Big 10k installation questions:

Nila · May 2013

Hi All,

I've setup some 2kw/3kw range projects for both personal/commercial clients here locally.

Now is the time to step up and do a 10KW installation for a commercial office.

Note we did not purchase anything yet.

Panel specs : 250Watts
30.72 - VMP
8.14 - IMP
37.8 - VOC
8.63 - ISC

Panels: 40

CC:
4 x 150 MPPT xantrex cc

3 in series / 4 parallel = 3000watts ,
3 in series / 4 parallel = 3000watts,
3 in series / 4 parallel = 3000watts
3 in series / 1 parallel = 750 watts ( we can extend this system to add more panels later on )

Wires: I believe UV coated 4MM for inter-panel connections
and then we can use 16MM for panels to CC and CC/Batteries and same for battery connections.

Panels and bank stay within 10 Meters.

Battery : 16 x 200AH 12V batteries only thing that is available here.
800AH at 48V.
4 parallel strings.

This is a day only office hence we need these batteries ONLY to support being a buffer for the day usage.
They wouldnt even use 50% of this battery bank whole day..

Inverter : 2 x 6048 xantrex. and
no I am not getting that distribution boxes,etc.. I have a local company who would make all the panels,etc for lower cost and is approved/certified.

Loads: they have close to 10KVA of loads but most of them are not switched on simultaneously..
We believe the total wattage required all the day would even drain 50% of bank at no sun.

We do not know the exact loads yet as its a new building but rough calculation was like around 8000-10000 watts if every fan and light and most computers were on .
I would actual say it would be far lesser may be like 6-7kw max.

Grid: is available but very faulty and very inconsistent and they need to run everything off the sun whenever possible.

Support : we have support available from Xantrex for inverter/battery installation and recommendations.. same thing for Charge controllers so they would send their personnel to help setup most for a cost.

Questions:

1. Bank: is battery bank TOO low? is that low like it would be waste of panels .. or is it like low enough to kill batteries with over charge?
I wouldn't care of its like we have too much panels.. as this system will have potential to extend more/ add more batteries.
I would be concerned if they would kill batteries.

I think theoretically we should get about 150AMPS minus CC/wire/panel losses. This should require atleast 1500 amps battery bank. however I do not trust these panels to deliver the same output.. and they wouldn't spend more on batteries

Should we add like 1 more strings.. 4 strings already is looking bad when reading the other posts around.

2. Disconnects..
a. panels to the CC's . I believe we need one each for each CC.
b. battery to CC ? or to inverter? would xantrex have any sort of disconnect in their 6048, i hope no.

Also would love to know what ratings is safe to buy.

3. Fuses. What sort of rated fuse should we use on the battery bank parallel strings? and how many.
These fuses are supposed to save the battery wiring and battery right? when in case of short or something/?

We can make our own combiner boxes with fuses for the panel strings.. I think 15A DC fuse would do there.

4. Safety/ considerations before beginning to under take this project. Most of the work would be done by the professionals ..
we would just co-ordinate to get this done .

Although this is a 10KW installation.. I tried to set this up like 3x 3KW kind of setup so it is safer/easier to work with for us. Also makes it flexible for future changes.
GridTied setup would have been much better for this kind of wattage.. but crappy grid makes it impossible.

Please help answer these questions.. Please yell at me if i didn't explain something clearly and English is my 2nd language so excuse grammar/etc.

Regards

NorthGuy · May 2013

Re: Big 10k installation questions:

There must be some mistake here.

7kW load will drain the batteries to 50% SOC in 2 hours, or it'll drain them completely in 4 hours.

Nila · May 2013

Re: Big 10k installation questions:

NorthGuy wrote: »

There must be some mistake here.

7kW load will drain the batteries to 50% SOC in 2 hours, or it'll drain them completely in 4 hours.

Its not a big deal as we have the EB Grid as backup and we would use Sun power ONLY when the batteries are full . I think Xantrex has the export grid facility.
And our power cuts here wont last more than 3 hours at the stretch.

Also the 7KW load is very very approx.. and should be even lower office time is 9-5 and that time Sun must be available all around the year and grid is available whenever batteries go below 60%.

NorthGuy · May 2013

Re: Big 10k installation questions:

Nila wrote: »

Also the 7KW load is very very approx.. and should be even lower office time is 9-5 and that time Sun must be available all around the year and grid is available whenever batteries go below 60%.

For that sort of system with intermittent outages and skewed usage pattern you need to be very careful with sizing because it is very easy to undersize or oversize. This is not an easy task. Hour-by-hour simulations of system performance in various scenarios may help.

Did you tell your clients that your system will fail if there's 4-hour grid outage on a cloudy day?

BB. · May 2013

Re: Big 10k installation questions:

Hi Nila,

I am not sure exactly what to do here. Without the loads known, all we can do is take the numbers and estimate its performance and tell the owner what it will do. That way, they will not be surprised if they try to get more from it, or are concerned about price/cost of system.

Since the system is grid tied--I would design it as a normal backup system with battery bank and AC battery charger(s). And then add solar panels as it make sense for the customer. At this point, is the solar array to reduce power costs or provide longer backup power for extended outages? And will there be a backup generator?

By the way Nila, you probably have several systems working now. How are they working out for you? Meeting your expectations/needs? Over/under designed? I would guess you are pretty happy so far as you are getting pretty deep into these new designs.

Starting with the pair of XW 6048 units. Personally, I prefer to have one battery central battery bank and power the pair of XWs. You don't get into needing to balance power between two independent systems (one has more load than the other).

The two rules of thumb that we use for sizing the battery banks for the inverter is A) the average kWH per day/session,

the maximum continuous power, and C) the surge power. So, lets pick some numbers.

Say 7kW peak average loads, and 3 kW "average", and 14 kW maximum surge (starting pumps, A/C, water pumps, etc.). By the way, this does raise the question of a pair of XW's vs one XW and some conservation. Again, this is more of a cost sensitive issue vs reserve power/functionality. If you can get the maximum sustained loads ot 6 kW or less, that could be a cost/complexity savings here. Otherwise, I will just do the math ignoring the one/two XW questions.

First is A) sizing to load. 1-3 days of backup power, 50% maximum discharge. 2 Days of backup seems to give a nice capable system without too much capacity/costs for battery bank.

3,000 Watt average load * 8 hour per day power loss * 1/0.85 inverter efficiency * 1/48 volt bank * 2 days backup * 1/0.50 maximum discharge = 2,353 AH @ 48 volt battery bank.

, maximum draw assuming 7 kW and a C/8 (8 hour discharge rate) would be the limit that could over heat the battery bank (generic flooded cell deep cycle batteries). We use the 20 Hour Rate for battery AH capacity numbers, an 8 hour rate will knock 20% or more off the capacity numbers--But obviously, the battery bank would close to 3 hours of energy at maximum design load of 7 kW:

7,000 watt * 1/0.85 invtr eff * 8 maximum current draw * 1/48 volt bank = 1,375 AH @ 48 volt battery bank minimum based on maximum draw

C), maximum surge current of 14 kW and C/2.5 maximum draw for flooded cell battery ensure battery does not collapse under load and varying states of charge (i.e., meets design surge at 100% SOC but not a few years down the road at 50% SOC and some aging on the battery):

14,000 Watts * 1/0.85 invtr eff * 2.5 maximum surge * 1/48 volt battery bank = 858 AH @ 48 volt battery bank.

So, your 800 AH @ 48 volt battery bank would be a limiting factor here... Of course, the rules of thumbs here are designed for an off grid system--which this is not. And, I would guess there is a backup genset available/outages are usually only a few hours at at time. So, using your bank and the above rules, such a bank would perform as (note, for real life, you should use the try bank capacity at the XX hour discharge rate--i.e., if the 3kW ends up being ~4 hour discharge rate, then use the 4 hour estimated battery capacity to give hours of support to the customer). Note, these are still conservative numbers.

800 AH * 48 volt battery bank * 1/0.85 invtr eff * 1/3,000 watt average * 50% max discharge = 7.5 hours of load support

And guessing that the ~15 hour battery bank capacity is 700 AH @ 15 hour discharge rate (just making numbers up here as an example):

700 AH * 48 volt battery bank * 1/0.85 invtr eff * 1/3,000 watt average * 50% max discharge = 6.6 hours of load support at re-estimated bank capacity and 50% maximum for long(er) battery life.

The estimated recommended peak power output (C/8 discharge rate):

800 AH * 48 volts * 0.85 ivntr eff * 1/8 hour discharge rate = 4,080 watt maximum continuous output

The maximum surge power (C/2.5 discharge rate):

800 AH * 48 volts * 0.85 ivntr eff * 1/2.5 hour discharge rate = 13,056 watt maximum continuous output

You and your customer will have to decide what number(s) is/are "important" to them. But, at this point, I would suggest that an 800 AH @ 48 volt battery bank will serve a single XW 6048 at this point.

Break into a second post---To be continued.

-Bill

BB. · May 2013

Re: Big 10k installation questions:

OK, we have an 800 AH @ 48 volt battery bank for now. Lets look at charging it from AC and Genset.

