Heavy Duty Energy Banking

Heavy Duty Energy Banking

 

Introduction  1

Some facts and implications  3

The Principles  5

More practical Alternatives and Elaborations  8

If a little is good, then a lot... 10

Getting More Serious About Practical Problems  10

Mole's eye view   14

If it still don't work, I gets a bigger 'ammer  16

What's your tipple?  18

Illustrations of some possibly useful configurations  20

 

Introduction

A few years ago, eheu fugaces, I wrote on the topic of energy storage in submarine tents of compressed air.  
http://fullduplexjonrichfield.blogspot.co.za/2011/01/energy-storage-renewable-energy-sources_04.html 
One’s mind tends to congeal around preconceptions (mine does anyway) so it was a long time before I seriously considered alternatives, and I tended to sneer at ideas concerning say, energy storage in the form of elevated heavy masses; they struck me as limiting and inelegantly mechanical, bulky and small of scale.

Bear in mind incidentally, that it is important not to confuse the concept of energy storage with that of energy generation; that is a perennial hazard in trying to convey the value of energy storage. The assumption is not that your storage device will produce energy, any more than your flask will produce wine, or even produce water to turn into wine. It does however leave room for recognising that as long as you can get the water or wine in the first place, a flask can be more useful for storing that water (or wine) than say, cupping it in your hands till you need it. Similarly, whereas a wind turbine can generate power while the wind blows, or a solar cell can do so while it is in sunlight, either or both would be useless in a calm in the dark, and if one needs a constant power supply, then it is well to have a means of storing the necessary energy against power-hungry intervals of still air at night.

Recently however, an on-line discussion stimulated me to think a little more flexibly, and as I now see things, it seems that there is room for reflection. Most such schemes take the form of raising water and letting it drive power turbines when necessary. Increasingly however, there is a trend towards the storage of electricity in the form of batteries. And there are other schemes, such as flywheel storage.

All those storage media have their points, and I do not pretend that the principle I describe will supersede all of them, but it does have points of interest and in special circumstances it might possibly be of practical value. The necessary structures would have very few working parts and should last indefinitely with minimal maintenance. 

Oh yes, you ask, then why don't we raise water for all our energy storage problems? And if it isn't in fact such a valuable way to store energy, what possesses us to use it in all our great power generating dams? 

The answer is that we take advantage of the fact that until recently we have had great tracts of land available to cover with water, and huge amounts of water to cover them with, and people didn't fuss much about the consequences of repeatedly flooding and exposing large areas of dam floor. But nowadays, both monetarily and ecologically, the costs of land and water are everywhere rising catastrophically. 

So maybe there is no harm to keeping an eye open for future alternatives to water storage dams for energy, especially if those alternatives have particular advantages such as: 

• not destroying or consuming resources such as land and water on a large scale;
• not causing pollution as fossils fuels and emptied dams do;
• not decaying in storage or in use as batteries do; and
• not seriously reducing the intensity or quality of the power they yield as they approach the end of their store of energy.

In this last attribute, dams, most forms of compressed air, flywheels, and batteries fall short.

So much for the sales talk.

 

Some facts and implications

Now, in this discussion, I will be cavalier with my arithmetic and my engineering assumptions, and without apology; if the argument were dependent on great precision, it would not be of much practical use, because most practical engineering has to deal with dirty, noisy, approximate conditions and we can’t afford systems that go haywire every time a passer-by sneezes. For example, for convenience I choose figures so rough as to be nearly fictitious, not least because I don’t want to bog down the discussion with pointless recourse to calculators. And some parameters I hardly even consider, such as the price of digging deep holes for construction; the devils in details vary with circumstances, and an engineer who cannot make allowances for such considerations in practice is not worth his salt. (Yes, I realise that some engineers are women, but I generally assume that a woman who can make it in the engineering world probably is very much worth her salt.)

Anyway, for convenience I assume for example that the density of lead is 10, whereas I fully realise that it is in fact closer to 11, and I take other liberties with the facts too when I reckon that they are close enough for jazz. But generally I am overly conservative, so you need not think that I am sneaking in unjustified assumptions under a cloak of innumeracy.

