Sourcing energy storage capacitors in the UK

[John Beardmore]

We are looking to put together a low voltage energy store (probably 24
or 48 volt) to store energy to run a 750 watt pump for a few minutes at
a time.

Can anybody recommend any good sources of appropriate capacitors,
ideally in the UK please, and relate any good or bad experiences with
such projects ?


Cheers, J/.

 

[Roger Hamlett]

We are looking to put together a low voltage energy store (probably 24
or 48 volt) to store energy to run a 750 watt pump for a few minutes at
a time.

Can anybody recommend any good sources of appropriate capacitors,
ideally in the UK please, and relate any good or bad experiences with
such projects ?

Try talking to Norfolk capacitors. They were able to help me in the past
with a rather smaller energy demand system.
There are a number of big questions that you are going to have to answer. If
the voltage needs to be stable, than there is going to need to be regulation
between the capacitor and the motor. You will need some fairly complex
energy management electronics if the load is to be kept small when the
supply is restored. There is obviously a balancing act between this, and a
battery based UPS system.

Best Wishes

 

[John Beardmore]

Try talking to Norfolk capacitors. They were able to help me in the past
with a rather smaller energy demand system.
There are a number of big questions that you are going to have to answer. If
the voltage needs to be stable,

Not particularly. The control electronics needn't run at capacitor
voltage. We can buck down to something sensible.

than there is going to need to be regulation
between the capacitor and the motor.

Not needed for the primary load.

You will need some fairly complex
energy management electronics if the load is to be kept small when the
supply is restored.

Sorry, don't see what this is getting at ? Am I being thick ?

We aim to store power in the caps, then when we have a worthwhile
amount, dump it into the primary load until the capacitor voltage falls
to near the load stall voltage, at which point we cut off the load, and
let it all charge up as and when.

Might also need an input power regulator if the load isn't needed or
fails, but none of this seems too problematic.

There is obviously a balancing act between this, and a
battery based UPS system.

Main reason we are going for capacitors is that we hope they'll last
longer than batteries even though they will store less energy.


Cheers, J/.

 

[Tim Wescott]

Not particularly. The control electronics needn't run at capacitor
voltage. We can buck down to something sensible.




Not needed for the primary load.

Really? The salient behavior of a capacitor is that the voltage is
proportional to charge, and to the square-root of energy stored -- if
you just connect the cap to your motor you'll see the voltage on the
caps fall exponentially, and the motor will see wildly varying power levels.

Sorry, don't see what this is getting at ? Am I being thick ?

If you have a constant-voltage source and just apply it straight to the
caps you'll have a _big_ current surge. You'll need to control the
current going into the capacitors one way or another. This is probably
the "input power regulator" that you mention below.

We aim to store power in the caps, then when we have a worthwhile
amount, dump it into the primary load until the capacitor voltage falls
to near the load stall voltage, at which point we cut off the load, and
let it all charge up as and when.

Might also need an input power regulator if the load isn't needed or
fails, but none of this seems too problematic.




Main reason we are going for capacitors is that we hope they'll last
longer than batteries even though they will store less energy.


Cheers, J/.

I don't know about the reliability of big energy-storage caps, but I
_do_ know that a properly managed set of Nickel-Cadmium batteries can
last a long time with a 10-minute discharge cycle, and 750W is quite
reachable. Check out what the model airplane folks are doing with their
NiCads, it's pretty amazing.

 

[John Beardmore]

Try talking to Norfolk capacitors. They were able to help me in the past
with a rather smaller energy demand system.

Thanks ! Found the website. Now drafting them an email.


Cheers, J/.

 

[John Beardmore]

John Beardmore wrote:

Really? The salient behavior of a capacitor is that the voltage is
proportional to charge, and to the square-root of energy stored -- if
you just connect the cap to your motor you'll see the voltage on the
caps fall exponentially, and the motor will see wildly varying power
levels.

Yes, that's fine. It's just there to pump water into a small wetland.

If we end up having to use an inverter and mains pump rather than a DC
unit we might need to do more, but then again, some inverters will cut
off when the input voltage is too low and restart a few volts higher.
That might pretty much do the job for us.