Our normal rate of charge is ~5% to 13%, and, perhaps up to 25% if there is thermal management for the battery bank (hot batteries require lower charging voltage, a remote battery temperature sensor feeding back to the battery charger for safety--and you should monitor/alarm the battery bank for over temperature (say 110-120F or 43-49C maximum).

The XW's have a very nice AC battery charger in them... I don't remember the exact specifications on efficiencies, but I will assume some pretty high efficiency numbers here (I have to let you have some of the fun here reading the manuals

).

Assuming 90% charger efficiency and 0.95 Power Factor and 13% rate of charge--pretty fast but should be well within the battery bank's capabilities. The 800 AH battery bank would draw around (note, I think a single XW 6048 has a maximum of 100 amp charge rate--so you will need to "fix" numbers below, depending on how you configure your system) (and NEC is US national electric code wiring/breaker/fuse deratings--makes for reliable installation anywhere--but your codes/hardware may be rated differently):

800 AH * 0.13% rate of charge = 104 Amperes rate of charge
800 AH * 58 volts charging * 0.13% rate of charge * 1/0.90 charging eff = 6,702 watts of AC charging power
800 AH * 58 volts charging * 0.13% rate of charge * 1/0.90 charging eff * 1/0.95 power factor = 7,055 VA to size branch circuit/generator
7,055 VA * 1/0.8 NEC derating * 1/210 VAC minimum line voltage = 42 amp minimum branch circuit rating

You can program the XW's to use less than maximum charging current if wiring/genset need to be smaller--But I would suggest 10% rate of charge is nominal and never to go under 5% rate of charge for battery health.

We can stop the design here if we are just looking at backup power. But since you talked about solar, we plough on.

So, we will have roughly four criteria to meet because this is a "day time" power usage vs our normal off grid home which charges during the day and discharges at night (assuming "off grid" solar power--since you have a genset and AC grid power, your solar array can be anything from zero to maximum wattage).

A) is the sizing the array to the battery bank.

is sizing the array to battery bank plus "average" daytime loads (need to meet battery minimum rate of charge + sustained loads to stay above 5% rate of charge)
C) maximum cost effective array + average loads (again, this is really an "off grid" design and will make the array way larger than is probably cost effective for your customer's needs).
D) maximum array to charge battery bank (cost effective)--Probably would recommend this number to start with as a "maximum" array. Could get into issues with problems during weekend use (i.e., array becomes too large for battery bank and can over voltage on weekends where there are no office loads).

First A), assumed load and hours of sun. 3,000 watt average load * 4 hours of backup per day (8 hours is probably way over kill here) and 4 hours minimum sun (use your numbers here):

3,000 watts * 4 hours usage * 1/0.77 panel+controller eff * 1/0.85 inverter efficiency * 1/0.80 battery efficiency * 1/4 hours minimum sun noontime equivalent sun per day = 5,769 Watt array (assuming charge during day and use power when sun is down)
3,000 watts * 4 hours usage * 1/0.77 panel+controller eff * 1/0.85 inverter efficiency * 1/4 hours minimum sun noontime equivalent sun per day = 4,615 Watt array (assuming charge during day and use power during the day--i.e., battery bank not cycling)

Then

, if this was a true off grid system, then we would need to both meet the battery bank's minimum 5% rate of charge and supply the loads. And we would not want too big of array either. So, the numbers would look something like this:

(800 AH battery bank * 58 volts charging * 1/0.77 panel+controller eff * 0.05 rate of charge) + (3,000 watt load * 1/0.77 panel+cntrllr eff) = 6,909 Watt array

This is a "huge" array and really depends on the customer's power usage (amount, time of day, mostly when the sun is up and shining) and what length/time of power outages would be supported (before firing up the genset--if there is one).

C) The maximum cost effective array for a battery bank--I use ~13% rate of charge--Above that rate, the battery bank gets recharged quickly and the array no longer has to supply any more power--a waste of capacity (money). So, using the equation from

with the rate of charge raised to 13%:

(800 AH battery bank * 58 volts charging * 1/0.77 panel+controller eff * 0.05 rate of charge) + (3,000 watt load * 1/0.77 panel+cntrllr eff) = 11,733 Watt array maximum

That gives you a maximum cost effective array size...

D) We talk about battery banks not being able to supply enough current for loads (C/8 continuous, C/2.5 surge). It turns out that a battery bank can also have problems sinking too much current when they are full. And MPPT charge controllers sometimes can have "issues" where they they "sweep" the array to measure Vmp and Imp of the the array, and they dump the "extra current" in the battery bank--which can cause a "too small" battery bank to exceed 72 volts (XW inverter's shutdown voltage).

More or less, I believe that 1,000 Watts of solar array per 100 AH of 48 volt battery bank is a good number (and also happens to be the recommendation here by several people for the maximum AC inverter rating per AH of battery bank--i.e., an 800 AH @ 48 volt battery bank can reliably source ~8,000 watts maximum continuous). And for a Hybrid Inverter like the XW where it can "sell power" to the utility (a whole new set of questions for you and your utility), that would be a good maximum array sizing too. So, the maximum array for an 800 AH @ 48 volt battery bank (I would suggest) would be:

800 AH * 1,000 Watt / 100 AH = 8,000 Watt solar array (and maximum AC inverter load)

I will stop here at the moment--Until we have some agreement on the numbers/system requirements, picking hardware is a bit early in the game.

But given your 800 AH @ 48 volt battery bank, I would suggest that your 10kW array and pair of XW AC inverters are not an optimum match at this point... Either increase the battery bank AH rating and/or reduce the expectations/hardware installation.

Nila, am I helping or causing more confusion here.

Warmest regards,
-Bill

Nila · May 2013

Re: Big 10k installation questions:

I am not sure exactly what to do here. Without the loads known, all we can do is take the numbers and estimate its performance and tell the owner what it will do. That way, they will not be surprised if they try to get more from it, or are concerned about price/cost of system.

This is a new building and hence we are not able to get full load data, From what i have seen there, continous load will be just 3-4KW however occasional surges will be there like Xerox machine/ Printer . There is no other loads like motors/pumps,etc.

Since the system is grid tied--I would design it as a normal backup system with battery bank and AC battery charger(s). And then add solar panels as it make sense for the customer. At this point, is the solar array to reduce power costs or provide longer backup power for extended outages? And will there be a backup generator?

Solar array is there for many reasons, One of them is govt requirement to use solar on commercial places .
There is generator/ faulty grid available.

Biggest reason however is to be independent of the grid and to have long backup period without using genset.

By the way Nila, you probably have several systems working now. How are they working out for you? Meeting your expectations/needs? Over/under designed? I would guess you are pretty happy so far as you are getting pretty deep into these new designs.

I think we are meeting expectations so far.
We have setup some auto circuit to cut the Grid charging on inverters(cheaper ones) all the morning times and so far every setup is holding good.
The battery is full when it is evening and loads on morning are all running straight off sun.

Starting with the pair of XW 6048 units. Personally, I prefer to have one battery central battery bank and power the pair of XWs. You don't get into needing to balance power between two independent systems (one has more load than the other).

yeah we were about to do it.

The two rules of thumb that we use for sizing the battery banks for the inverter is A) the average kWH per day/session, the maximum continuous power, and C) the surge power. So, lets pick some numbers.

Say 7kW peak average loads, and 3 kW "average", and 14 kW maximum surge (starting pumps, A/C, water pumps, etc.). By the way, this does raise the question of a pair of XW's vs one XW and some conservation. Again, this is more of a cost sensitive issue vs reserve power/functionality. If you can get the maximum sustained loads ot 6 kW or less, that could be a cost/complexity savings here. Otherwise, I will just do the math ignoring the one/two XW questions.

I would say the peek average loads would be even less more like a 5KW may be.
Reason we have a pair is because this is supposed to be a 10KW system !.. so i cannot have 1 Xantrex even if that is enough. and we dont have to worry about conservation.

First is A) sizing to load. 1-3 days of backup power, 50% maximum discharge. 2 Days of backup seems to give a nice capable system without too much capacity/costs for battery bank.
3,000 Watt average load * 8 hour per day power loss * 1/0.85 inverter efficiency * 1/48 volt bank * 2 days backup * 1/0.50 maximum discharge = 2,353 AH @ 48 volt battery bank.

, maximum draw assuming 7 kW and a C/8 (8 hour discharge rate) would be the limit that could over heat the battery bank (generic flooded cell deep cycle batteries). We use the 20 Hour Rate for battery AH capacity numbers, an 8 hour rate will knock 20% or more off the capacity numbers--But obviously, the battery bank would close to 3 hours of energy at maximum design load of 7 kW:
7,000 watt * 1/0.85 invtr eff * 8 maximum current draw * 1/48 volt bank = 1,375 AH @ 48 volt battery bank minimum based on maximum draw

C), maximum surge current of 14 kW and C/2.5 maximum draw for flooded cell battery ensure battery does not collapse under load and varying states of charge (i.e., meets design surge at 100% SOC but not a few years down the road at 50% SOC and some aging on the battery):
14,000 Watts * 1/0.85 invtr eff * 2.5 maximum surge * 1/48 volt battery bank = 858 AH @ 48 volt battery bank.

