaks and move to valleys?

You can either use 3-phase or multiple 2-wire 208V inverters. Depending on your kW and kWh needs, several 48V batteries could be just fine. The EG4 PowerPro looks attractive, 14 kWh in a weatherproof package.

SolArk, EG4, SMA, Schneider, Tesla, others ... a number of options for AC coupled batteries.

Also see what Current Connected offers - they have a weatherproof SOK enclosure and sell several brands.

 

[wpns]

One thing I am not sure of is whether the Sol-Ark 30K is adaptable to 3rd party batteries.

If your AHJ wants UL 9540 (not 9540A) you are going to have to do some digging. I know the EG4 18Kpv plus PowerPro ESS does that but now you are into three separate inverters(*) which you said you’d like to avoid.

The other advantage of the PowerPro is that it’s new, so you have some hope of expanding for the next few years.

(*) are your heavy loads split across all three phases or can you do ESS on a single phase and shave the peak using a single-phase system? #JustBrainstorming

 

[mciholas]

Further details:

Typical monthly usage is 9000 KWH. Varies from about 7000 KWH to 11000 KWH. Summer is usually higher, but not dramatically so. This is 300 KWH per day. I don't yet have data on time of day profile, working on that, but it will be higher during work days and between 0900 and 1800 local time. The HVAC system uses an average of 2500 KWH per month so out seasonal variations are not extreme relative to the other loads.

Demand charge peaks are typically 22 to 32 KW. This is the highest 5 minutes of demand during the month. Considering we have 288 KW service, we hardly peak at 10% of service rating. This is due to the low power usage we have from our building design, LED lights, etc. We really don't need the 288 KW service, but the cost to put in less was not much less, so it doesn't matter.

My charges, typical per month:

Meter charge: $15
Usage: $1260 ($0.14/KWH)
Demand: $125 ($5 per peak KW)

Average month: $1400, ~$17,000 per year.

Net metering is about 25% and is instantaneous. I pay for every KWH from the grid, they pay 25% for every KWH pushed into the grid.

I'm going to guess our typical during the day usage is about 16 KW and about 10 KW all other times. It looks like I need about 100 KWH of battery to survive the typical night. That won't get me through a cloudy day, though. I will also assume a typical sunny day is 4 hours of effective insolation.

Conceptual system:

96 panels, 400 watts, in 8 strings. Assuming $100 per panel, $10K.

Sol-Ark 30K. $15K.

120 KWH battery bank. Assuming $300 per KWH, $36K.

Misc racking, wires, mounts, etc: $15K.

Materials run about $76K, so maybe a $100K system all in with labor.

If it cuts my utility bill by 80% (allowing cloudy days, lower sun in winter, etc), the pay back period is 7.5 years. if I can push enough excess into the grid, maybe get closer to 100%?

Given my peak demands haven't gone over 32 KW, a 30 KW inverter sure seems like a decent fit. The issue might be I am short on panels and the Sol-Ark is maxed out. I might not be able to pump 120 KWH into the battery in a single sunny day.

What I really need is a 50 KW array that charges a 120 KWH battery and then a 30 KW inverter. Maybe another 20 KW array and a charger could be added to the Sol-Ark 30K system?

Mike C.

 

[mciholas]

If your AHJ wants UL 9540 (not 9540A) you are going to have to do some digging. I know the EG4 18Kpv plus PowerPro ESS does that but now you are into three separate inverters(*) which you said you’d like to avoid.

It is an option none the less.

(*) are your heavy loads split across all three phases or can you do ESS on a single phase and shave the peak using a single-phase system? #JustBrainstorming

Our loads are balanced among the phases. HVAC are 208 VAC single phase, so they straddle the various phase pairs. Most everything else is 120 VAC single phase and spread across the building. No one phase is really loaded more than the others.

Mike C.

 

[wpns]

Net metering is about 25% and is instantaneous. I pay for every KWH from the grid, they pay 25% for every KWH pushed into the grid.

So if you oversize your array by 4X and do "The Grid Is My Battery", you can eliminate the need for batteries entirely(*) and null out your bill? That might be a good starting point! Batteries are a lot less than they used to be, but I can only imagine battery prices will continue to drop...

(*) If your grid is stable and power failures are infrequent, this might be worth contemplating.

 

[wpns]

Usage: $1260 ($0.14/KWH)
Demand: $125 ($5 per peak KW)

So your Demand charge is only 10% of your usage charge, that makes KWHR the low-hanging fruit, and Demand much harder to get payback for (because it requires batteries, which are the expensive part).

 

[mciholas]

So if you oversize your array by 4X and do "The Grid Is My Battery", you can eliminate the need for batteries entirely(*) and null out your bill? That might be a good starting point! Batteries are a lot less than they used to be, but I can only imagine battery prices will continue to drop...

That is a thought.

I'd have to check with the utility. Sometimes the KWH balance is limited to what you draw, so you can't shove substantially more KWH into the grid than you use. In that case, you can only get about 25% of that back. Sometimes the money credit can't be positive, you can't get paid by the utility, so you don't want to shove so much power you aren't get paid for.

This sort of strategy is dependent on the net metering rules, and thus it makes it fragile to rule changes. I will investigate since panels plus grid tie inverter is a lot cheaper.

(*) If your grid is stable and power failures are infrequent, this might be worth contemplating.

A large part of my parent's needs are backup power for nuisance trips and for hurricane events (in Florida). That is only a small part of the issue for my building where power is relatively stable (maybe one outage per year, typically less than 2 hours). I would like a little bit of backup for my server room, but I'm not backing up the entire building. Backup is thus not a major factor, but would be nice to have on one subpanel that feeds the server room.

Mike C.

