Search in book...
Toggle Font Controls
Create new playlist

Name your new playlist

Playlist description (optional)
Sign In

Email address

Password

Forgot Password?

or

Continue with Facebook

Continue with Google
Sign Up

Full Name

Email address

Confirm Email Address

Password

or

Continue with Facebook

Continue with Google

Chapter 7

Underwater Compressed Air Energy Storage

Andrew Pimm

Seamus D. Garvey Faculty of Engineering, University of Nottingham, Nottingham, United Kingdom

Abstract

At the center of every compressed air energy storage installation is the vessel, or set of vessels, that retains the high-pressure air. Normally, high-pressure air storage also dominates the cost of the installation, and its characteristics play a key role in determining performance. Underwater storage of pressurized air is characterized by three important attributes: (1) it has the potential to achieve very low cost per unit of energy stored, (2) it naturally tends to exhibit an isobaric (constant pressure) characteristic of pressure versus fill-level, and (3) in stark contrast to underground air storage, it is feasible in many locations to establish economically competitive compressed air stores at relatively small scales—measured in tens of megawatt-hours rather than gigawatt-hours.

Keywords

CAES

UWCAES

offshore energy storage

energy bags

isobaric

1. Introduction

Compressed air energy storage (CAES) is an energy storage technology that is centered on the concept of storing energy in the form of high-pressure air. The offshore environment provides several ideal conditions for storage of compressed air. By storing pressurized air in an underwater vessel the pressure in the air can be reacted by the surrounding water, greatly reducing loading at the air/water barrier. This simplifies the design requirements of containment vessels and presents potential for their inexpensive manufacture, relative to the cost of a vessel required to handle full pressure without the support of surrounding water.

The density of seawater is around 1025 kg m^–3, less than half that of the Earth’s upper crust (around 2700 kg m^–3), meaning that the storage pressure under a given depth of water is less than half that under the same depth of rock, assuming that compressed air is stored at a pressure roughly equal to the surrounding hydrostatic/geostatic pressure. However, unlike underground stores that require the mining of a cavern, underwater CAES (UWCAES) stores can be manufactured and installed at potentially low cost, though the method of installing and securing the vessels to the seabed requires some development. UWCAES vessels are hidden from view, and suitably deep water can readily be found close to many coastlines, not only in saltwater oceans/seas but also in some freshwater lakes (e.g., Loch Morar in the UK, which has a maximum depth of 310 m). Compared with geological CAES, UWCAES also has the advantage of isobaric characteristics.

In an underwater vessel, compressed air is stored at approximately the same pressure as the hydrostatic pressure in the surrounding water, so the water provides the reaction to the pressure of the compressed air, and the storage vessel can be very low in cost. The load that must be reacted is primarily that of buoyancy, which is proportional to volume, and as a result the cost of a vessel for UWCAES is roughly independent of depth. While the vessel cost is roughly independent of depth, energy storage capacity increases with depth, so it follows that the deeper the vessel is fixed, the lower is the cost per unit of energy storage capacity (in other words the deeper the vessel, the better).

The gauge pressure in seawater at a depth d is given by:

$p = ρ_{sw} g d$ $p = ρ_{sw} g d$

(7.1)

where ρ_sw is the density of seawater (typically 1025 kg m^–3) and g is acceleration due to gravity (9.81 m s^–2). Using equations from chapter: Compressed Air Energy Storage, it is possible to obtain curves of energy density against depth for an underwater compressed air store, assuming the air is stored at a pressure equal to that of the surrounding seawater. These curves are shown in Fig. 7.1. In both cases (isothermal and adiabatic expansion) the energy density increases more than linearly with depth because of the compressibility of air.

Figure 7.1 Energy density available in a UWCAES vessel against depth.
Black solid line, adiabatic expansion with exhaust at ambient temperature; red dashed line, isothermal expansion.

At 500 m depth the energy density is between 5.6 kW h m^–3 and 10.3 kW h m^–3, depending upon how the air is reheated before/during expansion. The lower limit on energy density at this depth is over three times the energy density in the 600 m high upper reservoir at Dinorwig pumped storage plant in the United Kingdom. At depths of the order of hundreds of meters, wave action decays to a very low level and the seabed is typically bare. There is a lack of information on biofouling patterns with depth, and systematic studies are commonly restricted to man-made structures within 150 m depth [1].