So, your 800 AH @ 48 volt battery bank would be a limiting factor here... Of course, the rules of thumbs here are designed for an off grid system--which this is not. And, I would guess there is a backup genset available/outages are usually only a few hours at at time. So, using your bank and the above rules, such a bank would perform as (note, for real life, you should use the try bank capacity at the XX hour discharge rate--i.e., if the 3kW ends up being ~4 hour discharge rate, then use the 4 hour estimated battery capacity to give hours of support to the customer). Note, these are still conservative numbers.
800 AH * 48 volt battery bank * 1/0.85 invtr eff * 1/3,000 watt average * 50% max discharge = 7.5 hours of load support

And guessing that the ~15 hour battery bank capacity is 700 AH @ 15 hour discharge rate (just making numbers up here as an example):
700 AH * 48 volt battery bank * 1/0.85 invtr eff * 1/3,000 watt average * 50% max discharge = 6.6 hours of load support at re-estimated bank capacity and 50% maximum for long(er) battery life.

The estimated recommended peak power output (C/8 discharge rate):
800 AH * 48 volts * 0.85 ivntr eff * 1/8 hour discharge rate = 4,080 watt maximum continuous output

The maximum surge power (C/2.5 discharge rate):
800 AH * 48 volts * 0.85 ivntr eff * 1/2.5 hour discharge rate = 13,056 watt maximum continuous output

You and your customer will have to decide what number(s) is/are "important" to them. But, at this point, I would suggest that an 800 AH @ 48 volt battery bank will serve a single XW 6048 at this point.

Break into a second post---To be continued.

-Bill

Looks like we certainly need to be adding batteries to this system if we are going to be fetching about like 7000 watts of load .

However I think it is safe for now discharge batteries at the rate of 4-5000w shared between 2 xantrex.. hopefully this is right?

I understand that we are running 2 Xantrex here Just for the same of doing it lol but this lets us expand system later on as the loads increase.

Will continue to reply your other post here.

Huge thank you for all the help .

Nila · May 2013

Re: Big 10k installation questions:

OK, we have an 800 AH @ 48 volt battery bank for now. Lets look at charging it from AC and Genset.

Our normal rate of charge is ~5% to 13%, and, perhaps up to 25% if there is thermal management for the battery bank (hot batteries require lower charging voltage, a remote battery temperature sensor feeding back to the battery charger for safety--and you should monitor/alarm the battery bank for over temperature (say 110-120F or 43-49C maximum).

The XW's have a very nice AC battery charger in them... I don't remember the exact specifications on efficiencies, but I will assume some pretty high efficiency numbers here (I have to let you have some of the fun here reading the manuals ).

Assuming 90% charger efficiency and 0.95 Power Factor and 13% rate of charge--pretty fast but should be well within the battery bank's capabilities. The 800 AH battery bank would draw around (note, I think a single XW 6048 has a maximum of 100 amp charge rate--so you will need to "fix" numbers below, depending on how you configure your system) (and NEC is US national electric code wiring/breaker/fuse deratings--makes for reliable installation anywhere--but your codes/hardware may be rated differently):
800 AH * 0.13% rate of charge = 104 Amperes rate of charge

800 AH * 58 volts charging * 0.13% rate of charge * 1/0.90 charging eff = 6,702 watts of AC charging power

800 AH * 58 volts charging * 0.13% rate of charge * 1/0.90 charging eff * 1/0.95 power factor = 7,055 VA to size branch circuit/generator

7,055 VA * 1/0.8 NEC derating * 1/210 VAC minimum line voltage = 42 amp minimum branch circuit rating

You can program the XW's to use less than maximum charging current if wiring/genset need to be smaller--But I would suggest 10% rate of charge is nominal and never to go under 5% rate of charge for battery health.

Since we have more than enough amperes from the Sun, We would like to disable the Grid charging totally during all the morning
Lets disconnect the Grid totally every morning and let grid be used for charging or running loads ONLY during rainy/cloudy days.

OR is it better to let grid be connected and then use the grid export facility to use the extra sun power be exported?

We cannot sell back to government.. so the exported power will it help running the loads? or is it okay to export it back to another nearby circuits?

Which is better idea.. you disable grid and let the inverter use the sun/battery power always? or you have grid and export back into it?

We can stop the design here if we are just looking at backup power. But since you talked about solar, we plough on.

So, we will have roughly four criteria to meet because this is a "day time" power usage vs our normal off grid home which charges during the day and discharges at night (assuming "off grid" solar power--since you have a genset and AC grid power, your solar array can be anything from zero to maximum wattage).

A) is the sizing the array to the battery bank.
is sizing the array to battery bank plus "average" daytime loads (need to meet battery minimum rate of charge + sustained loads to stay above 5% rate of charge)
C) maximum cost effective array + average loads (again, this is really an "off grid" design and will make the array way larger than is probably cost effective for your customer's needs).
D) maximum array to charge battery bank (cost effective)--Probably would recommend this number to start with as a "maximum" array. Could get into issues with problems during weekend use (i.e., array becomes too large for battery bank and can over voltage on weekends where there are no office loads).

First A), assumed load and hours of sun. 3,000 watt average load * 4 hours of backup per day (8 hours is probably way over kill here) and 4 hours minimum sun (use your numbers here):
3,000 watts * 4 hours usage * 1/0.77 panel+controller eff * 1/0.85 inverter efficiency * 1/0.80 battery efficiency * 1/4 hours minimum sun noontime equivalent sun per day = 5,769 Watt array (assuming charge during day and use power when sun is down)

3,000 watts * 4 hours usage * 1/0.77 panel+controller eff * 1/0.85 inverter efficiency * 1/4 hours minimum sun noontime equivalent sun per day = 4,615 Watt array (assuming charge during day and use power during the day--i.e., battery bank not cycling)

Then , if this was a true off grid system, then we would need to both meet the battery bank's minimum 5% rate of charge and supply the loads. And we would not want too big of array either. So, the numbers would look something like this:
(800 AH battery bank * 58 volts charging * 1/0.77 panel+controller eff * 0.05 rate of charge) + (3,000 watt load * 1/0.77 panel+cntrllr eff) = 6,909 Watt array

I think we have more sun hours , even the minimum will be 5+. but our panels are not too efficient so that would make up for that lol.

This is a "huge" array and really depends on the customer's power usage (amount, time of day, mostly when the sun is up and shining) and what length/time of power outages would be supported (before firing up the genset--if there is one).

C) The maximum cost effective array for a battery bank--I use ~13% rate of charge--Above that rate, the battery bank gets recharged quickly and the array no longer has to supply any more power--a waste of capacity (money). So, using the equation from with the rate of charge raised to 13%:
(800 AH battery bank * 58 volts charging * 1/0.77 panel+controller eff * 0.05 rate of charge) + (3,000 watt load * 1/0.77 panel+cntrllr eff) = 11,733 Watt array maximum

That gives you a maximum cost effective array size...

D) We talk about battery banks not being able to supply enough current for loads (C/8 continuous, C/2.5 surge). It turns out that a battery bank can also have problems sinking too much current when they are full. And MPPT charge controllers sometimes can have "issues" where they they "sweep" the array to measure Vmp and Imp of the the array, and they dump the "extra current" in the battery bank--which can cause a "too small" battery bank to exceed 72 volts (XW inverter's shutdown voltage).

More or less, I believe that 1,000 Watts of solar array per 100 AH of 48 volt battery bank is a good number (and also happens to be the recommendation here by several people for the maximum AC inverter rating per AH of battery bank--i.e., an 800 AH @ 48 volt battery bank can reliably source ~8,000 watts maximum continuous). And for a Hybrid Inverter like the XW where it can "sell power" to the utility (a whole new set of questions for you and your utility), that would be a good maximum array sizing too. So, the maximum array for an 800 AH @ 48 volt battery bank (I would suggest) would be:
800 AH * 1,000 Watt / 100 AH = 8,000 Watt solar array (and maximum AC inverter load)

I will stop here at the moment--Until we have some agreement on the numbers/system requirements, picking hardware is a bit early in the game.

I think I like this .

[*]800 AH * 1,000 Watt / 100 AH = 8,000 Watt solar array (and maximum AC inverter load)

Im worried about the extra 2KW of the solar array , will that be useful when we export to grid or will that make matters worse?

But when i think off this in theory though it is sounding scary

10000 * 0.77 / 58 = 132.75 amps and we have an 800amp battery bank thats c/6 charge rate.

Will Charge controllers be able to limit the charging by itself or should we have to worry about this in more detail?

But given your 800 AH @ 48 volt battery bank, I would suggest that your 10kW array and pair of XW AC inverters are not an optimum match at this point... Either increase the battery bank AH rating and/or reduce the expectations/hardware installation.

You are correct if not for the requirements of the client to have a 10KW system on paper. we wouldnt have to use this big an array or 2 xantrexes lol.

Nila, am I helping or causing more confusion here.

Warmest regards,
-Bill

You are helping tremendously , I would need your help sizing the disconnects , fuses ,etc like i said on the first post.
I think im a bit confused though . Please let me know if i make things unclear.