How did lead get into this for example, you ask? Isn’t lead that nasty, poisonous stuff? Yes, but that is one of its advantages. Its main use in bygone years was in high-octane fuel, which rightly was discontinued, causing a slump in the price of lead, which nowadays is used mainly in lead-acid batteries and building materials such as roofing. And in my opinion, its use in batteries is likely to wane in the next decade or two anyway.

For us it has other advantages too; technologically it is well understood; properly used, it is easy to fabricate, and chemically fairly inert, which takes care of a lot of safety concerns. Also, it is not flammable in bulk. 

And it is dense.

Proverbially dense in fact.

There are denser materials, some practically twice as dense, such as depleted uranium, tungsten, osmium and so on, but all those are far more expensive and not nearly as plentiful as lead. Mercury is nice, but much too expensive, and even more poisonous than lead. In fact among common dense materials lead is unique in its combination of favourable features at affordable prices. Iron probably would rank next, and some of the suggestions could have been based on iron instead, which has a density nearer to 8. But I’ll assume lead for most purposes.

The figure 10 for the density is convenient because if we take a tonne of lead as 1 cubic metre, then we can take the pressure exerted by a column of lead one metre high as exerting a downward pressure of about one bar. And why do I use a deprecated unit like bar? It simply is convenient; a bar is about 1 atmosphere and accordingly easy to visualise, and it is 100000 Pascal, for those who prefer to think in Si units.

Accordingly a ten-metre column of lead exerts a pressure of ten atmospheres, and so on. Roughly speaking, without making allowance for the cost of heavier-duty machinery and structures, the higher the pressure, the greater the efficiency and the higher the amount of energy that can be stored in a unit of a given design.

The Principles

First principle: any time you can build a place high enough in which to store a sufficiently dense fluid, you can in principle use it as an energy store by constructing a suitable generator at its bottom outlet, and filling it with that fluid. In this essay I'll assume an electricity generator, which is general enough for our purposes. Suit yourself about thinking in terms of alternative forms of energy.

The denser the fluid and the higher the drop and the higher the pressure, the larger the energy capacity. Furthermore, such combinations potentially increase both the energy density and the efficiency.

So far we have achieved the notional engineering sophistication of the backyard mechanic who understands three principles.

• I reckon it's working, so it aint broke, so don't fix it, and certainly don't even think of trying to improve it.
• If it really, really won't work then look for someone who can lend you a hammer, and explain to you which way round to hold it.
• If it still don't work, look for someone with a bigger hammer.

An example of a device to store energy according to such principles might be a cylindrical tower (yes, yes, I know, but other shapes such as cones or mushrooms have cons as well as pros, and are less scalable in our context. Our towers will be prismatic in principle and cylindrical in practice; just watch this space). 

Pump a fluid, say water, into the tower and when it is full enough, we can withdraw energy by using it to drive our turbine or other hydrodynamic generator; just open its tap to drive the turbine, and close the tap or drive the turbine in reverse when we want to store more energy.

Simple.

Also problematic; the higher the tower, the higher the cost, and it would take a big tower to store any energy. There are some inconvenient facts.

Devices to store energy by raising weights require some sophisticated engineering, and in most circumstances they might be expensive to build to supply more than a few megajoules. The fundamental problem is that it takes a lot of mass suspended at a considerable height to store many megajoules.

Oh. What are megajoules?

A convenient measure of energy in any useful form.

Consider: 3.6 megajoules (MJ) equal 1 kilowatt-hour (kWh).

You might consume more energy than 1 kWh just in roasting a joint of meat for a family meal, and yet, just 1 MJ is how much energy it takes to raise a 10 Tonne mass 10 metres. That sounds energy-cheap of course, but unfortunately, the other side of the coin is that from a ten tonne mass raised ten metres, you can barely get enough energy to prepare a meal.

So it is hard to imagine the billions of tonnes of water that power utilities need for schemes to store power in elevated storage dams.

One tonne of water occupies about 1 cubic metre, and a ten-metre column of water exerts a pressure roughly the same as atmospheric pressure, and not surprisingly, as already mentioned, if we work with molten lead instead of water, then it takes a column about 1 metre high to exert a pressure equal to one atmosphere.

And the same goes for a block of solid lead 1 metre high.

But how does one pump solid lead, why not just forget it and stick to molten lead?

Well, molten lead presents its own difficulties, and there are other advantages to solid lead apart from coolness and inertness.