2. Storage vessels for UWCAES

Storage vessels for UWCAES can be split into two categories based on their walls: flexible and rigid. Both have advantages and disadvantages, as explained later.

2.1. Flexible Vessels

Flexible vessels (also known as “energy bags”) are made of coated fabric serving as the air/water barrier, with reinforcing straps to carry the main buoyancy loads. They can be manufactured in a variety of shapes, though development has focused on vessels with a single point of anchorage (Fig. 7.2) whose inflated shape is very similar to that of hot-air balloons [2]. Such vessels are very similar in design to lift bags used for underwater salvage operations and must be anchored to the seabed using gravity base anchors, seabed piles, or other anchorage means. Similar to lift bags a cluster of several balloon-shaped vessels can be attached to a single anchor point [3].

Figure 7.2 A 1.8 m-diameter energy bag prototype undergoing tank testing at the University of Nottingham, 2011.
The base of the bag can be seen attached to the base of the tank; in a real-world offshore installation the base would be attached to a gravity base anchor, that is, a large weight, or a pile installed in the seabed.

With careful manufacture, including welded seams, flexible vessels can be watertight and reliable over long periods; fabric air lift bags have been kept underwater for periods of several years and still remained functional [4]. While leakage can be kept to levels so low that the effect on roundtrip efficiency is negligible, it must be possible to prevent any buildup of water at the bottom of both flexible and rigid vessels, which could otherwise cause the vessel to fill up with water and reduce the storage volume. Therefore, both flexible and rigid vessels must have some kind of port in the base to insure that water can be forced out by internal pressure. In flexible vessels, this can take the form of a pressure relief valve, a nonreturn valve, or simply an open port; with the latter, water is allowed to enter the inside of the vessel during discharge, though this is not necessarily a bad thing as the water will be forced out of the vessel again during charge. All of these will also act as pressure relief devices to combat unintended overpressurization, (e.g., due to a valve in the airline that is stuck in the open position, or a pressure regulator accidentally set to a dangerously high pressure), though very small ports may not be large enough to allow air to flow out of the vessel at the rate necessary to prevent a dangerous buildup of pressure.

The airline that connects a vessel to the compression and expansion machinery must be attached to the top of the vessel. If the pipe were connected to the base of the vessel then any water ingress into the vessel (either because of leakage or because of an open base, as mentioned previously) would enter the pipe and block it, potentially rendering the vessel unusable until the water can be removed from the pipe.

Ignoring the mass and volume of materials (which are both relatively small for a fabric vessel) the buoyancy force on a flexible vessel is given by:

$F_{b} = (ρ_{sw} - ρ_{a}) g V$ $F_{b} = (ρ_{sw} - ρ_{a}) g V$

(7.2)

where ρ_sw is the density of seawater, ρ_a is the density of air, and V is the volume of the inflated vessel. This buoyancy must be resisted by some type of anchorage the possible nature of which is discussed later.

The buoyancy load can be carried through meridional straps which run from the top of the vessel down to the anchor, so that the fabric only needs to provide an airtight barrier between the air and surrounding water and react to the small differential pressure load. This rises linearly with height above the base of the vessel; the differential pressure across the air–water barrier at height h above the base of the vessel is given by:

$p = (ρ_{sw} - ρ_{a}) g h + p_{0}$ $p = (ρ_{sw} - ρ_{a}) g h + p_{0}$

(7.3)

where p₀ is the differential pressure at the base of the vessel. In an open-based vessel, p₀ can be no greater than zero and h must be modified to be the height above the water level inside the vessel rather than above the base of the vessel; the differential pressure across the air–water barrier at all points below the water level is zero [5].

The differential pressure load is reacted by curvature in the fabric. Assuming a thin wall, which is quite reasonable for a fabric structure, the differential pressure at a point is related to the tension and curvature in the fabric at that point according to the Young–Laplace equation, which is:

$p = \frac{F_{x}}{r_{x}} + \frac{F_{y}}{r_{y}}$ $p = \frac{F_{x}}{r_{x}} + \frac{F_{y}}{r_{y}}$

(7.4)

where p is the differential pressure across the fabric; F_x, F_y are the tensions (per unit length) in the fabric in the directions of the principal radii of curvature; and r_x, r_y are the principal radii of curvature.