Nila · May 2013

Re: Big 10k installation questions:

yeah we did and they have a genset to take care or they wil shut the computers/fans down and do it old style

Since its a new building its hard to do all the load estimates.

BB. · May 2013

Re: Big 10k installation questions:

Nila,

Looks like we certainly need to be adding batteries to this system if we are going to be fetching about like 7000 watts of load .

However I think it is safe for now discharge batteries at the rate of 4-5000w shared between 2 xantrex.. hopefully this is right?

I understand that we are running 2 Xantrex here Just for the same of doing it lol but this lets us expand system later on as the loads increase.

All of our rules of thumbs are based on "generic batteries"... And are intended to be conservative estimates. Obviously, I have no knowledge of the batteries available to you or how much local engineering help your battery supplier can provide--But it would be interesting to summarize the battery side of the rules of thumb and see what the engineers/battery experts think.

Pulling 5kW from an 800 AH @ 28 volt battery bank would be:

(5,000 watts * 1/0.85 invtr eff * 1/48 volt battery bank) * 1/800 AH bank = 0.15 ~ C/6.5

Probably OK for tens of minutes or an hour--But I would suggest that is too much load for 3-4 hours continuous loads. Obviously, if you plan on 50% maximum discharge than that the C/8 capacity may be closer to 700 AH (pure guess/example), your actual run time would be limited to:

(5,000 watts * 1/0.85 invtr eff * 1/48 volt battery bank) * 1/700 AH bank * 0.50 max discharge = 2.9 hours of run time

Which brings us into your whole battery life cycle cost estimates. For our forum, we focus on off grid power use and having a system that will last years/decade or more without battery replacement (remote systems, reliable power, allow batteries to lose upwards of 50% of capacity before replacement is required).

In your case, you may have more of the "forklift" model. Cycle deeply/heavily every day to 20% State of Charge. And fully/quickly recharge back to 90%+ every day. Batteries may only last 500-1,000 cycles or 1.5 to 3 years or so, then replacement. You have ~1/2 the size of battery bank that if it is monitored and replaced every 3 years vs 2x the batteries replaced every 6 years may be a better model for you (smaller battery bank, less cells to maintain, forces better monitoring/maintenance practices, less room for error, etc.).

Another possible advantage is that Lead acid batteries are very energy efficient when charging from 0 to ~80% State of Charge and not near as efficient when charging that last ~20%. Takes longer time, more electrical energy, more waste heat, more damage to battery bank (hydrogen/oxygen gases forming, using distilled water, erodiing plates, causing oxygen to be forced into positive plates and plate grids--corrosion, etc.).

There is a theory/practice out there where lead acid battery banks are cycled 50-80% and not "fully recharged" back >90% except every 5-10 days or so. We talk about "sulfates hardening" process starting after a few hours or day--But it appears that cycling batteries delays the whole sulfation process to where you can go 5-10 days between "full charges".

Given that you are looking at business systems, it may be interesting to research setting up a system to only recharge to 80% during the week and doing the "full" to 90%+ charging for the weekend (when there is no power usage anyway, and lots of excess solar power). And even supporting deeper cycling... Help to reduce battery bank size/costs, and energy "waste" when charging battery bank. Of course, this relies on you have detailed knowledge of energy usage and monitoring to ensure that you don't ever take the battery bank below ~20% SOC (where battery cells may be permanently damaged and customer may need to switch on the backup genset more often).

But it certainly leads one to, perhaps, not over sizing the battery bank--But run closer to "nominal guesses" of power needs and let the bank cycle deeper if the customer's power needs are higher than initially planned. And see what happens for overall battery life/cycle costs.

Since we have more than enough amperes from the Sun, We would like to disable the Grid charging totally during all the morning
Lets disconnect the Grid totally every morning and let grid be used for charging or running loads ONLY during rainy/cloudy days.

OR is it better to let grid be connected and then use the grid export facility to use the extra sun power be exported?

We cannot sell back to government.. so the exported power will it help running the loads? or is it okay to export it back to another nearby circuits?

Which is better idea.. you disable grid and let the inverter use the sun/battery power always? or you have grid and export back into it?

This gets into monitoring... If you cannot sell to grid, then what is "export to grid" mean?

If it means that the utility will not pay for your "excess power", but it is legal to send power do the grid--It may make sense. If the entire office is not on "UPS Power", but there are significant other power usages when the grid is up (say A/C for the office that is not battery backed), then exporting to grid can be "free power" to reduce other office loads.

This gets into the whole government rules vs actual electrical grid/metering capabilities. You have "grid stability" of a local "generator" feeding back into grid (can cause utility engineers headaches with voltage stability, and there are a very few grid designs where back feeding is a "fault" and can cause a local power outage--And in the US, people would plug meters in upside down 2 weeks a month to steal power--So utilities designed meters to "always charge for power" no matter the direction of current flow... So you could be charged for supplying power for grid).

I think we have more sun hours , even the minimum will be 5+. but our panels are not too efficient so that would make up for that lol.

It's the old trade off between solar array/battery bank costs vs fuel/maintenance for the backup generator.

I think I like this .
[*]800 AH * 1,000 Watt / 100 AH = 8,000 Watt solar array (and maximum AC inverter load)

Im worried about the extra 2KW of the solar array , will that be useful when we export to grid or will that make matters worse? But when i think off this in theory though it is sounding scary

10000 * 0.77 / 58 = 132.75 amps and we have an 800amp battery bank thats c/6 charge rate.

Will Charge controllers be able to limit the charging by itself or should we have to worry about this in more detail?

In general, no. Charge controllers cannot tell the difference between current going into recharging the battery bank vs supplying other DC loads/AC inverter inputs.

Midnite (Classic MPPT charge controllers) are just about to come out with an integrated battery shunt that measures current being provided to the battery bank. And allow the controller to terminate charging when the battery falls below ~1% rate of charge (0.1-2% is the typical charge termination signal for industrial chargers). I wonder if this could also be upgraded to include a C/8 battery charging limit? It may be worth your time to contact them with this question (Midite has an active forum, and the founders/engineers sometimes visit our board too).

Now--Several things. Warm/hot batteries require less charging voltage, so you should have a remote battery temperature sensor to help prevent thermal run away (hot batteries with too much charging voltage/current get hotter, eventually boil/melt down, etc.). So thermal/voltage management may help here anyway. And batteries that are deeply discharged (below ~80% SOC) are more efficient, so they generate less heat/gases during this phase of charging. C/6 at less than (approximately) 80% SOC may be perfectly OK... And you can adjust the charging set point voltage/even adjust the thermal offset to be more agressive to bring down charging voltage faster if you want the controller to backup charging voltages/current if the battery banks are hot (-5mVolt per degree C per cell is typical setting).

Termination charging current brings up another interesting point for you and your battery monitoring/service plan... I read a paper somewhere recently from a company that recovers battery capacity for UPS type applications (another issue/discussion). Their data says (suggests?) that many battery failures are easily predicted by monitoring this average termination current. For a new AGM battery bank, the termination current may be as low as 0.1% of Bank AH capacity. And, if I recall correctly, when the "charging tail current" rises to ~2% or higher, the battery bank is starting to fail and can, eventually, even catch fire (catastrophic failures).

You are helping tremendously , I would need your help sizing the disconnects , fuses ,etc like i said on the first post.
I think im a bit confused though . Please let me know if i make things unclear.

You are doing fine by me... We are covering a wide range of questions here--So I would suggest that we "focus" on one set until you get clarity (say sizing of system/battery issues). Then talk about details of hardware design (and we have some "firm" numbers for power usage/battery bank sizing). It will make it less confusing for all. Just keep a list of new questions off to the side for now (I am sure it will grow as we continue the discussions).

Hmmm... My wife and I love Indian food, and my wife loves to travel. I may be looking you up someday for a good dinner. LOL

-Bill

Nila · May 2013

Re: Big 10k installation questions:

All of our rules of thumbs are based on "generic batteries"... And are intended to be conservative estimates. Obviously, I have no knowledge of the batteries available to you or how much local engineering help your battery supplier can provide--But it would be interesting to summarize the battery side of the rules of thumb and see what the engineers/battery experts think.

Pulling 5kW from an 800 AH @ 28 volt battery bank would be:
(5,000 watts * 1/0.85 invtr eff * 1/48 volt battery bank) * 1/800 AH bank = 0.15 ~ C/6.5

Probably OK for tens of minutes or an hour--But I would suggest that is too much load for 3-4 hours continuous loads. Obviously, if you plan on 50% maximum discharge than that the C/8 capacity may be closer to 700 AH (pure guess/example), your actual run time would be limited to:
(5,000 watts * 1/0.85 invtr eff * 1/48 volt battery bank) * 1/700 AH bank * 0.50 max discharge = 2.9 hours of run time

This is hugely helpful. Now i can tell the client.. look if we have 16 batteries 800AH is going to give us only like 4-5k discharge at normal times.. surge can be max 7k but thats going to be it .

We do use deep cycle batteries which can go into 20% and last like 2-3 years easily and they have 5 years replacement warrenty lol so not worried about health actually
I think this helps it a bit.

