Battery and Inverter Efficiency

[KenB]

Ivan, Martin & List,

I posting in this section - but feel free to move to a more appropriate section. 

Lead acid batteries are commonly used in off-grid ystems for power storage and buffering.  Batteries have a charge/recharge efficiency of around 66%, and with some inverters running atpoor efficiency at low load, the losses soon mount up, meaning that a larger battery pack might be needed.

Here are some of my observations from my set-up, partly as a record of my developments but also so that others my benefit from my observations.

As some of you know I have acquired a load of old Hawker SBS 40 sealed lead acid batteries, that are now wired up and powering my 5kW pure sinewave inverter.

Yesterday I decided to do a battery indurance test, using a 500W halogen flood light as the test load.  I was re-arranging my workshop/engine shed and the extra lighting was most useful on such a dull day.  The floodlight presented a very stable 472W load.

I wanted to confirm that the battery pack was of sufficient capacity to power the household during peak times and during the night.

My inverter is about 10 years old and was salvaged from an emergency lighting system. Unfortunately it has a fixed overhead power drain on the batteries, which means at low output the overall efficiency can be very low.

My reason for using it was that it was cheap, bombproof up to 5kW and produces very good quality sinusoidal power.  This is ideal for smoothing out the lumpy power that is generated by the Lister, and the battery acts as a substantial store of power.

After 12 hours of running the inverter and floodlight load, the battery had reached 90V or a nominal 1.66V per cell.  The battery current to the inverter remained in the range of 8.4A to 9.4A during the 12 hour test.

My batteries consist of 4 parallel strings of 9 batteries, so the discharge current through each string was between 2.1A and 2.35A.

At the end of the 12 hours I went through the batteries to identify which were good and which were at a reduced voltage.  Most were still in the spread of 11.0V to 11.9V, but some had dropped as low as 9.0 to 9.5V and one pour mite was down at 7V.

I have concluded the following:

1. If you have a mixed bunch of 2nd hand batteries, separate out the weaker ones and put them in the one string - these can then be given a more rigorous charge until their capacity is satisfactory.

2.  3/4 of the batteries had substantial capacity remaining, and it is likely that the pack would have better overall performance if the weaker batteries were removed and I ran on 3 strings rather than 4.

3.  The inverter efficiency is fairly poor when running a 472W load.  Between 850W and 940W was being drawn continuously from the pack during the duration of the test.

4. The pack was placed on charge overnight and the charger is drawing about 1150W from the mains.  11.6kW has already been consumed in just over 10 hours. It is likely that the bulk charge will take around 12 hours before the charger reaches its float voltage and reduces the current.

5. Some figures (for the record)

Battery Pack   36 SBS 40 sealed lead acid monoblocks, ex-telecom/standby arranged as 4 strings of 9 batteries.

Temperature in battery shed  20 C

Float Voltage 124V

Initial On-load voltage     111V  (when drawing 2.1A per string).

Test Load     472W halogen flood light

Power into Inverter (111V x 8.4A)  = 932.4W

Power from Inverter                      = 472W

Inverter efficiency   (472/932.4)     = 50.6%

Energy drawn from inverter in 12hrs 0mins  5.66kWh

Energy drawn from batteries (estimated)  = 10.67 kWh

Energy needed to recharge pack tofloat voltage  = 12 hrs x 1150W =  13.8kWh

Efficiency of battery charging process  = (10.67/13. = 77.3% - to be confirmed at end of charge cycle


Conclusions.

Lead acid batteries provide useful storage but not without substantial losses in the charge/discharge process.  An inverter and battery bank could be used for time-shifting power from cheap-rate night tariff, to daytime use, if the inverter is a high efficiency unit.

The inverter losses are considerable and need to be investigated. Mains cooling fans could be replaced with dc fans to reduce losses at low load.

There is a strong argument to distribute low voltage dc around an off-grid household, to avoid inverter losses.

A smaller, more efficient inverter could be used for light loads, such as office PC, lighting circuits, with the larger unit wired into the kitchen ring main for occasional use of kettle, washer, dishwaster etc.

There is a strong case to keep the fridge/freezer on grid power, switching only to inverter power during grid failure.

As the batteries will be recharged from the Lister using a direct dc feed, this will avoid some of the transformer losses.  However the estimate of 2kWh of power per litre of veg oil will have to be lowered to almost 1 litre per kWh of clean inverter power. These losses clearly affect the overall efficiency of the electrical generation side of the Lister CHP system, suggesting that there will be a greater percentage of waste heat - for home heating, at the expense of fuel burned.