This equation shows us that the tensions in the fabric of an energy bag are proportional to the local curvature of the fabric. Therefore, it is advantageous to allow the fabric to form lobes between the meridional straps.

It is reasonable to assume that the cost of the anchorage C_a is directly proportional to inflated volume (itself proportional to some nominal geometric dimension, for example, diameter, D³).

$C_{a} = a D^{3}$ $C_{a} = a D^{3}$

(7.5)

It is also fair to assume that the cost of the meridional straps C_m is directly proportional to D⁴, since a force proportional to D³ must be carried a distance proportional to D.

$C_{m} = b D^{4}$ $C_{m} = b D^{4}$

(7.6)

We can assume that the cost of surface, C_s, is directly proportional to D².

$C_{s} = c D^{2}$ $C_{s} = c D^{2}$

(7.7)

Finally, like the anchorage cost the energy storage capacity E is proportional to D³:

$E = k D^{3}$ $E = k D^{3}$

(7.8)

Looking at the ratio of vessel and anchorage cost to storage capacity, we arrive at an interesting result. The ratio of cost to storage capacity is given by:

$\frac{C}{E} = \frac{a D^{3} + b D^{4} + c D^{2}}{k D^{3}}$ $\frac{C}{E} = \frac{a D^{3} + b D^{4} + c D^{2}}{k D^{3}}$

(7.9)

We can differentiate this with respect to D and set the result equal to zero to find the vessel diameter that minimizes cost per unit of storage capacity, D^*.

$\frac{\partial (C / E)}{\partial D} = \frac{1}{k} (b - \frac{c}{D^{2}}) = 0$ $\frac{\partial (C / E)}{\partial D} = \frac{1}{k} (b - \frac{c}{D^{2}}) = 0$

(7.10)

$D^{*} = \sqrt{c / b}$ $D^{*} = \sqrt{c / b}$

(7.11)

Substituting D^* into the equations for C_m and C_s shows that the optimal size of a vessel that minimizes cost per unit of energy storage is that which has equal costs of meridional straps and surface. With realistic materials costs, it was shown in [2] that the optimal vessel can cost less than £10 (kW h)^–1 of storage capacity. This is further reinforced by the costs of air lift bags, giving costs for 50 t (metric tonne) vessels of around £15 (kW h)^–1 (not including anchorage).

2.2. Rigid Vessels

Rigid vessels can be built with ballast in mind so that the vessels are heavy enough to remain on the seabed without any additional ballast material or connection to piles in the seabed. In 2005, Davies [6] proposed building modular steel caissons for UWCAES into the base of large offshore wind turbines, and in 2014 Hydrostor installed multiple rigid caissons at a pilot plant in Lake Ontario (Fig. 7.3).

Figure 7.3 Rigid vessels in dry dock, ready for installation at Hydrostor in Lake Ontario, 2014 [7].

Rigid vessels resist loads in bending because they cannot change shape in the same way as flexible vessels, and because of this it is generally considered that flexible vessels can be of lower cost than rigid vessels (Fig. 7.4). However, flexible vessels are more vulnerable to damage due to handling (though this danger can largely be removed with proper handling procedures and durable fabric), and ballast must be provided separately for a flexible vessel. Also, if some of the heat of compression was to be stored for later use inside the compressed air, rigid vessels would lose heat to the environment at a slower rate than flexible vessels because their walls provide more insulation.

Figure 7.4 Rigid storage vessel incorporating ballast.

To avoid large negative differential pressure, it is necessary to ensure that a rigid vessel is open based. As with an open-based flexible vessel, this allows water to fill the vessel from the base upward as the store is discharged and requires that the air pipe is connected to the top of the vessel.

It makes sense to design a rigid vessel as a gravity base structure, so that its resistance to buoyancy comes from its weight and no piles are required. The volume of concrete required in the structure depends upon the anchorage ratio γ (the ratio of anchorage capacity to buoyancy), and is given by:

$V_{c} = \frac{ρ_{sw} - γ ρ_{a}}{γ ρ_{c} - ρ_{sw}} V_{a}$ $V_{c} = \frac{ρ_{sw} - γ ρ_{a}}{γ ρ_{c} - ρ_{sw}} V_{a}$

(7.12)

where ρ_c is the density of concrete, or the average density of the walls.