Given that you are looking at business systems, it may be interesting to research setting up a system to only recharge to 80% during the week and doing the "full" to 90%+ charging for the weekend (when there is no power usage anyway, and lots of excess solar power). And even supporting deeper cycling... Help to reduce battery bank size/costs, and energy "waste" when charging battery bank. Of course, this relies on you have detailed knowledge of energy usage and monitoring to ensure that you don't ever take the battery bank below ~20% SOC (where battery cells may be permanently damaged and customer may need to switch on the backup genset more often).

But it certainly leads one to, perhaps, not over sizing the battery bank--But run closer to "nominal guesses" of power needs and let the bank cycle deeper if the customer's power needs are higher than initially planned. And see what happens for overall battery life/cycle costs.

This is sounding good too and most likely this is going to happen because we will have simultaneous loads going on so batteries may never have time to finish charging last stage.
Is it possible with Xantrex 6048 to like say Use GRID bypass and charging only if battery is lower than xx level? i think so.. so this way we dont have to worry about grid export features..

Just run everything off batteries and sun until the batteries go below 20% then ask the grid to support charging at nights/evening only if required. Weekend will automatically take care of full charging batteries ...

This gets into monitoring... If you cannot sell to grid, then what is "export to grid" mean?

If it means that the utility will not pay for your "excess power", but it is legal to send power do the grid--It may make sense. If the entire office is not on "UPS Power", but there are significant other power usages when the grid is up (say A/C for the office that is not battery backed), then exporting to grid can be "free power" to reduce other office loads.

This gets into the whole government rules vs actual electrical grid/metering capabilities. You have "grid stability" of a local "generator" feeding back into grid (can cause utility engineers headaches with voltage stability, and there are a very few grid designs where back feeding is a "fault" and can cause a local power outage--And in the US, people would plug meters in upside down 2 weeks a month to steal power--So utilities designed meters to "always charge for power" no matter the direction of current flow... So you could be charged for supplying power for grid).

This is a headache because our grid is run by morons who do not know what they are doing lol.. whatever i am going to ask will be met with answer NO or nothing.
I will give it a try though and see if they give us the meters that run both ways.. .

Midnite (Classic MPPT charge controllers) are just about to come out with an integrated battery shunt that measures current being provided to the battery bank. And allow the controller to terminate charging when the battery falls below ~1% rate of charge (0.1-2% is the typical charge termination signal for industrial chargers). I wonder if this could also be upgraded to include a C/8 battery charging limit? It may be worth your time to contact them with this question (Midite has an active forum, and the founders/engineers sometimes visit our board too).

Midnite has to be imported to India and do not have local support there like Xantrex does so we are probably tied to xantrex at this point.

Now--Several things. Warm/hot batteries require less charging voltage, so you should have a remote battery temperature sensor to help prevent thermal run away (hot batteries with too much charging voltage/current get hotter, eventually boil/melt down, etc.). So thermal/voltage management may help here anyway. And batteries that are deeply discharged (below ~80% SOC) are more efficient, so they generate less heat/gases during this phase of charging. C/6 at less than (approximately) 80% SOC may be perfectly OK... And you can adjust the charging set point voltage/even adjust the thermal offset to be more agressive to bring down charging voltage faster if you want the controller to backup charging voltages/current if the battery banks are hot (-5mVolt per degree C per cell is typical setting).

Im sure Xantrex mppts have the thermal sensor/management .
I think we can have their engineers tune for the 80%+ charging voltages,etc

Termination charging current brings up another interesting point for you and your battery monitoring/service plan... I read a paper somewhere recently from a company that recovers battery capacity for UPS type applications (another issue/discussion). Their data says (suggests?) that many battery failures are easily predicted by monitoring this average termination current. For a new AGM battery bank, the termination current may be as low as 0.1% of Bank AH capacity. And, if I recall correctly, when the "charging tail current" rises to ~2% or higher, the battery bank is starting to fail and can, eventually, even catch fire (catastrophic failures).

We use tubular lead acid batteries deep cycle. I dont think i have anyway to find the termination charging current actually. I will try to research and learn more here.

You are doing fine by me... We are covering a wide range of questions here--So I would suggest that we "focus" on one set until you get clarity (say sizing of system/battery issues). Then talk about details of hardware design (and we have some "firm" numbers for power usage/battery bank sizing). It will make it less confusing for all. Just keep a list of new questions off to the side for now (I am sure it will grow as we continue the discussions).

We are going to start with the 800AH battery bank and tell the client to not expect over 5kw draw at the moment and will monitor the system to see how much it can provide when there is sun's power available together with the battery to see if we need upgrade to batteries.

So lets size this system based on these numbers. Thanks for all the help but need some more pointers from you to size the other things.

Hmmm... My wife and I love Indian food, and my wife loves to travel. I may be looking you up someday for a good dinner. LOL

-Bill[

THis is the easiest part.. I would be happy to assist you and take you around if you ever come to south india,:)
I do love travelling a lot and let me know if you ever come to anywhere in SE Asia. I have been all around there and is short flight for me to reach anywhere there.
PM me when you have plans to visit .

NorthGuy · May 2013

Re: Big 10k installation questions:

Nila wrote: »

.. look if we have 16 batteries 800AH is going to give us only like 4-5k discharge at normal times.. surge can be max 7k but thats going to be it.

For that load, one XW6048 is probably enough, so you don't need two.

Nila wrote: »

Is it possible with Xantrex 6048 to like say Use GRID bypass and charging only if battery is lower than xx level? i think so.. so this way we dont have to worry about grid export features.

When connected to grid and not selling, XW6048 will always draw a small amount, about 1A, from the grid. It's about 6kWh every day. For two units, I guess, it's twice this amount - 12kWh every day. And it will fully re-charge batteries from the grid every time you come from a power outage.

There's a Charge Block setting which lets you diconnect the grid completely for specified day time. This may be useful for you because if the battery voltage drops too low durging Charge Block time, it'll still start charging, but regular charging settings won't apply.

It is possible that you won't be able to get from XW exactly what you want. So, I suggest you read their manual, figure out the settings you need and make sure they do what you want them to do before ordering XW.

robert Tarzwell · May 2013

Re: Big 10k installation questions:

you may need more batteries

150 210 watt evergreen solar cells 4 radian 8kw inverters 88 6 volt wet cell 190 ah off grid

BB. · May 2013

Re: Big 10k installation questions:

OK, assuming 1x 6048 inverter for the moment. 5kWatt peak average power and 10kW peak surge loads.

Sizing the wiring is usually based on several requirements:

Maximum Average Power (both wire size for code and voltage drop)
Maximum Surge Power (code and voltage drop)
Protecting parallel battery strings for over current
Maximum short circuit current from battery bank
Paralleling batteries in a bank

Nominally, the suggested minimum voltage for a battery bank under heavy load would be something like 11.5 volts (46 volts for a 48 volt bank) for around 5-10 minutes--then reduce loads/start generator (you will probably have to play with these numbers to see what your requirements will be to protect the battery bank--If you plan on discharging deeper than 50% and/or want to supply more current to customer and don't care "that much" about the batteries, you can pick your numbers like 44 volts for 30 minutes, etc.).

It is always a question of your customer's needs. For an off grid home, they want to ensure that the battery banks will be around for many years to come. For a telephone company, they want the equipment to operate to destruction, then clean up the mess after the emergency is over.

Regarding watts, volts, and current... On the AC side, we have two numbers to pay attention to. Watts and VA ("Volt Amps" or Volts*Amps). We always use Power=Volts*Amp... But this is not the "true" AC power equation, the real one is:

Power = Volts * Amps * Cosine of the phase angle between V and I

Which is sometime written as:

Power = Volts * Amps * Power Factor (PF/Cosine vary between 1.0 and 0.0--1.0 being "good").

Power Factor can be calculated for "linear loads" (such as electric motors and resistance heaters) using the Cosine function (current lags with an inductor, leads with a capacitor). The short answer is that resistive heaters (and power factor corrected power supplies) have a PF >~0.95 and uncorrected electronics and electric induction (AC) motors have a PF in the range of ~0.67 to 0.80 or so (and worst case have seen PF of 0.50 for some CFL/Florescent lighting). Your AC wiring has to be designed for "VA" which is always at least the Wattage of the loads and, worst case can draw upwards of 2x current (worst case) current with respect to Watts.

For non-linear loads (such as many/most computer power supplies), they draw very "spikey" AC current. The PF and "true wattage" can be measured (with the right tools). But--VA can be measured with a good quality AC Current Clamp Meter designed to read RMS current and voltage (root mean square--RMS reading meters are more expensive).

AC Inverters (and generators, utility grid) supply VA, but the DC side of the battery bank will draw power based on AC Watts (more or less). So, you have to design the AC side (wiring, transformers, switches, etc. based on "VA" ratings and DC side on "Watt" ratings (with respect to the AC loads).