Watch.

Consider a cylinder, say 100 metres, roughly as high as a thirty-story building. Expensive, but not enormously so. Now you pump it full of water and let it empty through a turbine. Suppose it has a diameter of about 3.6 metres (something like 12 feet or two fathoms for any Americans who still happen to prefer paces, palms, or perches); that rather eccentric figure I chose because it implies a  cross sectional area of about ten square metres. Then water to fill it weighs 1000 tonnes. And the energy you get out of it varies with the height of the water column; full power from the top, half power from water halfway up, and so on. The total output is the same as if you had dropped 1000 tonnes through 50 metres, not counting the fact that the last ten tonnes or so would yield practically no energy. (Hardly any pressure, see?) At 1 megajoule per ten metres per ten tonnes, that cylinder could yield something like 500 megajoules at best.

But suppose that instead of a 100 metre column of water, you had had a 10 metre high lump of lead in the top of the cylinder, and dropped that. There would be one awful crash of course, but this is just an illustration. The lead, all 1000 tonnes of it, would drop 90 metres, with a possible yield of some 900 megajoules. If we designed the cylinder so that the lead could stick 10 metres out of the top, we could get a neat, round 1000 megajoules.

But what about a more useful arrangement, though less exciting? We could dangle the lead slug on a chain wound onto a reel, so that it could wind down peacefully, spinning a dynamo to generate electricity.

Much nicer of course, but we still must try to do better. That arrangement would need a lot of moving parts with lots of wear and maintenance, and it is no joke lifting and controlling massive weights like 1000 tonnes even once. The gearing alone would be horrendous. In practice we might need to cycle the system a few times daily, and each cycle would stress the system drastically.

Not very practical. Get a bigger hammer.

More practical Alternatives and Elaborations

We most certainly are nowhere near to a practical design yet, but let's see whether we can improve matters.

Notice that the dangling mass certainly would have some tempting points. For a start, unlike the draining fluid, it would yield essentially full power from the word go and continue doing so right up to the end; the pressure would not peter out the way that the fluid pressure would. That simplifies a lot of the design and permits more economical installation and running.

But what about all that wasted space? The slug only occupies 10% of the tower. The fluid exploited the full volume, and unlike the slug, it smoothly distributed the stresses that it exerted on the wall and base of the cylinder.

So. Retain the fluid. Make the cylinder smooth on the inside. Manufacture the slug to fit into the cylinder with a sliding seal like a piston, but no chains, no suspension, unless it seemed advisable to be able to stop the mass at the top with an adequate set of actuated lugs that could keep it parked when fully charged. That should not be a very demanding challenge.

The system could pump fluid in at the bottom, raising the slug by fluid pressure. A ten-metre-long slug would require only about a ten-bar pressure to raise it smoothly; ten bars however would be the pressure only at the lower face of the slug; as the fluid level rose, its final mass, concentrated over a ten square-metre area, would add its thousand tonnes or so to the store, giving us a total pressure of about twenty bars. 

That is enough pressure to be efficiently useful, without being too high for realistic engineering. (I emphasise again that, being inspired by the example of the wall of Saint Donald Trump, I am working with convenient figures, not trying to preempt any professional's design.)

Now, this notional design, fluid and all, could yield about 1400 or 1500 megajoules, whereas even doubling the tower height without including the lead would only yield a total of about 1000 megajoules; and you could buy a lot of lead for the money saved by building a 100 metre tower instead of 200 metres. Furthermore, the presence of a 10-metre lead slug floating on top of the column of fluid would have a dramatically beneficial effect: it would guarantee a minimum pressure of 10 bars even as the last few litres flowed out, a pressure as good as the best that the fluid alone could yield with the tower chock full.

Starting to sound better, right?

Of course, it does assume some resources apart from the lead and the tower; the fluid for one thing. If the structure stood by the seaside or even by a perennial freshwater source, one simply could pump water in and run it out indefinitely. But in most places water is not free and we are becoming used to thinking of it in terms of responsibility. To waste 1000 tonnes of water per cycle would not be responsible. But there are alternatives. Most obviously there could be a reservoir into which the turbine discharged and from which it would draw water to recharge the tower. With a some filtration and topping up and a little disinfectant, it could stay good indefinitely.