If the structure were made of concrete with a density of 2400 kg m^–3 and submerged in seawater with a density of 1025 kg m^–3, and if an anchorage ratio of 1.5 is used, then for every cubic meter of stored air, 0.37 m³ of concrete is required. This means that in a storage vessel at 500 m depth, every megawatt-hour of energy storage capacity would require between 36 m³ and 66 m³ of concrete (depending on whether the system was adiabatic or isothermal).

It is possible to incorporate ballast into the design of rigid vessels in other ways. One proposal by Seymour was to create a vessel comprising an air storage unit at the base, above which is an open container filled with rocks or other ballast material [8].

3. Anchorage and installation

Vessels for UWCAES require anchorage capacity in proportion to their storage capacity, and useful plant sizes require significant amounts of anchorage. Economically attractive anchorage and installation techniques therefore provide the biggest challenges facing UWCAES. As mentioned previously, rigid vessels can have mass built into the design and so may not require other anchorage means, but this is not the case for flexible vessels.

The simplest method of anchoring a buoyant object to the seabed is using gravity base anchors (GBAs, also known as “deadweights”). A GBA is simply a large weight that holds the vessel on the seabed. To maximize the lifetime of a GBA underwater, metals should be avoided where possible and so concrete, sand, and rocks make an obvious choice. One interesting idea for GBAs is to design them so that they can be prepared elsewhere then towed to site, using the air storage vessel as support. This approach would require ballast to be added to the GBA before it is sunk; the simplest way to do this would be to tow the sealed GBA to site with air inside, and then open a port in the top of the GBA once it is in position above the site, thus allowing it to fill with water. The GBA/bag combination would then start to sink, and descent must be controlled using a cable attached to a crane or winch on a workboat while also blowing air into the bag (e.g., using a compressor on the workboat), allowing the excess air to flow out of the base of the bag as bubbles. Air must be blown into the bag because as the GBA/bag combination is lowered the surrounding water pressure increases, and thus the bag will tend to reduce in volume and buoyancy unless more air is added. This approach is laid out in Fig. 7.5.

Figure 7.5 Installation technique with gravity base anchor allowing ballast to be floated to site within a container [9].

Such a GBA could then be floated at a later date (e.g., for recovery) by connecting an air pipe to the base of the GBA and filling it with air. It would also be necessary to inflate the energy bag and insure that there are holes at the base of the GBA and energy bag to allow air to flow out as the combination rises into lower pressure water.

Alternatively, anchorage can be provided through piles installed in the seabed. These can be hammered, screwed (screw, or helical, piles), dropped (torpedo piles), or sucked (suction cup) into the seabed. Screw piles and suction cup anchors are of particular interest for UWCAES since they both minimize reliance upon friction and screw pile, in particular, are under consideration for offshore wind foundation applications. They comprise a number of helical steel plates attached to a steel shaft and they are installed into the seabed by applying a torque to the shaft. The anchorage capacity of a screw pile increases with the number of helices, the helix diameter, and the sand compactness [10]. Research into their large-scale application in the seabed is ongoing.

4. System configurations

Discussion in the chapter has so far mostly covered the nature of compressed air stores underwater. We now look at the machinery and infrastructure required for a UWCAES plant and how this might be configured in terms of location and technology. We first consider heat management.

As discussed in chapter: Compressed Air Energy Storage, as well as storing energy as compressed air, CAES plants may also store energy as heat. The reason for this is simple: when a gas is compressed, its temperature increases because thermal energy is generated. Part of the work that went into compressing the air has been converted to thermal energy (i.e., heat) and ideally this energy should not be wasted. In existing underground CAES plants the heat of compression is simply taken out of the compressed air and vented to atmosphere as waste heat.

When a gas is expanded, its temperature drops, and so when the compressed air is allowed out of storage it must be heated before entering the turbine (otherwise moisture in the air would freeze, blocking the pipes). In existing CAES plants, compressed air is mixed with natural gas and burned in a gas turbine, emitting greenhouse gases.An alternative to venting the heat of compression to atmosphere and later burning gas to heat the compressed air is to store the heat of compression in a dedicated thermal energy store and later use it to reheat the air as it enters the turbine. Known as adiabatic CAES (A-CAES), this approach improves utilization of the store and reduces (or removes) the need to burn gas. The Adiabatic CAES Storage for Electricity Supply (ADELE) project, currently under construction in Germany, will be the world’s first commercial A-CAES plant [11].