Note that most AC inverter are rated for maximum Watts=VA... You will see that larger commercial gensets may be rated in VA and Watts may be Watts=VA*0.80 PF -- I.e., their VA rating is higher than there Watt rating (or kVA and kW) to recognize that most AC loads do not have PF=1.0

Picking these numbers, plus the usual minimum inverter DC input voltage of ~10.5 volt/42.0 volts gives us another design point. Both for supported maximum voltage drop (wiring/breakers/connections) and the maximum input current for the inverter (remember that AC inverters are "constant power" devices--P=V*I--If the inverter sees low input voltage, it will increase current draw until it can supply the output or collapse the battery bus voltage.

This gives us some answer. Using the above guesstimates for power usage and average inverter efficiencies:

5,000 Watts * 1/0.85 inverter eff * 1/42 volts batt cutoff voltage = 140 Amps DC max estimated
10,000 Watts Surge * 1/0.85 inverter eff * 1/42 volts batt cutoff voltage = 280 Amps DC max estimated surge current

Now we can figure out the minimum wire size for the DC circuits... I am sorry, I don't remember if you are "metric" or US gauge/electric code... I will do this in US code since it what I know--But we can talk your numbers if needed.

So, minimum battery voltage of 46 volts at load (continuous or surge?) and 42 volt AC inverter cutoff... That gives us a maximum of 4 volt drop for all wiring/breakers/connections. And if we assume worst case surge current is 2x rated current... We should use the 280 Amps for our 4 volt drop (and maybe even less than 4 volts if you want to operate the batteries below ~46 volts under short term loads). But use 4 volts for now (engineering is layers and layers of rules of thumbs and safety margins--at some point you have to stop layering on safety margins as it makes everything too costly/impractical to build).

Using a generic voltage drop calculator (US type--note that some calculators use "one way" wire run, and others use "2 way wire run"), we can estimate wire size needed to support ~280 amps and 4 volt maximum drop (remember that fuses/breakers have drop too--So, you may have to add this for your "real system", hot batteries run lower output voltage, batteries increase internal resistance as they age, warm copper wire has higher resistance than cold wire, etc.--again layers of margins issues--You could try some experiments with a single string of batteries and an AC inverter--Start loading down at various test conditions and see how everything really performs--You will be ahead of 99% of the other installers/designers out there).

280 amps with 3 meter (10 feet) one way run and 4 volt max drop => 6 awg wire gives ~2.7 volt drop

Now, that wire diameter is much lighter than I expected--Longer wire runs will require heavier wire (2x longer run, ~3 AWG heavier wire, etc.). But, we still have to look at the "code" requirement for carrying current.

In the US/NEC (National Electric Code) we use a 1.25x safety factor (or we only run fuses/breakers/wiring circuits at 80% of their rated capacity--if you run at >80% to 100% of rated capacity, the fuses/breakers may eventually trip).

280 amp inverter DC circuit * 1.25 NEC safety factor = 350 Amp fuse/breaker/wiring design minimum

We have two guidelines to look at... The NEC requirements would suggest a minimum wire gauge of 4/O or larger diameter wire (depending on insulation, conduit fill, etc.).

Or you can look at the ABYC (boating) requirements. They would suggest a minimum of 3/O cable (slightly smaller, and I believe ABYC uses SAE Wire Gauge which is slightly smaller than American Wire Gauge--aren't you guys happy you went metric?).

If you parallel cables, NEC only allows you do to this with heavy gauge wiring (I forgot the exact number--it is in the code somewhere--The do not allow paralleling of smaller gauge wiring--if you get a wire/connection break, it is not obvious and you could get over heated wiring for the remaining cables so I suggest not paralleling unless you have to).

And, you need to pick fuses/circuit breakers rated to meet the Voltage (and current) requirements of your circuits. DC current is much harder to interrupt vs AC current (DC arc welders are "very nice" units). So similar DC rated switches and breakers are much "larger" than their AC counter parts. Also, your battery bank will operate at ~60 volts nominal maximum--Many breakers/fuses are rated at 12 or 24 volts (automotive/boating industries). You may have to look at telecom equipment suppliers to find >60 VDC rated devices.

Plus there is the Amps Interrupt Current (AIC?) rating... A dead short on a residential AC pole transformer in the US is rated for 10,000 amps... So, all of our main panel breakers are rated for at least 10,000 AIC. A too small breaker (voltage or AIC) will simply arc over internally and catch fire.

http://www.midnitesolar.com/pdfs/MidNite%20PV%20Combiners%20explained%20w%20diagrams.pdf
http://www.youtube.com/watch?v=0MPbWQ-FRoU
http://www.youtube.com/watch?v=CERTxi_ZH_o (solar panel short testing)

And pulling fuses (and popping open touch safe fuse holders) can easily catch fire if under load when opened.

Your DC Battery banks are going to be capable of many 1,000's of Amperes into a dead short--and many DC protection devices are only rated for 1,000 AIC (+ or -). Many times it is easier to find fuses that have the correct AIC rating--But then that gives you the problem that you still will need a disconnect switch.

Battery bank wise, if you have a 4 parallel string bank, it would be nice to assume that all battery strings share the load equally. I.e., 1/4 of 140 amps = 35 amps per string.

In real life, I would would not assume that... My own unscientific work with parallel power cables (used a lot in electronics) would indicate that they share more in a 1/N factor... I.e., if I have four circuits, the total available (long term reliable) current based on a 35 amp per string rating would be:

35 amps/1 + 35/2 + 35/3 + 35/4 = 72.9 amps total "reliable"

So, personally, if I was to design such a system--I would suggest that for a 140 amp continuous load with 4 parallel strings of batteries, that I would design each string for a rated ~70 amps each (for a total of ~145.8 amps rated battery bank output).

Now, since in the US we only run at 80% of rated current, the fuse/breaker/wiring for such a string should be rated at:

70 amps * 1.25 NEC margin = 87.5 amp wiring/breakers per string (round up to 100 amp standard fuse/breaker)

The above is conservative (I think) and what I would do based on my (limited) experiences. Note that doing this "over design" does allow the battery bank to still support your loads if there is a failure in one of the strings (at least for the short term). So it builds in a bit of redundancy too.

Nila, I will stop here for the moment--Comments/Questions? Where would you like to go next in the discussion(s)?

-Bill

Nila · May 2013

Re: Big 10k installation questions:

OK, assuming 1x 6048 inverter for the moment. 5kWatt peak average power and 10kW peak surge loads.

I guess I can use these knowledge I ve gained in this thread to calculate for the 2 xantrexes easily later on if needed.

Sizing the wiring is usually based on several requirements:
Maximum Average Power (both wire size for code and voltage drop)

Maximum Surge Power (code and voltage drop)

Protecting parallel battery strings for over current

Maximum short circuit current from battery bank

Paralleling batteries in a bank

I never knew there is these many factors in sizing the wire lol.. Here the norm is either u choose 4 / 6/ 8 / 16MM and mostly thats going to be it.. Just assuming one would work!
I can learn hopefully become one of the best installers here thanks to you!.

Nominally, the suggested minimum voltage for a battery bank under heavy load would be something like 11.5 volts (46 volts for a 48 volt bank) for around 5-10 minutes--then reduce loads/start generator (you will probably have to play with these numbers to see what your requirements will be to protect the battery bank--If you plan on discharging deeper than 50% and/or want to supply more current to customer and don't care "that much" about the batteries, you can pick your numbers like 44 volts for 30 minutes, etc.).

It is always a question of your customer's needs. For an off grid home, they want to ensure that the battery banks will be around for many years to come. For a telephone company, they want the equipment to operate to destruction, then clean up the mess after the emergency is over.

We are going to be in the telephone company mode here..Lets not care much about bank here

.

Power Factor can be calculated for "linear loads" (such as electric motors and resistance heaters) using the Cosine function (current lags with an inductor, leads with a capacitor). The short answer is that resistive heaters (and power factor corrected power supplies) have a PF >~0.95 and uncorrected electronics and electric induction (AC) motors have a PF in the range of ~0.67 to 0.80 or so (and worst case have seen PF of 0.50 for some CFL/Florescent lighting). Your AC wiring has to be designed for "VA" which is always at least the Wattage of the loads and, worst case can draw upwards of 2x current (worst case) current with respect to Watts.

Ive learnt some of this from your previous post helping me size my office setup. Thanks again anyway.

For non-linear loads (such as many/most computer power supplies), they draw very "spikey" AC current. The PF and "true wattage" can be measured (with the right tools). But--VA can be measured with a good quality AC Current Clamp Meter designed to read RMS current and voltage (root mean square--RMS reading meters are more expensive).

I have 2 meters,, one very costly DC and AC Clamp Meter which is true RMS. and
I also have a cheap AC Clamp meter which would probably calculate only the current and I dont think it would calculate the RMS,etc.

AC Inverters (and generators, utility grid) supply VA, but the DC side of the battery bank will draw power based on AC Watts (more or less). So, you have to design the AC side (wiring, transformers, switches, etc. based on "VA" ratings and DC side on "Watt" ratings (with respect to the AC loads).

Note that most AC inverter are rated for maximum Watts=VA... You will see that larger commercial gensets may be rated in VA and Watts may be Watts=VA*0.80 PF -- I.e., their VA rating is higher than there Watt rating (or kVA and kW) to recognize that most AC loads do not have PF=1.0

AC side wirings I think are not really my scope but I would just tell our VA and their electricians will take care of that. so not worried hugely there.
Helpful though to know all this.