If a little is good, then a lot...

One of the attractions of the power cylinder units is that n of them can be linked into the equivalent of n times the height or n times the cross sectional area of the cylinder, according to preference. Ten cylinders accommodating ten 10-metre-long pistons would in many ways be equivalent to one much higher cylinder with a single 100-metre-long piston, but without the forbidding pressure problems.

In short, many of the desirable attributes of such devices are linearly scaleable over wide ranges of sizes, and they offer pretty fair economies of scale too, before the dis-economies of scale begin to bite.

Such principles of scaling could be exploited in many ways and should improve both the versatility and the maintainability of a large  system.

Such a battery of cylinders could be designed easily to accommodate a million tonnes of lead or cast iron in say a square kilometre of land, and to offer building space on top. In cylinders with a working height of 100 metres it should be possible to store something like a million megajoules by raising a million tonnes, something like 300 gigawatt-hours, enough to run quite a large city for a week or two.

After all, hubris is one of humanity's major virtues, not so?

More to the point such stored energy could buffer realistic power generation shortfalls for months of routine operation perhaps. For that sort of function a square kilometre is pretty compact, especially when the area above the battery need not be wasted, but be available for office space or industrial construction or the like.

Getting More Serious About Practical Problems

The devil is nowhere more present in the details, than in simplicity. Let us consider a few complications.

To begin with, you object, if I think of suspending 1000-tonne lead weights at the tops of towers 100 to 200 metres high and deep, I surely should be running for president and building walls around national borders, instead of wasting your time with such nonsense.

Even constructing such a lump on the ground would be a non-trivial task, let alone manipulating it. It would not be just a scaled-up version of casting fishing sinkers, even if one tried to do it in situ. And what about the necessary precision? If leakage past the piston were not completely trivial, the system could hardly store the output of a hamster wheel for a day.

True.

But once again, what might be in the details?

Let's start with the concept of precision. It shouldn't be too bad a problem. Given that the engineers will have designed the towers to take the necessary pressure, I suspect that a steel or a pre-stressed concrete cylinder lined with steel sheet, and either polished or lined with a suitable, carefully smoothed polymer, possibly epoxy or silicone, would do very well. 

Then the slug — it need only fit to within a mm or two for a thoroughly manageable manufacturing job. But supply it with the equivalent of piston rings of suitable polymer all the way up, and the fit could be very tight and resilient. I suspect that for wear resistance and low friction self-lubrication, piston rings of ultra-high-molecular weight polyethylene (UHMWPE) would last indefinitely. With piston rings of such materials, no gap, such as we have in the typical metal piston ring, should be necessary.

But again, those are details.

And yet, you suspect that I still am skirting the problem of handling thousand-tonne slugs, don't you?

True.

But divide and conquer, say I.  Fabricate the floor, walls, and top of the piston of suitable steel. Whether our steel be stainless, or coated with a polymer such as a suitable epoxy, or both, I do not mind. It should supplied with a suitable lid that is proof against leakage of the cylinder's working fluid, and it should be installed empty, but complete with piston rings. Such a unit shouldn't weigh more than a few tonnes. Once installed in the cylinder, it then could be floated to a comfortable depth for the operation, and then it could be filled with lead.

We had better never want such a completed piston taken out for maintenance you say? Maybe, but that was the good news. Such a piston filled with a thousand tonne slug of lead would be a long term disaster waiting to happen, maintenance or no maintenance.

Well, in my opinion lead (or just possibly cast iron) would in fact be the right material, but we need not think in terms of using it as a single slug. There are two obvious ways of filling the piston without having to destroy it to remove the content: either pack it with lead segments, or fill it with lead pellets or rods, then fill the gaps with a suitable grease, soft wax, or heavy oil.

Now, the lead pellet idea is not so attractive because of the low density to be expected, though a suitable choice of combinations of pellet diameters would have attractions if the pellets were packed with vibration to settle them closely under the oil filler.

Instead of the pellets and their attractions, I prefer the idea of large metal segments. A notionally perfect packing would be possible, with segments of shaped lead of convenient size for handling, say a tonne or perhaps just 50 kg each. Prisms could be cheaply extruded, possibly on the spot, though that is a detail, and installed in vertical arrangement in the steel piston. Special sections could be installed to fill in peripheral gaps. Oil could fill the spaces, ease the fitting and remain to protect the metal.