Another choice of system configuration that relates to heat management is the number of stages of compression and expansion, and what intercooling/reheating takes place between compression/expansion stages. UWCAES systems can take many different configurations, including those considered or used for underground CAES (such as single/multistage and adiabatic/diabatic), as effectively the main difference between an underwater CAES system and a fixed volume underground CAES system is the pressure characteristic, which only serves to improve the roundtrip efficiency of an underwater system over the underground equivalent.

In a UWCAES plant the ideal choice of location for the conversion machinery and any heat storage depends upon the distance between the vessels and shore and the relative costs of transmitting power as electricity in a cable and as high-pressure air in a pipe. For nearshore plants, it makes sense to site the machinery onshore, making it easily accessible and removing the need for any underwater power transmission. In such cases an air pipe runs between the machinery and the storage vessels. In some parts of the world there are limited locations where deep water, for example, >400 m, is found within a few kilometers of shore, so in those locations it would make more financial sense to site the machinery on a floating offshore platform moored above the storage vessels, with an air pipe running between the floating platform and the vessels, and a power transmission line running between the floating platform and an onshore electrical substation. Locating the compression and expansion machinery on the seabed could reduce the air transmission distance and hence increase roundtrip efficiency, but it has not been seriously considered because of the sealing requirements and, in particular, the immense difficulty this would introduce to access and maintenance (Fig. 7.6).

Figure 7.6 Shore-based and platform-based machinery.

In a near-shore UWCAES plant with onshore conversion machinery a thermal energy storage unit could take exactly the same form as the thermal energy storage unit for underground CAES plants. In a far-offshore plant a thermal energy store would likely be located on the same platform as the conversion machinery. In such cases the size and weight of the required thermal energy stores must be given serious consideration. The weight of thermal energy storage material required for a given storage capacity can easily be calculated: assuming that granite (c_p ≈ 800 J K⁻¹ kg⁻¹) is used to store heat at 500 °C above ambient, 9 kg of rock would be required for a kilowatt-hour of thermal energy storage capacity. For 100 MW h of thermal storage capacity, 900 t of rock would be required, taking up about 550 m³ (assuming a packing factor of 60%). While the weight of machinery and containment would also need to be taken into account, this kind of size and mass is not excessively large in terms of the capacity of seafaring vessels. In deep water a tension leg platform could be used for this task, similar to those used for offshore oil/gas platforms (Fig. 7.7).

Figure 7.7 Temperature after single-stage adiabatic compression from 283.15 K.

Various aspects of the offshore environment can be advantageous for thermal energy storage. One interesting proposal for offshore thermal energy storage made by Ruer [12] is to locate a pressurized heat store underwater. This could be used in a system incorporating direct contact between the hot compressed air and the thermal storage medium and could use the pressure of the surrounding water to provide resistance to the pressure of the air passing through the heat store. The concept was originally developed for coastal underground CAES but would be equally applicable to underwater CAES (Fig. 7.8).

Figure 7.8 Proposed offshore adiabatic CAES system with underwater thermal storage [12].

One other aspect of the offshore environment that could be beneficial to UWCAES is the fact that the sea effectively provides an infinite thermal sink to aid near-isothermal compression, with water also having much higher thermal conductivity than air [0.58 W (m K)^–1 compared with 0.024 W (m K)^–1].

5. Locations

Suitable locations for UWCAES are those with deep water close to shore. In a study of coastal waters with depth greater than 400 m around Europe and North America, it was found that many locations have such depths within reasonable distance of shore, particularly islands [13]. An example of suitable locations in California is shown in Fig. 7.9. A minimum depth of 400 m was used because the water pressure at this depth is about 4 MPa (∼40 bar), similar to the inlet pressure of the high pressure turbines at the existing underground CAES plants [14,15]. California is particularly interesting for energy storage because of the mandate approved by the California Public Utilities Commission in late 2013 for 1325 MW of energy storage to be operational by the end of 2024.

Figure 7.9 Map of the central and southern California coastline, showing areas within 5 km of land which have a water depth greater than 400 m.