Picking these numbers, plus the usual minimum inverter DC input voltage of ~10.5 volt/42.0 volts gives us another design point. Both for supported maximum voltage drop (wiring/breakers/connections) and the maximum input current for the inverter (remember that AC inverters are "constant power" devices--P=V*I--If the inverter sees low input voltage, it will increase current draw until it can supply the output or collapse the battery bus voltage.

This gives us some answer. Using the above guesstimates for power usage and average inverter efficiencies:
5,000 Watts * 1/0.85 inverter eff * 1/42 volts batt cutoff voltage = 140 Amps DC max estimated

10,000 Watts Surge * 1/0.85 inverter eff * 1/42 volts batt cutoff voltage = 280 Amps DC max estimated surge current

Now we can figure out the minimum wire size for the DC circuits... I am sorry, I don't remember if you are "metric" or US gauge/electric code... I will do this in US code since it what I know--But we can talk your numbers if needed.

We are metric, due to british rule may be . It would be easy for me to speak metric however I can always use online calculators to change between these

So, minimum battery voltage of 46 volts at load (continuous or surge?) and 42 volt AC inverter cutoff... That gives us a maximum of 4 volt drop for all wiring/breakers/connections. And if we assume worst case surge current is 2x rated current... We should use the 280 Amps for our 4 volt drop (and maybe even less than 4 volts if you want to operate the batteries below ~46 volts under short term loads). But use 4 volts for now (engineering is layers and layers of rules of thumbs and safety margins--at some point you have to stop layering on safety margins as it makes everything too costly/impractical to build).

I might go as low as 44V and I think the margin would be like 3v for us.

Using a generic voltage drop calculator (US type--note that some calculators use "one way" wire run, and others use "2 way wire run"), we can estimate wire size needed to support ~280 amps and 4 volt maximum drop (remember that fuses/breakers have drop too--So, you may have to add this for your "real system", hot batteries run lower output voltage, batteries increase internal resistance as they age, warm copper wire has higher resistance than cold wire, etc.--again layers of margins issues--You could try some experiments with a single string of batteries and an AC inverter--Start loading down at various test conditions and see how everything really performs--You will be ahead of 99% of the other installers/designers out there).
280 amps with 3 meter (10 feet) one way run and 4 volt max drop => 6 awg wire gives ~2.7 volt drop

I think the batteries would be almost right next to the inverter so even shorter length..
may be 7 feet and 3 v drop may be? I think 6w again would do.. but looks like it is only 4mm wire

Actually we would have used the 16MM as that is the standard here lol but that I think would help when we go with more batteries? another invertenr,etc.

In the US/NEC (National Electric Code) we use a 1.25x safety factor (or we only run fuses/breakers/wiring circuits at 80% of their rated capacity--if you run at >80% to 100% of rated capacity, the fuses/breakers may eventually trip).
280 amp inverter DC circuit * 1.25 NEC safety factor = 350 Amp fuse/breaker/wiring design minimum

Thats going to cost a fortune ? 350A DC 60 or 72V breaker..

http://products.schneider-electric.us/products-services/products/circuit-breakers/direct-current-dcrated-circuit-breakers/500-vdc-thermalmagnetic-circuit-breakers/lhldc/
something like this? would do

If you parallel cables, NEC only allows you do to this with heavy gauge wiring (I forgot the exact number--it is in the code somewhere--The do not allow paralleling of smaller gauge wiring--if you get a wire/connection break, it is not obvious and you could get over heated wiring for the remaining cables so I suggest not paralleling unless you have to).

We have only those 12V 200AH batteries here so we have to obviously parallel

.

Your DC Battery banks are going to be capable of many 1,000's of Amperes into a dead short--and many DC protection devices are only rated for 1,000 AIC (+ or -). Many times it is easier to find fuses that have the correct AIC rating--But then that gives you the problem that you still will need a disconnect switch.

I do not know about the AIC rating values here. I think I can leave it to their AC electrical team.

I wouldnt even touch the fuses when on load, I saw those videos before haha. Hopefully 48V is not that big of a voltage to cause big problems. Can cause fire but not an arc? may be im wrong.

Why do we need a fuse and a disconnect? if we have a breaker instead?

Battery bank wise,

In real life, I would would not assume that... My own unscientific work with parallel power cables (used a lot in electronics) would indicate that they share more in a 1/N factor... I.e., if I have for circuits, the total available (long term reliable) current based on a 35 amp per string rating would be:
35 amps/1 + 35/2 + 35/3 + 35/4 = 72.9 amps total "reliable"

So, personally, if I was to design such a system--I would suggest that for a 140 amp continuous load with 4 parallel strings of batteries, that I would design each string for a rated ~70 amps each (for a total of ~145.8 amps rated battery bank output).

I would just take this as roughly Half the total current lol.. instead of calculating it like n/1+n/2,etc.. because most of our big systems will be 48V here.

Now, since in the US we only run at 80% of rated current, the fuse/breaker/wiring for such a string should be rated at:
70 amps * 1.25 NEC margin = 87.5 amp wiring/breakers per string (round up to 100 amp standard fuse/breaker)

The above is conservative (I think) and what I would do based on my (limited) experiences. Note that doing this "over design" does allow the battery bank to still support your loads if there is a failure in one of the strings (at least for the short term). So it builds in a bit of redundancy too.

This is helpful too this is a sytem that may expand in future so yeah..

It took 2 days to really read all your words and then digest .. it is hugely helpful.

When you parallel like 4 wires .. please tell me how you would combine them all.. and which would be the charge points..

I can draw connections and something and show you how would i do it if you think you want to see it.

Nila, I will stop here for the moment--Comments/Questions? Where would you like to go next in the discussion(s)?

-Bill

1. I would ask you the best structure to connect batteries

2. When we have many charge controllers.. should we take all these output wires.. combine them and then add it to the battery bank via 1 Breaker?

or would you have like few smaller breakers one from the each charge controller? and then add them all to battery bank?

May be it is easier to create some drawings to discuss the structuring of the system..

P.S we are exceeding the limits of this system I got this error , so I would remove some of your quoted text to get around for now

The text that you have entered is too long (13398 characters). Please shorten it to 12000 characters long.

BB. · May 2013

Re: Big 10k installation questions:

Nila wrote: »

I guess I can use these knowledge I ve gained in this thread to calculate for the 2 xantrexes easily later on if needed.

There is a "problem" with "oversized" inverters (grid tied and off grid)... An Inverter will usually output its maximum rated current (or power) in a "controlled" manner.

So, if you have a battery bank that you want to limit to 5kW output (just for sake of discussion). If you install one XW 6048 inverter, then you can count on that one inverter limiting the maximum current from (or to if charging) the battery bank. And many of these inverter/chargers are highly programmable... You can set several different limits (AC input current, output charging current, etc.) to further "protect" the battery bank and wiring.

However, if you install two 6kW inverters--At this point (as far as I know) there is no way to limit the combined output power of these inverters. You could have balanced loads of 2.5 kW each, then somebody adds a "new 2.5 kW AC load". Now you have 2.5+5.0kW loading from the inverters--But they don't know that they are exceeding the total battery bank rating.

So, you are left with no way to "limit" bank current to 5kW when you have two 6kW independent AC inverters connected (and the loading may rotate--One inverter is near 5kW and the second near 0 kW, then they switch)... So, where the two inverters "combine" you have a choice of putting in a Breaker rated at YYY amps which will blow if the total load exceeds 5-6 kW, or your battery bank wiring needs to be capable of 2xYYY amps so that you don't run the danger of causing a wiring fire.

We have similar issues with Grid Tied AC inverters... The inverter output limits are "hard" limits by the design of the inverter hardware... But the Solar panels have no "hard" output power limits. In colder climates (especially with snow on the ground), it is not uncommon for solar arrays to exceed their maximum rated output current/power. So, if you have an "oversized" GT inverter, the AC output wiring needs to be sized for the Array's maximum Not to Exceed output (which is usually defined as 1.56x Isc rating).

So if I had a GT inverter, I can size the wire to its "hard limit". Or, I have to size the wiring to the "soft limit" of the solar array--which ever is the "smaller" limit (i.e., 1.25x GT inverter current or 1.56x Isc-array current).

So--that is why I recommend sizing the (single in this case) Inverter to the loads--It keeps you from having to oversize the DC power bus--Or you run the risk of popping fuses on the main bus if the two (or more) inverters are loaded more to more power than the DC bus can supply.

From the bus to each inverter--Again, you have the option of wiring/breakers sized to the expected maximum load--Or to the maximum input rating of the inverter (batteries supply more current than designed, or breaker pops and takes down their backup AC).

Failure modes are what engineers spend the most time on... 20% of our time is designing the "straight line" everything is fine portion of the project. The other 80% of the work is "error handling"--Whether that is AC Power of Software Engineering--That is where you will put most of your effort (and the customers, if you do it correctly, will never know it because the system will never fail, or at least fail gracefully).