If you have a constant-voltage source and just apply it straight to the
caps you'll have a _big_ current surge. You'll need to control the
current going into the capacitors one way or another. This is probably
the "input power regulator" that you mention below.

The source will probably be solar panels. Shorting them isn't a great
idea, but the inrush current will only be risky on initial connection,
and that can always be done in the dark.

PV charging applications are usually 'limited' by a forward biased diode
and panel output.

The interesting question is perhaps, should we use an 'optimal power
point tracking regulator' to make transfer of power more efficient ?

Also, should we make the solar array track the sun ?

I don't know about the reliability of big energy-storage caps, but I
_do_ know that a properly managed set of Nickel-Cadmium batteries can
last a long time with a 10-minute discharge cycle, and 750W is quite
reachable. Check out what the model airplane folks are doing with
their NiCads, it's pretty amazing.

OK. I guess lead acid could do the job too, and with discharge rated to
less than 10% of rated capacity, I guess sulphation could be kept to a
reasonable minimum.


Cheers, J/.

 

[Ian Stirling]

Yes, that's fine. It's just there to pump water into a small wetland.

If we end up having to use an inverter and mains pump rather than a DC
unit we might need to do more, but then again, some inverters will cut
off when the input voltage is too low and restart a few volts higher.
That might pretty much do the job for us.

You probably want to define the problem more closely.
And consider the pump carefully.
Some common pumps are 10% efficiant.

For example, I have a
"600 watt 13000l/h pump."
It's a foul water centrifugal pump.
Measuring it, it draws 360W running.
Looking at the output graph, it will pump 6000l/h to 6m.

6000l/h is 1.6l/s.
1.6Kg raised by 6m takes 96J, so the efficiancy is around 25%.

Lifting a ton of water a day ten meters takes around a watt minimum.
What head/flow do you need?

You almost certainly do not want a big pump, but a tiny pump running for
hours.

You may want to consider a better design of pump, if there is room.
Something like a big wheel with buckets could closely approach 100%.

 

[John Beardmore]

You probably want to define the problem more closely.

I hope not - it's been done to death internally already !!

And consider the pump carefully.
Some common pumps are 10% efficiant.

For example, I have a
"600 watt 13000l/h pump."
It's a foul water centrifugal pump.
Measuring it, it draws 360W running.
Looking at the output graph, it will pump 6000l/h to 6m.

6000l/h is 1.6l/s.
1.6Kg raised by 6m takes 96J, so the efficiancy is around 25%.

Lifting a ton of water a day ten meters takes around a watt minimum.
What head/flow do you need?

Aiming for 2000 cubic meters per year, mostly in summer.

You almost certainly do not want a big pump, but a tiny pump running for
hours.

Agreed, but there is the matter of what is available off the shelf.

You may want to consider a better design of pump, if there is room.

This must be vandal proof and ideally fairly compact.

Something like a big wheel with buckets could closely approach 100%.

While we'd like an efficient pump, we have to recognise that that time
to do innovative R&D is significant.

It will very probably be cheaper to buy more PV than make an innovative
pump that is also very reliable and cheap to maintain.

This said, all recommendations for efficient pumps that can lift say 4
meters, DC or AC would be welcome, especially in the next day or
two !

Keep in mind that as most of the water is needed in summer the use of
PVs is quite appropriate as energy supply maps well on to demand, but on
the other hand, the bulk of the work has to happen during the relatively
short periods of good sunshine, perhaps totalling 500 to 600 hours per
year with a non tracking collector system.


Cheers, J/.

 

[Ian Stirling]

Well, at least here

Aiming for 2000 cubic meters per year, mostly in summer.

What sort of head?

So, 2000 tons, in 100 days, or 20 tons a day.

How often do you envisage servicing this?

20 tons a day, let's say it's running for 5 hours, so that's
4 tons an hour, or about a liter a second. (this implies fairly good
insolation and a fixed panel, or a slightly less good panel with a tracker.

Ok...
You later say the head is 4m, so that's 40W of 100% efficiant pump.
160W would probably be a reasonable worst-case guesstimate for the 'right'
AC centrifugal pump.