My own preference however, would be for pieces in the shape of sectors of circular slabs, of a size to fit into the inner space of the piston like pie slices. In narrower cylinders entire disks might be better. Units might be contoured for sophisticated fitting, though I doubt that would be necessary or desirable. If a central vertical passage were to accommodate a closed pressure tube, it would be possible to expand such a tube to force the sectors against the steel piston wall to force a good contact between the piston rings and the cylinder wall. Whether that would be important in practice, I cannot guess in advance. It might be better instead to fill the packed piston's ullage with oil or grease under pressure. But in any case such segments should give the piston a density practically as good as solid metal.

The thickness of the pie-slice sectors would depend on the most convenient mass to handle; I suspect that pie slices weighing several tonnes each would be convenient, conveyed by a crane and positioned by hand. The upper face of each slice would have a hollow above its centre of mass to engage the lifting cable; for the rest it would be as smooth as practical. As the best design would not have the upper and lower face of equal shape, only the upper face has such a recess, thereby rendering incorrect installation less likely. Given a gap of diameter 3.6 metres in the piston into which the lead must fit, each metre of thickness would weigh one hundred tonnes. A layer 20 cm thick would weigh twenty tonnes. Split that into ten slices of 36 degrees, and each slice would weigh two tonnes; a convenient figure for a man to push around for fine positioning at the end of the lowering cable, but if desired, the thickness and angle of the slices could be adjusted either up or down.

But, still assuming two-tonne slices, five layers of ten slices would be fifty slices in a 1-metre layer of lead, weighing a total of 100 tonnes, and a thousand-tonne ten-metre piston would require 500 slices. Tedious, but not forbidding. Installation of the lead, or removing it for maintenance, should take less than a week I reckon. For really large installations with multiple cylinders and pistons, robot gantries instead of human controlled ballast handling probably would be more economical and safer, both for installation and for maintenance.

The speed of laying the slices, plus the precision of placement and stability of each layer could be improved by so contouring and lubricating the upper and lower surfaces of the slices that they would mate and instantly settle into position as soon as they were put down anywhere within several cm of the proper position, and that position should alternate in angle between layers so that the slices in each layer would be held firmly in place without slipping or sliding about and each slice would lie above the contact line between the slices immediately beneath, and not immediately above any single slice. The options for suitable patterns are very wide and trivial to design, so I'll not discuss them here, except to remark that the base of the piston container should match the underside contour of the slices.

The steel jacket could be proof against accidents that would ruin a solid lead piston.

Once installed in the cylinder on top of the working fluid, the piston should never have anything to do but float up and down, passively, slowly and smoothly without much friction to speak of. Its working life without maintenance should be indefinite, even if the fluid needed occasional replacement. 

A possible exception might be when the piston must be held in position either for some maintenance work or for static storage of energy when there is no room to store more and there is no demand. It seems to me that the best way to handle this would be with sets of static detents in the form of say, rectangular steel beams recessed radially into the cylinder walls at appropriate heights. At need they could be projected say, some tens of cm into the cylinder. Two detents in each set would be adequate in theory, though three would be better, and six or eight better still, but again, let the engineers decide. The faces of the detent beams would be contoured to match the inner surface of the cylinder to let the piston pass harmlessly when they are not in use. Their vertical dimension would be far less than the length of the piston, and preferably less than the width of the piston ring face, so that either damage or the vertical escape of fluid past the recess should be minimal.

Mole's eye view

Mass and height are crucial to the function of the units, but in other respects most of the system is very simple. But even simplicity entails complexity in unexpected ways.

To begin with, we are speaking of a lump of lead reciprocating in a smooth cylinder, than which not many ideas could be simpler. Still, we must recognise that one complication already touched on is that of scale. To store and deliver interesting amounts of energy we need a big lump of lead and a big cylinder, both in diameter and height. And the cylinder also needs to stand up to sizeable internal pressures. Even at modest pressures such as twenty bars, one would not like to be too near a cylinder a few metres across if that wall ruptured.

But large, precise, rupture-proof walls are expensive.