Most of the suitable locations for UWCAES in Europe are found within the Mediterranean Sea and the fjords of Norway. The bathymetry of fjords is particularly interesting because they are typically characterized by a phenomenon known as overdeepening, whereby glacial erosion has caused them to become deeper than the surrounding sea. Within the Mediterranean, heavily populated areas with deep water close to shore include the Côte d’Azur and the Amalfi Coast. Fig. 7.10 shows the areas in the central Mediterranean Sea which are within 5 km of land and have a depth greater than 400 m.

Figure 7.10 Map of the central Mediterranean Sea, showing areas within 5 km of land which have a water depth greater than 400 m.

Table 7.1 shows the area of water with depth >400 m within the territorial waters of each country in Europe, categorized by distance from shore. Almost all of Norway’s vast areas of near-shore deep water are found within the fjords. Aside from Norway, it is typically the case that island nations have the largest areas of deep water close to shore.

Table 7.1

Areas of Water Within European Territorial Waters with a Depth Greater Than 400 m, Categorized by Distance from Shore

Area with depth >400 m/km²
Country	0–22.2 km	0–10 km	0–5 km	0–2 km	0–1 km
Norway	4 787.19	2 692.07	1 810.26	733.01	149.43
Greece	76 346.16	22 473.24	5 289.61	440.02	28.07
Azores	19 561.23	5 746.84	1 472.01	168.61	27.92
Canary Islands	23 340.65	5 746.81	1 171.44	51.42	16.70
Italy	37 577.86	8 460.96	1 663.27	139.94	15.40
Turkey	22 813.23	5 385.08	725.03	27.48	2.26
France	7 934.15	1 887.44	430.85	25.65	1.23
Madeira	9 141.76	2 444.39	513.05	19.91	0.49
Georgia	2 123.36	512.82	99.98	3.83
Gibraltar	231.24	75.90	17.90	0.95
Portugal	1 583.32	159.82	21.83	0.45
Russia (to 65°E)	3 173.24	269.55	23.20	0.43
Cyprus	10 173.09	2 593.82	445.98
Spain	12 645.34	1 039.71	67.38
Monaco	54.33	14.07	1.02
Albania	585.17	51.44
Ukraine	1 182.22	48.80
Malta	970.83	41.95
Iceland	411.76
Croatia	107.13
Totals	235 783.68	59 934.84	13 823.85	1 612.52	241.51

6. Cost and efficiency

With suitable pipework the roundtrip efficiency of an underwater CAES plant can be very similar to that of an underground CAES plant. The effective roundtrip efficiency of a CAES plant was discussed in chapter: Compressed Air Energy Storage. It is not unusual for compressors to achieve adiabatic efficiencies of 88% and for turbines to achieve efficiencies above 92%, so the basic elements of the system point to a roundtrip efficiency in excess of 80% being achievable.

Assuming that compressed air is stored at a similar temperature to the surroundings—as is the case at Huntorf and at McIntosh), the additional losses introduced by underwater storage are those associated with leakage and pressure drop. With a well-manufactured vessel, leakage losses should be small. The leakage rate in the first prototype offshore flexible vessel was estimated at 1.2% per day at the end of a 3-month deployment [16], with air loss mainly found around stitching of seams and areas of repaired trauma due to handling. It is anticipated that future deployments would benefit from improved handling procedures and more heavy-duty material and so would not be damaged on deck.

The pressure drop in the pipework is likely to be the main source of loss in a UWCAES system. Table 7.1 shows the pipe diameter required for various percentage pressure drops per kilometer of pipe in a UWCAES system taking cool compressed air at 100 MW. Two points should be noted here: (1) the mass flow rate of air is determined here by pessimistically assuming that the air will be expanded isothermally, and (2) the actual losses in output power are much lower than the losses in pressure. As an example of the latter point, at 2.0 MPa (20 bar) a 1% drop in pressure causes only a 0.33% drop in output power, and at 7.0 MPa (70 bar) a 1% drop in pressure causes only a 0.24% drop in output power.

Table 7.3 gives the pipe masses corresponding to the diameters given in Table 7.2, assuming a 10 km length of cylindrical steel pipe with density of 7800 kg m^–3, working stress of 100 MPa, and a constant wall thickness. Associated costs are also given, assuming a cost of £2000 per tonne of steel (about five times the cost of the raw material to account for welding and other working). The pressure drop has not been taken into account, as to fully account for its effects on air density (and hence net buoyancy) would require knowledge of the seabed profile along the length of the pipe, which varies by location. In reality because the highest differential pressure across the pipe would be seen by the parts closest to the surface of the sea it would be possible to reduce the total pipe mass by using a pipe whose wall thickness reduces in stages as it reaches greater depths.