Agreed, but there is the matter of what is available off the shelf.

There are many shelves on the internet.

It will very probably be cheaper to buy more PV than make an innovative
pump that is also very reliable and cheap to maintain.

Very possibly.

This said, all recommendations for efficient pumps that can lift say 4
meters, DC or AC would be welcome, especially in the next day or
two !

This is almost certainly the wrong group.
Hmm.

I'd look at the various alternative energy, and solar groups.
Look for newsgroups with photovoltaic in the name, and you should be
good.
Alternatively, have a browse round the fishpond websites.
You want to find a site with flow curves for pumps.
Find one that does a bit over your desired flow (I suspect you'r looking
for one with 8m maximum head, and 8000l/h flow (the efficiancy generally
peaks at around half the max head)), and of course the minimum wattage.

Unfortunately, the pump makers often put a bigger wattage sticker on the
side of the pump than it really needs, which is a problem for you.

Keep in mind that as most of the water is needed in summer the use of
PVs is quite appropriate as energy supply maps well on to demand, but on
the other hand, the bulk of the work has to happen during the relatively
short periods of good sunshine, perhaps totalling 500 to 600 hours per
year with a non tracking collector system.

You have looked at the various solar FAQs, and got the information for
your location as to expected solar gain?

However, 250W of solar panel sounds like it would probably work, with
a 160W pump.

If servicing every 2 years is OK, I would go with lead-acid.
I'd want to keep the battery over 75% charge, so you want to size the
battery so that a really bright day can charge it up to 100%, and then
it'll keep going a bit during the night until the voltage drops below
a threshold (lead-acid will have best cycle life if kept partially charged,
75% may be conservative.) (read the sci.chem.battery.electrochem FAQ)
Say the solar panel can provide 90W over the demands of the pump for
8 hours, so that's 720Wh, which would lead to a 240Ah 12V battery.
On a good day, the pump would be running for around 4.5 extra hours.

I think you'r at about where you'd have to look whether a solar tracker
is really worth it.
Yes it might allow you to shave a few hundred dollars off the solar panel
price, but it's not going to be free.

You probably want a peak-power tracker with battery charger, something
to cut out the pump when the battery voltage gets too low and if
you can't find DC pumps, an inverter.

 

[John Beardmore]

Well, at least here
Possibly...


What sort of head?

So, 2000 tons, in 100 days, or 20 tons a day.

How often do you envisage servicing this?

As little as possible.

20 tons a day, let's say it's running for 5 hours, so that's
4 tons an hour, or about a liter a second. (this implies fairly good
insolation and a fixed panel, or a slightly less good panel with a tracker.

Ok...
You later say the head is 4m, so that's 40W of 100% efficiant pump.
160W would probably be a reasonable worst-case guesstimate for the 'right'
AC centrifugal pump.

Yes. Provisionally I'm assuming perhaps 320W of PV, a 357W single phase
250V pump and a 500W sine wave inverter.

There are many shelves on the internet.
Yes.



This is almost certainly the wrong group.
Hmm.

I'd look at the various alternative energy, and solar groups.
Look for newsgroups with photovoltaic in the name, and you should be
good.
OK.


Alternatively, have a browse round the fishpond websites.
You want to find a site with flow curves for pumps.
Find one that does a bit over your desired flow (I suspect you'r looking
for one with 8m maximum head, and 8000l/h flow (the efficiancy generally
peaks at around half the max head)), and of course the minimum wattage.

Unfortunately, the pump makers often put a bigger wattage sticker on the
side of the pump than it really needs, which is a problem for you.
OK.



You have looked at the various solar FAQs, and got the information for
your location as to expected solar gain?

Didn't know there were any FAQs on this in general. Then again, it's
been a long time since I looked, and usually I'm writing them not
reading them.

However, 250W of solar panel sounds like it would probably work, with
a 160W pump.

Sounds as if we are basically taking the same line but we've built in
more safety margin and assumed less efficient off the shelf kit.

If servicing every 2 years is OK, I would go with lead-acid.
OK.