One way to reduce the risk would be to drill vertical shafts into hard ground to accommodate the lower regions of the cylinder, preventing any dramatic ruptures. And the deeper the shaft, the more energy could be stored. However, shaft sinking is an expensive activity, so one would not wish to dig more deeply than necessary. And no matter how deep the shaft, the higher one could extend the cylinder economically above ground, the greater the economy and usable scale. And one could increase the practical height by retaining the rocky spoil from the shaft to pack round the base of the above-ground part of the cylinder. And if we were to increase the scale by constructing a battery of daisy-chained cylinders in a fairly close array, with perhaps five-or-ten metre spacing, we could manage some serious height, say fifty or 100 metres down and up.

One could of course make use of existing excavations such as abandoned quarries, mine shafts, or open-pit mines that thoughtless conservationists tend to be rabid about remediating instead of re-using. (Have a read in Wikipedia under headings such as "Environmental remediation" and "Remediation of contaminated sites with cement".) 

Such opportunities tend to be in short supply, so they should not be regarded as routine resources, but where they are available, and there in no foreseeable temptation say, to fill them with water as reservoirs or lakes, there also is no reason that they should not accommodate cheap, deep, and very large-scale gravitational energy-storage installations, both underground and above-ground, together with whatever facilities could be established beside or on top.

Even where one just uses rock walls for supporting one side of a bank of cylinders several hundred metres high, that could be extremely valuable for reducing costs. The cost of digging recesses into the walls to accommodate cylinders would be only a fraction of the cost of sinking shafts of comparable size. And many abandoned open-pit mines are several hundred metres deep.

And the scope for batteries of cylinders capable of supporting entire major cities for weeks if necessary, could transform the prospects for exploiting renewable energy resources such as wind, wave, or solar.

Admittedly, though there are major opportunities for similar facilities at sea, I am interested mainly in inland facilities, because I think that submarine tents offer better and less resource-hungry prospects. See for example: http://fullduplexjonrichfield.blogspot.co.za/2011/01/energy-storage-renewable-energy-sources_04.html

But it is not sensible to omit notional prospects out of hand, so ...

If it still don't work, I gets a bigger 'ammer

The entire principle so far has been to raise the largest possible mass to the greatest possible height (Balloons anyone? Dirigibles?) and dropping it to the greatest possible depth. The hydraulic approach has special advantages for driving turbines, but it is not the only promising principle. For instance, instead of inert pistons, we could have linear motors/dynamos inductively coupled to circuitry in the cylinder walls. Notionally that would obviate the need for any fluid at all, and even the need for any cylinder at all — all sorts of shapes would do better, and no piping and the like — but that is another story and would require more adventurous technology. For present purposes I shall stick to discussing the more pedestrian option of hydraulic power transmission.

However, the constraints include the cost of building high, and the cost of digging deep. Also there is the cost of attaining the strength and the diameter of the cylinder.

Other constraints are the mass of fluid, and the size of the piston. So what do engineers do with the parameters during the conceptual design phase? They juggle the parameters and try to optimise the compromises for given applications. If we can increase the mass of the lead sufficiently, we can reduce the height and depth and even the diameter of the cylinder without affecting the output.

Well, why should we have a top on the cylinder at all, beyond what is necessary to keep out weather and dirt? If the piston is allowed to project out of the top, then there is no immediate limit to its length, so there is no reason it should not be as long as the cylinder, filling the cylinder completely when it reaches the bottom. This yields ten times the pressure that a column of water would, and twenty times the energy.  There are of course limits to how high the column could reach without overbalancing, but they would not be very close, and struts to support the column would be relatively cheap.

On the same principle, there is no reason the piston should not be still longer than the cylinder, thereby increasing the mass without having to dig deeper or build higher.

In case one wished to avoid really high structures however, then instead of making the piston simply cylindrical, the part that never descended into the bore could extend out sideways, instead of vertically. Triple the diameter, and it would be nine times as massive as the same diameter of shaft, and easier to balance.

The engineers designing the shaft would of course have to balance the choice of materials and the structures for safety, reliability, durability, and manageability, but that is what engineers are for.

Bless the dears!

 

What's your tipple?