Table 7.2

Pipe Diameters Required to Transmit 100 MW of Cool Air for Various Percentage Pressure Drops

		Required pipe diameter D/m
Pressure (10⁵ Pa) (bar)	Mass flow rate/(kg s^–1)	0.05% loss per kilometer	0.2% loss per kilometer	1% loss per kilometer	4% loss per kilometer
20	415.2	2.182	1.648	1.192	0.904
30	365.7	1.761	1.331	0.962	0.730
40	337.2	1.518	1.147	0.830	0.630
50	317.9	1.356	1.025	0.741	0.562
60	303.8	1.237	0.935	0.676	0.513
70	292.8	1.145	0.866	0.626	0.475

Data from Ref. [17].

Interestingly, the pipe masses shown are almost independent of internal pressure because the required pipe diameter for a given pressure drop decreases with greater internal pressure. The only reason pipe mass is not constant for all pressures is because air density (and hence differential pressure across the pipe wall) increases with pressure inside the pipe. While the pipe masses shown are large, it should be remembered that the pipe connecting the vessels to the machinery is also a store of compressed air.

The underwater pipes would need to rest on the seabed, so it is necessary to take into account the pipe buoyancy and pipe masses to calculate the net buoyancy of the pipes, which must be counteracted with ballast. The net buoyancy of the 10 km-long pipes referred to in Tables 7.2 and 7.3 is shown in Table 7.4. Clearly higher storage pressures used in deeper water allow narrower pipes to be used for a given pressure drop, hence reducing the net buoyancy of the pipe.

Table 7.3

Pipe Mass m and Approximate Costs Assuming 10 km Length of Steel Pipe, Calculated Assuming Constant Wall Thickness from Surface Down to Storage Vessels

	Pipe mass for 10 km length/t
Pressure (10⁵ Pa) (bar)	0.05% loss per kilometer	0.2% loss per kilometer	1% loss per kilometer	4% loss per kilometer
20	11 792	6 709	3 502	2 011
30	11 463	6 540	3 412	1 963
40	11 334	6 465	3 383	1 948
50	11 293	6 449	3 368	1 937
60	11 270	6 436	3 363	1 936
70	11 261	6 439	3 364	1 936
Approx. cost/£	23 × 10⁶ (23 m)	13 × 10⁶ (13 m)	7 × 10⁶ (7 m)	4 × 10⁶ (4 m)

Table 7.4

Net Buoyancy m* in tonnes for 10 km of Pipe, Using Pipe Diameters from Table 7.2 and Pipe Masses from Table 7.3

	Net buoyancy, m*, for 10 km length/t
Pressure/(10⁵ Pa) (bar)	0.05% loss per kilometer	0.2% loss per kilometer	1% loss per kilometer	4% loss per kilometer
20	25 590	14 615	7 654	4 405
30	12 576	7 193	3 762	2 168
40	6 300	3 602	1 889	1 089
50	2 595	1 487	779	449
60	135	80	43	25
70	−1 620	−924	−482	−277

The existing underground CAES plant at Huntorf has the machinery house located almost directly above the caverns, so for each cavern there is only 300 m of surface-based pipework and ∼500 m of downpipe. UWCAES plants with offshore machinery could have the machinery located on a platform directly above the storage vessels, and so would only require a pipe directly down to the seabed. Near-shore plants would clearly need to have a longer pipe run than offshore plants and as such would have greater transmission losses.

UWCAES has an advantage over underground CAES in that the pressure characteristic in an underwater CAES vessel is roughly isobaric. This contrasts with a fixed volume containment (such as a cavern), in which the air pressure varies strongly with the amount of stored energy. As such, CAES plants that use underground caverns must incorporate pressure regulators to ensure that the turbine inlet pressure is below a certain level. Pressure regulators have energy losses associated with them, losses which are not seen by an underwater system (which needs not use pressure regulation, or certainly does not need such a severe pressure drop). Fig. 7.11 shows a curve of pressure against stored energy for the air in an underwater CAES plant at 400 m depth and the curve for an equivalent plant using a fixed volume container (e.g., an underground salt cavern). The isobaric characteristic of UWCAES is clear. The equivalent underground cavern (of the Huntorf/McIntosh type, with 4 MPa (40 bar) maximum turbine inlet pressure) requires the air to be throttled down to 4 MPa (40 bar) during discharge until roughly two-thirds of its energy content have been removed, and then at lower states of charge the turbine inlet pressure reduces below 4 MPa (40 bar). Both of these characteristics of fixed volume containment serve to reduce efficiency.