I'd want to keep the battery over 75% charge, so you want to size the
battery so that a really bright day can charge it up to 100%, and then
it'll keep going a bit during the night until the voltage drops below
a threshold

This is OK up to a point, but our assumption is that batteries will
start to sulphate when discharged by as little as 10%.

We thus aim to charge up to 100%, then discharge to 90%.

100% charge can be determined by terminal voltage under charge, and 90%
discharge assumed after running a known load for a know time, though
perhaps with a low voltage cut off just in case.

I'm assuming we can always use more energy that we generate.

This way, the battery will only have to hold its charge for a few
minutes, and will never be left over night at less than 90% charge.

(lead-acid will have best cycle life if kept partially charged,
75% may be conservative.)

As I understand it, % discharge without sulphation is a bit of marketing
issue. 10% seems to be accepted in the academic community, do for
example you can safely take 10AH out of a 100AH battery.

Sometimes however, a say 200AH battery might get marketed as 100AH, but
able to accept 20% discharge without sulphation.

(read the sci.chem.battery.electrochem FAQ)

Thanks ! Didn't know there was one !

Say the solar panel can provide 90W over the demands of the pump for
8 hours, so that's 720Wh, which would lead to a 240Ah 12V battery.
On a good day, the pump would be running for around 4.5 extra hours.

See what you mean. Run at night as well as day.

Could do, but that takes you into longer term energy storage.

I must confess I had thought to do the job entirely during daylight
hours.

I think you'r at about where you'd have to look whether a solar tracker
is really worth it.

Yes, though the far that we'll be up a 15 meter mast with an existing
yaw bearing makes it more appealing than it might be !

Yes it might allow you to shave a few hundred dollars off the solar panel
price, but it's not going to be free.

No, though it could be as simple as

some studding used as a lead screw,

a nut on the studding welded to a hinge on the panel frame,

a couple of limit switches.


Could run the tracker closed loop, or could just take the more naive
position of starting facing East and turning to face West over the
course of eight hours.

You probably want a peak-power tracker with battery charger,

Yes. (Assuming you mean an 'optimal power point tracking (OPT)
regulator' ?)

something
to cut out the pump when the battery voltage gets too low

Yes, easy.

and if
you can't find DC pumps, an inverter.

Possibly, and many inverters will shut down if the input voltage is too
low.

Also add a way to isolate the PVs if the battery voltage is too high.


Cheers, J/.

 

[John Beardmore]

Try talking to Norfolk capacitors. They were able to help me in the past
with a rather smaller energy demand system.

Having chased them up, they claim to start at 20kV.

They have no problem with our wanting 60kJ, but the 48V is a problem.

What sort of system were you working on ?


Cheers, J/.

 

[Ian Stirling]

Yes. Provisionally I'm assuming perhaps 320W of PV, a 357W single phase
250V pump and a 500W sine wave inverter.

Didn't know there were any FAQs on this in general. Then again, it's
been a long time since I looked, and usually I'm writing them not
reading them.

alt.solar.photovoltaic.
There are sites where you can look up the location, and it'll give
you the optimal tilt, and average power output for the correctly tilted
panel, and average power for a tracker, taking into account weather
conditions.

This is OK up to a point, but our assumption is that batteries will
start to sulphate when discharged by as little as 10%.

The idea is to be able to use all the energy the solar panel provides,
and one way of ensuring this is to (probably better to discharge just
before the morning) discharge enought that you can store the excess
that may build up over an especially good day.

We thus aim to charge up to 100%, then discharge to 90%.

100% charge can be determined by terminal voltage under charge, and 90%
discharge assumed after running a known load for a know time, though
perhaps with a low voltage cut off just in case.

Batteries do have a temperature coefficiant, which you probably need
to look at if you haven't already, it alters the set-points a bit.

I'm assuming we can always use more energy that we generate.

Hmm, I see where you'r going.
This would need a significantly smaller battery, but has the problem that
a significant proportion of your energy is going to go into and out of
the battery, incurring losses (some 30-40% IIRC) along the way.
You'r also going to be charging and discharging the battery quite hard,
something which is not good for life.

This way, the battery will only have to hold its charge for a few
minutes, and will never be left over night at less than 90% charge.