I would be interested hearing from anyone else running a battery/inverter system in the 3kW to 5kW range, for their experiences of inverter losses. 




Ken

[KenB]

List,

As a follow up to above.

The battery charger has now entered float mode after 11 hours, with 11.85kWh drawn from the mains.

In this mode it slowly starts winding down the charge current to keep the batteries at their recommended float volatge of about 2.29 volts per cell.

The power used by the charger slowly reduces until the charger is only drawing about 110W.

I have also noticed that the power losses in my cable runs from the house to the workshop are not insignificant.  I happened to have a plug in power meter both in the house and in the workshop.

The house one read 12.35kWh and the workshop one 11.85kWh.

This means that 0.5kWh has been lost overnight as resistive losses in the  temporary extension cable that feeds the workshop - about 4% loss!

Fortunately I will be upgrading the workshop power feed to 6mm2, back to the meter, and using a larger cable size to bring the Lister/Inverter power back to the house.


Ken

[KenB]

List,

Whilst browsing the forum, I came across Paul_h_boats recent solar installation using a plug-in Soladin 120 grid tied inverter. This has to be one of the easiest ways to get your solar pV panels onto the grid.

I then looked at their other inverters, including the Soladin 600 inverter,  which can handle dc input voltages up to 125V, and claims 91 to 93% efficiency.

To me this seems a good way for my Lister genset and battery bank to make an appreciable contribution back to the grid.

The Soladin 600 also comes  with an optional PC Link and logging software so that you can keep track of your exports.

At around £450 it could be paid for in 2 years from the savings in electricity.

http://www.energyenv.co.uk/Soladin%20Solar%20Inverter.asp

http://www.mastervolt.com/com/53/prg/1035/solar_inverters.html

Anyone know of any other simple grid tied inverters?  Might Windy boy or sunny boy have a suitable dc input?



Ken

[Ivan]

Very useful information.

When I was at school I decided it could be very cost effective to charge batteries at night on economy 7 and re-use the electricity during the day - thus running on cheap electricity. Maybe even grid-tie to feed back into the grid during peak hours, and end up with free electricity. I had not discovered the poor charge/discharge efficiency of batteries, nor had I considered the lifespan of the batteries in terms of charge/discharge cycles. The grid is not particularly efficient, but you do have the advantage with small -scale production, of producing power close to its intended points-of-use.

Yes, the Sunny Boy would take power generated from the CHP unit. There are three modes of operation - solar PV (which is MPPT), wind (which is a voltage/power line graph) and 'constant voltage' - the third is probably the best mode, but I think it should work reasonably well in 'wind' mode too. There are two types of Windy Boy - high voltage - these are the cheaper, and ideal for coupling to 220v dc (rectified and smoothed ac) alterantors. (not trying to plug navitron products over standard alternators, but you need to get ripple below 10% - which is hard to do when generating a lot of power - with a Navitron 3phase 12pole alternator, you have higher frequency than a standard alternator, and smoothing is not an issue you need to worry about. If you use a 50Hz single phase alternator, you will need to spend around £400 on capacitors, which will get very hot, and have a lifespan of 2-3years).

I would not recommend the LV option (24-60V) of the Sunny Boy GT inverters. If you go over 65v, you will ruin the inverter - so it requires that you build protection electronics (Ken, if you want to check this - and it would be useful to do so - let the Lister spin your alternator open circuit and measure the voltage - assuming it is nominally 48v, I would expect it to be in the region of 80v). By comparison the high voltage model is able to cope with input voltages in the range 200-600v.

There are other GT inverters on the market, but I only have experience of the SMA Sunny Boy units. The Mastervolt unit (600W) is getting a lot of good press at the moment, so might be worth looking into.

BTW, it was quite cold today - I decided I must get my lister finished off - it only needs a silencer and the grid-connection. Suddenly, autumn does not seem to be too far away!


Ivan

[KenB]

Ivan,

Thanks for your suggestions.

You make valid points about the inefficiencies of batteries and the fact that they are limited to a finite number of charge/discharge cycles and their replacement cost has to be factored into the equation.

However - I like the idea of using a MPPT solar inverter as a way of getting peak power or maximum efficiency from a battery bank, or as you are intending, to use a standard wind turbine and grid tied inverter driven by the Lister and not the wind. 