I never knew there is these many factors in sizing the wire lol.. Here the norm is either u choose 4 / 6/ 8 / 16MM and mostly thats going to be it.. Just assuming one would work! I can learn hopefully become one of the best installers here thanks to you!.

I believe India has been working on a National Electrical Code since ~1985--So you might get some good stuff out of there. You might be able to get a copy of the NEC from the US to get an idea of how deep they go into sizing wiring.

And, just to be clear, your 4/6/8/16mm wires are probably mm² (square mm, not diameter) I am guessing. You can certainly come up with a begining chart for system design:

4 mm² ~ 11 AWG 25-30 amps in conduit (depending on insulation NEC); ~38-45 amps (boating)
6 mm² ~ 9 AWG 30-40 amps; 51-60 amps
8 mm² ~ 8 AWG 40-55 amps; 68-80 amps
16 mm² ~ 5 AWG 55-75 amps; 102-120 amps

The above are just conservative guesses (mm <> AWG). But gives you an idea of what a 6kW @ 48 volt inverter draws:

6,000 watts * 1/0.85 inverter eff * 1/42 volt cutoff = 168 amps maximum current
168 amps * 1.25 NEC derating factor = 210 amp minimum rated branch circuit / breaker / fusing

Which would be on the order of ~85 to 107 mm² wiring. These big systems are a whole deeper level of design problems. Looking at ~6x 16mm² wires in parallel (per power leg).

If you were near a major sea port--I wounder if there are any electrical shops that specialize in electric fork lift/marine DC gear of this size.

We are going to be in the telephone company mode here..Lets not care much about bank here.

A very reasonable decision... So that gets back into "paralleling" to large AC inverter to one battery bank. I assume that you don't want full power on one and partial power on the other to pop a single breaker that takes down both AC inverters at the same time (there goes your redundancy). So you want the DC Bus to be rated higher than the maximum power you expect your two AC inverters to ever use. And each inverter should have a breaker that is less than the main bus breaker/wiring (i.e., you want the individual AC inverter to be knocked off-line, you don't want both inverters knocked off together if one over loads.

I also have a cheap AC Clamp meter which would probably calculate only the current and I dont think it would calculate the RMS,etc.

Usually, the "cheap meters" just measure the Peak voltage of the sine wave and then multiply by 1/sqrt(2) --- The RMS integration of the V² under the sinewave curve when related to a steady DC signal. Non-sine wave forms have a different conversion factor.

We are metric, due to british rule may be . It would be easy for me to speak metric however I can always use online calculators to change between these

Electric Codes, if done correctly, are all interelated... Wires get hot (hot wires have higher resistance which drives them hotter). Different insulation types have different max temperature ratings. Placing wires in conduit reduces the amout of heat transfer to outside air--and the more wires in a conduit, the less ability to dissapate heat. Wires in Sun get hotter ambient temperatures than wires in offices/homes., etc...

It would be great if you can get the Indian Electric Code and study on how they do thing. Note that codes are not always "safe" and regarding DC power systems, they sometimes are very wrong.

I think the batteries would be almost right next to the inverter so even shorter length..
may be 7 feet and 3 v drop may be? I think 6w again would do.. but looks like it is only 4mm wire

Back to the maximum AC Watts out, then how much DC current from battery bank (minimum battery voltage, nominal AC inverter efficiency, can the AC output be programmed to a maximum ouput power/current), etc...

Actually we would have used the 16MM as that is the standard here lol but that I think would help when we go with more batteries? another invertenr,etc.

For big systems, you are going ot be looking at cables >100 mm² for even 48 volt 6 kW systems... Then you start looking at "real computer room" ups systems with 180+ volt battery banks--again to keep the copper wiring reasonable size...

Beware--This stuff is DANGEROUS. High current/high voltage can create nasty ARC FLASH events. Ever hear a guy say "he saw the sun" during an electrical short? He really did and may have gotten badly burned too.

Thats going to cost a fortune ? 350A DC 60 or 72V breaker..

Yep.

http://products.schneider-electric.us/products-services/products/circuit-breakers/direct-current-dcrated-circuit-breakers/500-vdc-thermalmagnetic-circuit-breakers/lhldc/
something like this? would do

Schneider makes it almost impossible to link to their web pages... I found the breaker and that is what would be required for a large UPS system (your link is correct, it just does not allow deep linking).

We have only those 12V 200AH batteries here so we have to obviously parallel .

So, you would want each string to be capable of supplying the 1/N or ~1/2 of that 5kW load or ~100-125 amps per string (assuming 4 parallel strings and ~200 amp rated circuit).

On the other hand I suggest starting with C/2.5 maximum surge current so:

200 AH / 2.5 = 80 amps maximum surge current

So--even then, ~100 amps (80 amps * 1.25 NEC derating = 100 amp circuit)--Would be about the maximum that could be justified--Really depends on the design of the batteries too--Deep Cycle flooded cell batteries tend to not support as high of surge current (per Amp*Hour rating) as a car/marine type battery (thinner plates designed to start a car/motor with light weight). (AGM/GEL batteries are different yet).

If you are going to do a lot with these 12 volt 200 AH batteries--I would look into getting a good sized 12 volt battery charger and a ~2000 watt "cheap" MSW AC inverter. And experiment.

Get some electric heaters (some people even use 100-200 watt filament light bulbs) to act as loads. And excessive the heck out of that single battery (charging/discharging) while monitoring voltage/specific gravity/current flow, etc... and come up with your own figures of merit.

Break into another post...

-Bill

BB. · May 2013

Re: Big 10k installation questions:

Post continues:

I do not know about the AIC rating values here. I think I can leave it to their AC electrical team.

Just remember that AIC also applies to DC breakers.

I wouldnt even touch the fuses when on load, I saw those videos before haha. Hopefully 48V is not that big of a voltage to cause big problems. Can cause fire but not an arc? may be im wrong.

More or less, sustained arcs are possible at ~12 volts (does depend on the metals involved).

Arcs are plasma and have low resistance... So an arc on a voltage source (like a battery, they can draw enough current to pop a breaker. With solar panels, because they are current sources (operating current is pretty close to short circuit current), so with solar panels, short circuits can turn into sustained arcs and become an ignition source.|

With large current sources (batteries, large AC mains), there are hundreds to thousands of amp available--And Arc Flash burns are much more likely if there are people nearby.

Why do we need a fuse and a disconnect? if we have a breaker instead?

Your choice... In some cases, fuses have higher AIC ratings vs breakers (industrial office parks near the main distribution transformer have to use expensive fuses instead of breaker because of the low resistance from the main panel back to the main power transformer).

I would just take this as roughly Half the total current lol.. instead of calculating it like n/1+n/2,etc.. because most of our big systems will be 48V here.

Yep... Works for me. Again, this was my own personal observation of computer equipment and test gear (logic analyzers and such) that had lots of 5 volt current and they used lots of parallel ribbon cable and such to carry power. From what I saw, it was the "low resistance" connections/connectors that had the most problem--They carried the most current and heat=I²R so if a connection carried 2x current, it got 4x the heating--and failed first (burned connections/wiring).

When you parallel like 4 wires .. please tell me how you would combine them all.. and which would be the charge points..

Bolted connections and using compression (crimped) connection for wiring is the standard. A properly done crimp is "gas tight" and so should not fail over time. Also, if a connection gets hot, a crimp does not let the wire fall out (vs soldering wire into a copper end where the solder can melt and drop the wire and cause more problems).

Use a DC current clamp meter and make sure that your paralleled connections are all sharing current "equally". In the US electric code, they only allow paralleling of heavy wiring (larger diameters). They do not allow paralleling of smaller diameter wire.

Also, wire resistance can actually help "balance" the current among the parallel cables--I.e., short parallel cables have poor current sharing (quality of connections "steer" the current). Longer cables have more resistance in the copper wiring so minor variations in connection quality does not affect current sharing nearly as much.

I can draw connections and something and show you how would i do it if you think you want to see it.

There are people here who have much more experience than I with high current power systems--Perhaps one of them can give you some better pointers.

1. I would ask you the best structure to connect batteries

2. When we have many charge controllers.. should we take all these output wires.. combine them and then add it to the battery bank via 1 Breaker?

or would you have like few smaller breakers one from the each charge controller? and then add them all to battery bank?

May be it is easier to create some drawings to discuss the structuring of the system..

Hmmm... I am somewhat familiar with US/North American power connections (as used in phone companies). I really do not have any idea what is available in India.

You are getting to a point where a single very large AH battery bank powering 2-3-4-or more 6kW inverters may not be a good idea. And it may be better to "isolate" into one or two 6kW inverter battery sets.

Anytime you add lots of batteries and "relatively" low voltages (48 volts), you are going to be talking about a lot of copper and brass bus bars and cable connections. There is no way to do this "cheap".

For a large system, you are probably looking at something like this:

Attachment not found.

This is getting well out of my comfort zone... See if you can find friends/suppliers that perhaps can introduce you to the wonderful world of high current power management.

P.S we are exceeding the limits of this system I got this error , so I would remove some of your quoted text to get around for now

I have never had that happen to me before...

-Bill

Big 10k installation questions:

Comments

Categories