As I understand it, % discharge without sulphation is a bit of marketing
issue. 10% seems to be accepted in the academic community, do for
example you can safely take 10AH out of a 100AH battery.

Sometimes however, a say 200AH battery might get marketed as 100AH, but
able to accept 20% discharge without sulphation.

Thanks ! Didn't know there was one !

Probably an idea to ask about charge philosophies there too.

Yes, though the far that we'll be up a 15 meter mast with an existing
yaw bearing makes it more appealing than it might be !

Is this mast partially underwater, or is the pump at the base?

No, though it could be as simple as

<snip>
It could be, but can you ensure that it can take several thousand cycles
when exposed to wind/rain/...

Yes. (Assuming you mean an 'optimal power point tracking (OPT)
regulator' ?)
I do.

Possibly, and many inverters will shut down if the input voltage is too
low.

You certainly don't want to take the battery down to the non-changable
cut-off points of many inverters. (10.6V is common)

 

[Frank Bemelman]

"John Beardmore" <[email protected]> schreef in bericht

[snip]

No, though it could be as simple as

What about a mirror, and a stirling motor directly driving
a mechanical pump?

 

[John Beardmore]

"John Beardmore" <[email protected]> schreef in bericht

[snip]

No, though it could be as simple as

What about a mirror, and a stirling motor directly driving
a mechanical pump?

Harder to armour against vandals than a PV, and hard to get good
Stirling engines off the shelf or at a sensible price ?

Keen to know if I'm wrong about that though, though if I am, please cc
to email as my news service is iffy at the moment.


Cheers, J/.

 

[Rich Grise]

"John Beardmore" <[email protected]> schreef in bericht

[snip]

No, though it could be as simple as

What about a mirror, and a stirling motor directly driving
a mechanical pump?

Harder to armour against vandals than a PV, and hard to get good
Stirling engines off the shelf or at a sensible price ?

Keen to know if I'm wrong about that though, though if I am, please cc
to email as my news service is iffy at the moment.

Would it make sense to focus it on an actual boiler of some kind?

The working fluid wouldn't have to be water, although it's got
about the highest heat capacity of anything I know of - freon
is probably not too hard to get, and would work at a much lower
temp. Then, a little steam engine.

I'd like to try that one myself, actually. ;-)

Cheers!
Rich

 

[John Beardmore]

Thanks for the pointers. I've quit this n.g. now due to time
pressure...

alt.solar.photovoltaic.
There are sites where you can look up the location, and it'll give
you the optimal tilt, and average power output for the correctly tilted
panel, and average power for a tracker, taking into account weather
conditions.

Thanks ! Will have a look !

The idea is to be able to use all the energy the solar panel provides,
and one way of ensuring this is to (probably better to discharge just
before the morning) discharge enought that you can store the excess
that may build up over an especially good day.

I think our plan is to use the power to do useful pumping ASAP rather
than cycling batteries more deeply than we need to.

Batteries do have a temperature coefficiant, which you probably need
to look at if you haven't already, it alters the set-points a bit.

Good point as this will be running at out door ambient temperature plus
a bit. In the UK I guess that's potentially a -15 to +30 range.

Hmm, I see where you'r going.
This would need a significantly smaller battery, but has the problem that
a significant proportion of your energy is going to go into and out of
the battery, incurring losses (some 30-40% IIRC) along the way.

Yes, if the pump routinely uses more than the panels generate.

You'r also going to be charging and discharging the battery quite hard,
something which is not good for life.

OK, though it could be made relatively large, charged less 'hard' and
cycled less deeply..

Probably an idea to ask about charge philosophies there too.
OK.



Is this mast partially underwater, or is the pump at the base?

Pump at base, base above canal.

<snip>
It could be, but can you ensure that it can take several thousand cycles
when exposed to wind/rain/...

With all the snipping I'm not sure what had to take it. Battery I guess
?

You certainly don't want to take the battery down to the non-changable
cut-off points of many inverters. (10.6V is common)

Indeed.

Capacitors by the way proved to be a non starter for this. 120F @ say
60V wildly expensive ! No magic wand there !


Cheers, J/.