Whilst my pack is a nominal 108V  (9 x 12V)  it has a float voltage of 124V, and during a normal discharge you will see the voltage soon settle to about 111.0V and then slowly descend until a recommended terminal voltage of 97V.

Although the dc power is produced from a battery, not a solar panel, the MPPT inverter will not mind. All it sees is a steady falling dc voltage, which would happen on a solar pV array, when the sun fades at the end of the day.

The Soladin 600 would be a much more efficient way of getting 600W or so, of inverted power back to the grid, to offset the normal consumption.

I wonder what happens at night, when your household consumption drops well below 600W.  I guess a modern electricity meter (with non reverse ratchet) is not going to turn backwards,  but will remain stationary.

It would therefore be advisable to invest in an export meter, and see if you can find any electricity supplier who is prepared to pay for your exported power.

Does anyone know/recommend any utility that offers a good rate for exported power?


Ken

[Ivan]

Hi Ken,

When you say MPPT - you are referring to the ability of the GT inverter to 'match' the voltage produced with the mains to keep efficiency high. Whatever function you use with the GT inverter (wind, constant voltage or solar) you will get this. The disadvantage of the MPPT option is that it hunts up and down trying to match the load to the power produced. With solar, if 'volts x amps' drops to a lower level, then the MPPT goes the other way, so hopefully reaching the maximum power voltage/current combination (I'd call this impedance matching, if it was working with audio or radio signals - not sure if you can use the same terminology with mains power-matching). If you do this with batteries, you'll just get a short (internal resistance ('impedance'?) of lead-acid batteries is very low) - actually, it will max out at whatever the GT max output is. I played around with the SMA LV inverter and the Soladin 120, both did not respond well to this. You could use it as a kind of dump though - so if the voltage goes above a certain level, it will start grid-feeding the excess. (Hence 'constant voltage' mode, might be a good option). If you are looking at SMA, check the input voltage range - I think the 700W or 1100W would be most suitable. The 2500W unit needs a minimum of  250v (I think) to get started, although it will continue to grid-feed down to 197v, if you set it to do so.

If you want to export and get paid for it, most meters have a rachet, so usually you need an export meter. I bought mine from Western Power (they do the export meters for most distributors) - cost £80 including installation (I own it, whereas the import meter is owned by the distribution company). Things have changed somewhat - see Paul_boats comments on Npower - they are paying him 1/3 of his output on estimate, as they can't be bothered to fit export meters! Many utilities will pay the same as they charge - but shop around some don't. I think npower's renewable tarriff is a good choice - costs same as standard, they pay the same as they charge, and you are buying renewable electricity when you do import

Ivan

Ivan

[KenB]

Ivan,

Thanks for the clarification regarding MPPT.

If we assume my battery has a pack voltage of 100V, and the MPPT presents a load to it that draws 1 A - then the power would be 100W.

Are you suggesting the MPPT then lowers its impedance - say to draw 2A, in an attempt to transfer 200W from the source.

It keeps lowering its impedance (if connected to a pV) until the point where the pV is running at its peak power - ie the product of amps and volts is a maximum.

With a well charged lead acid battery,  the voltage seen by the MPPT will not fall significantly as the MPPT lowers its impedance. This  might well confuse it if it is expecting the V I behaviour to match a solar pV.


Ken

 

[Ivan]

I think it's a little more complex than that, but that't the basic idea. If you use the wind power mode - it waits until voltage goes above a certain level, then progressively increases power as voltage increases. Which might be useful. You can tailor this slope in the SMA setup. Constant voltage tries to maintain a certain voltage. eg. if you set it to 250v, if the voltage decreases, power draw will stop. If the voltage increases above 250v, it will progressively draw more current to try to bring the voltage back down to 250v. I think this would work well for wind power, but never had the site to test it.



Ivan

[KenB]

Ivan, List,

Here's an update to my rather sorry looking inverter efficiency.

AC Load = 2410 Watts  (deep fat fryer 2350W & fluorescent lamp 60W)

Power into inverter  112.8V @ 22.9A = 2583.12 W

Power into battery from generator  = 24.7A @ 112.8V  = 2786.16W

Battery "losses"   =  203W

Inverter losses   = 173 W


Percentage from generator being converted to ac = 86.5%

Percentage from battery being converted to ac = 93.2%


Now that's a bit better 

All I have to do is find big loads for the inverter to maintain its efficiency.



Ken