The choice of fluid in the system is not necessarily automatic. In principle it need not be a liquid, and even some gases would do. Of course, I do not deny the possibility of using any gas at all, and elsewhere I have agitated for the use of air for underwater power storage. However, gases do have their shortcomings. For our purposes the first one is that they are not very dense, and we really value density in our working materials. Another is that compressed gases tend to be more dangerous than pressurised liquids, because in the event of a rupture they expand explosively, hurling debris dangerously for long distances. Thirdly, because of their ability to waste power in expansion and contraction, they entail many inefficiencies.

So, suit yourself, but I shall not consider gases seriously in this application.

I have mentioned water from time to time, for example where it is available for the pumping, such as beside the sea or similar water bodies, and water, especially free water, definitely has its attractions. A slightly concentrated brine with a small admixture of biocides such as zinc and boron salts should remain indefinitely clean and safe to handle, but however well you looked after it, it would inevitably present problems of corrosion. It also is not a very attractive heavy-duty lubricant.

So though I do not deny its possible value, I am not enthusiastic about the use of water.

Instead of water, we could use a suitable organic fluid, say diesel fuel or kerosene; it would be only about 0.8 times the density of water, and would cost a good deal more, but it could have its attractions all the same. For example, some such fluids are more mobile than water, requiring less energy to pump. Also, at the end of its period of use, such liquid still would retain much of its original monetary value. And with suitable selection of structural materials, suitable organic liquids could aid in protecting and lubricating seals and moving parts. It would be interesting to contemplate the use of semi-solids such as greases instead of liquids, but that is a matter of engineering detail that we need not discuss yet.

If we were willing to work with suitable materials under suitable pressures at low enough temperatures, we even could think in terms of liquid SO₂ (density 1.4) or the like. However, having had to deal with quite modest quantities of SO₂ myself in my murky past, I hesitate to recommend it to anyone in quantities of tens to thousands of tonnes.

Chlorinated organic compounds such as CCl₄ and larger molecules, preferably non-volatile, have their own attractions and are comparatively cheap and fireproof, but I do not offhand know of any that is at once cheap enough and otherwise attractive. Also, in spills, they could be very troublesome to clean up in comparison to simple hydrocarbons.

Of course, even oils are destructive to various materials, and they produce vapours that are harmful to workers and present fire hazards, but such problems are well understood and manageable.

Subject to good arguments to the contrary, my favoured option remains something along the lines of a light oil mix that will not present problems of freezing or leaking at ambient working temperatures. 

 

Illustrations of some possibly useful configurations

To illustrate some of the principles, I here present some strictly schematic representations of notional designs. They are of course not even vaguely to scale and do not represent viable mechanisms, but I hope they prove helpful to readers who found my wording confusing.

I omit some versions, especially some that I regard as unpromising, such as telescopic compound pistons. However, they should not be dismissed out of hand; See for example: https://fullduplexjonrichfield.blogspot.com/2017/04/heavier-duty-banking-appendix-supplement.html

 

This is a simple minded version in which the weight of the fluid and the lead combine to give a high output of energy per cylinder. It is however not  very compact. Its pressure drops as it empties, but not unacceptably.

Variation on the open circuit, with less constraint on cylinder height and piston mass, which could be any length within reason, filling the cylinder, or longer, even mushroom-shaped, see below. The greater mass of lead raised as high as possible gives very high energy capacity.

This version uses the fluid almost purely to raise the piston; its contribution to the mass is negligible and its volume minimal. The cylinder is as narrow as practical and its wall must be proof against the much higher pressures than other designs achieve. One advantage is that it could greatly reduce the required volume of working fluid.

In some ways the closed circulation piston is more efficient. For one thing, it is more compact in terms of the real estate it occupies, because it requires no reservoir. On the other hand the raised fluid does not contribute to the gross output, but only to the raising of the next batch of fluid. It does however have two  merits: it contributes to the output pressure; and it keeps the working output pressure very constant. It also improves the options for daisy-chaining cylinders by feeding the output of upstream cylinders in above the downstream cylinders (see below). Because the unit is sealed, it is suitable for fluids other than water, such as hydrocarbons.

This is one approach to the design of batteries of cylinders, such as could be used in abandoned open-pit mines. In the configuration  shown here, the cylinders are as deep as may be practical, partly in drilled shafts, and partly covered by spoil from the drilling. They are sealed. The outputs from cylinders may be directed variously to impose pressures in headspaces of downstream cylinders in series, or yield power in parallel with other cylinders.