Figure 7.11 Pressure against percentage charge for a 10 m-high underwater compressed air store at 400 m depth and an equivalent underground store with maximum and minimum cavern pressures of 7 MPa (70 bar) and 2.5 MPa (25 bar), respectively.

Using reasonable costs for materials, it was shown in [2] that total costs of flexible vessels and ballast for UWCAES can be less than £10 (kW h)^–1 (assuming storage at 500 m depth). Installation and machinery will increase the total cost. Through discussions with a major lift bag manufacturer, it has been found that currently the costs of flexible vessels could be less than £15 (kW h)^–1 at 500 m depth.

The design life of a UWCAES system will be at least 20 years. Concrete structures underwater have lasted much longer than this (concrete reacts well with seawater), and inflatable vessels have been used underwater for significant periods of time without malfunction.

7. State of development

After early development at the University of Nottingham, Hydrostor Inc. of Toronto have pushed forward commercialisation of UWCAES technology, working with researchers at the University of Windsor. A 750 kW h/1.5 MW h pilot plant was installed at a depth of 80 m in Lake Ontario in 2014, and the first commercial plant will be installed off the coast of Aruba in 2015/2016 (Aruba is one of the Lesser Antilles located in the southern Caribbean Sea and is close to the coast of Venezuela). The pilot plant combines flexible and rigid storage vessels.

Other companies working on or considering UWCAES include Bright Energy Storage, Brayton Energy, Exquadrum, AGNES, DNV, and Moffatt-Nichol. One company working on a novel method of moving the air down to the seabed is OC Energy, who propose the use of an inclined Archimedes screw coupled to a surface-mounted electrical machine [18].

The areas that are in particular need of development are installation procedures and the securing of vessels to the seabed. Rigid vessels are designed such that they have ballast mass built into the structure and, to date, flexible vessels have been held to the seabed using gravity base anchors (i.e., large weights, such as concrete blocks on a steel tray) that were transported to site from shore. However, the use of deadweights could become prohibitively time consuming and costly if carried out for commercial-scale installations of large amounts of storage capacity. By way of example, let us consider a 1 GW h store at 500 m depth. This would require up to 179 000 m³ of storage capacity, and hence it would be necessary to react 172 000 t of buoyancy force. With a 2:1 ratio of holding-down force to buoyancy force, 344 000 t of holding-down force would be required. Using concrete as a deadweight, approximately 250 000 m³ would be required to provide this ballast force underwater. To put this in perspective the largest concrete dam in the United Kingdom (at Clywedog reservoir) has a height of 72 m, a length of 230 m, and contains about 200 000 m³ of concrete [19].

Other than using deadweights, alternative methods of securing vessels to the seabed include driven piles, suction piles, suction-embedded anchors, torpedo piles, and screw (or helical) piles. Of these, screw piles are of particular interest because they provide some amount of positive drive. To minimize cost per unit of anchorage capacity, anchorage which can be installed without use of remotely operated underwater vehicles (ROVs) is of particular interest.

8. Concluding remarks

UWCAES is a developing storage technology which is a natural extension of CAES for coastal environments. It is very similar to underground CAES in all aspects but the energy store. Compared with a fixed volume underground store an underwater store brings the benefit of isobaric containment, raising the system’s roundtrip efficiency. Around the world there are many coastal locations suitable for UWCAES, particularly around islands, and UWCAES plants could either have the machinery based on shore or on an offshore platform; decisions over the number of compression/expansion stages and whether to use heat storage or burn gas are largely based on the same factors as underground CAES. The main challenges currently facing UWCAES are cost-effective access (including deployment, maintenance, and recovery) and anchorage, and a number of companies and organizations are working to develop solutions to both.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.

Table of Contents for Chapter 7: Underwater Compressed Air Energy Storage

Create new playlist

Sign In

Sign Up