 

[wpns]

I would like a little bit of backup for my server room, but I'm not backing up the entire building. Backup is thus not a major factor, but would be nice to have on one subpanel that feeds the server room.

That's the point where I'd look at running the server room off a single phase, so you don't need a 3phase inverter with batteries

 

[Supervstech]

You certainly can build a system to cover current and future needs of the building… BUT…
Many states have solar farm programs, where you produce, and the power company pays you income on the system.
If you have acres of land to dedicate, hire a company to pitch a solar farm and generate serious income… keep the roof for your building needs.

 

[Hedges]

Typical monthly usage is 9000 KWH.

Average month: $1400, ~$17,000 per year.

Net metering is about 25% and is instantaneous. I pay for every KWH from the grid, they pay 25% for every KWH pushed into the grid.

$1400/9000 kWh = $0.16/kWh

Don't bother with PV.
Turnkey install would make power for $0.10/kWh (amortized over 20 years, ignoring time value of money)
DIY, free labor, hardware only costs $0.025/kWh (over 20 years.)

I'm sure you've got other things you can do with your time which will earn more for your business.


You can consider battery inverter to shave peaks. But besides height of peaks, need to consider how much kWh to ride through them until next valley.

Don't bother with getting through the night on battery (unless you want backup due to power failures). Just shaving peaks.

Some LiFePO4 batteries are claimed to last 6000 cycles. That's 16 years of one cycle per year. They may support 1C discharge, maybe only 0.5C. That means to shave a peak lasting 15 minutes, you need battery capacity 4x or 8x greater than consumption during that time. If you shave 4 or 8 peaks per day, might get your value out of the battery in this lifetime. If peak is 1 or 2 hours, then a better fit even shaving just 1 peak.

Batteries can cost $0.05 to $0.10 per kWh of cycle life. Not a good deal compared to your utility rates (refer to my comment not trying to operate through the night offgrid), but maybe avoiding demand charges they can be worthwhile.

So your Demand charge is only 10% of your usage charge, that makes KWHR the low-hanging fruit, and Demand much harder to get payback for (because it requires batteries, which are the expensive part).

In other words, don't bother with batteries and eliminating demand charge.
Pay the utility and focus on your business.

 

[12VoltInstalls]

don't bother with batteries and eliminating demand charge.
Pay the utility and focus on your business

Having ‘some’ battery as a buffer has advantages; overnight may make far less economic sense as you state. But covering the server room he mentioned and carrying the A/C for the 2-hours/year may be an attractive consumer choice.

Beyond that, it doesn’t sound as if the RTU’s are running under a software management platform. That may actually mitigate the peak demand and enable a predictable rolloff of that portion of the billing.

Then one can determine if adding batteries DOES actually make economic sense if the RTU aka a/c management is programmed to be ‘aware’ of the peak demand mitigation goal.

Pay the utility and focus on your business

I thought you were an EE but that’s pretty good management advice right there.

 

[kscessnadriver]

I have zero experience with them, but Growatt US has a commercial inverter listed, that might do what you want.

WIT 28-55K-HU/AU-US L2 | Commercial Storage Inverter | Growatt

WIT 28-55K-HU/AU-US L2 | Commercial Storage Inverter | Growatt

us.growatt.com

 

[mciholas]

So if you oversize your array by 4X and do "The Grid Is My Battery", you can eliminate the need for batteries entirely(*) and null out your bill?

I did some research and this is how they tariff export (outflow):

During the Month, Company shall measure the total kWh amount of Inflow and the total kWh amount of Outflow.

The Inflow kWh for the Month shall be billed in accordance with the Customer’s standard Rate Schedule, with all applicable rates and charges (heretofore defined as Standard Charges).

The Excess DG kWh (Outflow) for the Month shall be multiplied by the Marginal DG Price to determine the Rider EDG Billing Credit.

For each Month, the Customer will be billed the Minimum Monthly Charge as defined in the Customer’s applicable Rate Schedule. If the portion of the Customer’s bill for the Month attributed to the Rider EDG Billing Credit is in excess of the amount attributed to Standard Charges less the Minimum Monthly Charge, the amount in excess will be accumulated in a Rider EDG Billing Credit Balance for use in a subsequent period.

If the portion of the Customer’s bill for the Month attributed to the Standard Charges is in excess of the Rider EDG Billing Credit, any remaining Rider EDG Billing Credit Balance will be applied until the bill becomes the Minimum Monthly Charge or until the Rider EDG Billing Credit Balance becomes zero.

In accordance with IC 8-1-40-18, when Customer discontinues Rider EDG service and no longer receives retail electric service from the Company at the Premises, any unused and remaining Rider EDG Billing Credit Balance will revert to Company.
What is interesting is that demand charge is part of the Rate Schedule, so it seems I can over produce and cancel those charges if I want to.

If I don't care about backup power, then the simplest system is a straight grid tie and then produce power. Whatever I produce will offset local loads first, then I get credit at some rate. With no battery, and with a simple grid tie inverter, the cost of the system is way less. This can be done with microinverters, but that's expensive.

Something like this would work:

https://signaturesolar.com/content/documents/GROWATT/1519073-specs.pdf

Growatt 10KW for $1300.

Need at least 3 to balance the phases. That would be 60KW of production for $8K in inverters (6 units). $20K in panels (180 panels of 400 W each) would be 72 KW of PV. That won't clip due to the flat roof angle. The PVWatts estimator says I will get about 86 MWH per year from that array. My annual usage is about 108 MWH, so that won't cover it, but maybe it covers half my bill (accounting for the fact much of it will be export at low value). A $40K system would save me $8K per year, or have a pay back period of 5 years. A system that is 9 inverters would make 130 MWH per year and might save me $12K per year, so same payback period but making more energy.

The utility really controls the economics of this a great deal.

Mike C.