I don’t know if there has been any experimentation with a hybrid low head pumped hydro, along with CAES. But this may make sense, just to toy around with the idea. Since wind and wave generated compressed air could augment any pumped hydro system. In theory it may be a good marriage. Both the wind and water supply can be fickle, and each could hedge each other, again in theory.

Since the most efficient storage of compressed air, to me, is at depth at sea, then the problem of distance would need to be overcome with pipe diameter, which should not be a problem. At sea CAES means little to no wiring and all the distance issues with transmission. Pipe is cheap, but pipe dreams are cheaper. http://www.environmentalfisherman.com

The major issue, though, is how to manage the necessary ‘slip’ between the turbine rotor moving at full speed as required for grid synchrony and the water which is initially not moving at all. At startup the impeller blades have to move through the near-stationary water, while nearly all their swept volume goes around their edges. During normal pumping, or generation, running at only slightly lower pressure than the startup pressure, there can be only a few percent of the swept volume leaking past the blades if acceptable efficiency is to be achieved. I can’t think how this could be done with a fixed geometry for the rotor and housing. There has to be some way of allowing more bypass during startup. This problem is not unique to this scheme, and must have been solved for other pumped storage sysyems

Too much speculation here, I don’t think we can get much further unless a turbine designer comes along to help.

Good point. The 3 m/s is for discharge not recharge. I haven’t calculated the rate for recharge, but I’ll take your figure as probably correct.

You make a good point about the additional power needed for acceleration (i.e. additional to the power needed for constant rate pumping) is only 2% for 30 minutes acceleration period or 12% for 5 minute acceleration period.

However, I am still not sure about the capability of pumps to accelerate 7 million tonnes of water (per tunnel). I understand the pump’s speed is synchronised to the grid then begins pumping and its speed stays constant throughout. So I am not sure if it is feasible to accelerate the flow over a long period. I know very little about this, but I still wonder about accelerating this mass of water given that existing pumped hydro schemes have only relatively short tunnels.

You’ll notice on the map in the lead article the dog leg in the tunnel alignment. That is a high point in the surface topography and the best location for the surge shaft. I wonder how high the surge shaft would have to be above ground level to accommodate the surge caused by the relatively long duration acceleration and decelerations we are considering. Calculating this would be beyond my capability. Of course, this is just an added cost so it is does not preclude the concept. Also, there may be other ways, such as an underground pressure chamber instead. Again, just an additional cost.

For DBB, variable speed pumps are not used in large pumped hydro faculties, although they may become viable in the future. I’ve commented on some pumped hydro plants in Japan and Europe in previous comments on this thread.

For the nuclear + hydro case, how fast would you want to ramp up? You start with all the reactors running at near capacity and demand falling in the late evening. From the demand graph above the load ramp rate is about 1250 MW/hr, so you start the pumps at 23:00 but have to wait until 02:30 for there to be enough spare power to run at full rate. Then you ramp back down more steeply from 06:00 to 07:30. This scheme is so big, 18% of the average grid load, that it is constrained to move slowly to avoid over-stressing the grid it is meant to serve.

A possible error – is the 3 m/s flow rate for the discharge or recharge step? Back-of-envelope calculation gives 4.5 GW recharge as only 1.5 m/s, so all the energy and power estimates drop by a factor of 4.

The latest in large-scale engineering technology are variable speed machines for greater efficiency.

This is meaningless. What does it mean in the context of the pumped hydro we are talking about. Can I urge you to put the bits and pieces you come up with together. Otherwise it a bit like saying “an Atlas rocket is really big, but a Holden car is one of our biggest cars”. It’s just meaningless bits and pieces of unrelated snippets.

From
http://en.wikipedia.org/wiki/Francis_turbine
the highest head utilized so far is around 700 meters; I have no way to estimate whether operating with a 900 meter head is even possible with currently available materials.

The remaining issue is the valves used to control flow rates while generating. Again that great head height may be a concern. For smaller projects it seems that starting in under a minute is possible [from the first wiki article]. So one possiblity is to provide several smaller pump-generators, each with its own tunnel; only a suggestion.

In any case, I don’t see (so far) any insuperable difficulty regarding the electrical aspects. But as earlier comments have discussed, inertia means that flow is not going to quickly reverse.

The synchronous generators in the dams around here are rescheduled every five minutes and I’m not sure how much less time than that the units could be started or stopped. The ones I’ve seen rotate quite slowly despite generating around 200 MW each. In any case BPA routinely ramps in excess of 400 MW each rescheduling interval (but of course has many individual generators to participate in this load following).

Grand Coulee dam has has six pump-generators to deliver irrigation water (and someday do some peaking power gneration). These are of a design similar to the generators although only about 50MW each.

This is an excellent reality check. Very elegant. Thank you.

= 90 GW.s

Which is a lot, but is also only 10 s of full power output, or 20 s of charging input.

i.e. 9 GW for 10 s or 4.5 GW for 20 s.

In terms of power I think it is more manageable than this because there are three tunnels. I expect they would be started one at a time. Therefore, we’d need 3 GW for 10 s or 1.5 GW for 20 s (if the pumps had the capability to accelerate the mass of water in that time).

However, as you imply, the pumps would not have the power to provide this acceleration in addition to lifting the water against 875 m of hydraulic head. Hence your second “perspective”.

I’ve redone your calculations for the situation with three tunnels. I calculate the time to accelerate the water in each tunnel would be the same as you calculated – say 30 minutes – but the three tunnels would probably not be started in parallel. So the start up time to full pumping rate would probably be longer, and perhaps up to 90 minutes.

This seems to me to be an excessively long duration for pumped hydro. As far as I know the big pumps that would be needed on a project like this have to run at constant speed. So, I suspect a long acceleration time is unlikely to be viable (however, this is outside my area of expertise). I’d note that most hydro plants boast they can go from zero to full power generation in less than a minute. I am not sure how long the pumps take to start. Perhaps someone else could comment.

My gut feeling remains that a pumped hydro scheme with 53 km of tunnel and 875 m of head is not viable. For comparison, Tumut 3 has 488 m of surface penstock (i.e. steel pipes) and 150 m of head. Wivenhoe has 420 m of tunnel and 100 m of head (from memory). Therefore, the Tantangara-Blowring tunnel is more than 100 times as long and the hydraulic head is about 6-10 times greater than Wivenhoe and Tumut 3.

Think about the power and energy needed to accelerate 20 million tonnes to 3 m/s

OK, Energy = 1/2 * m * v^2

20 * 10^6 * 10^3 kg * (3 m/s)^2 = 10 * 9 * 10^9 J = 90 GW.s

Which is a lot, but is also only 10 s of full power output, or 20 s of charging input.
Another perspective, use energy = pressure * volume. If the pumps can manage ~5 bar/50m head/0.5 MPa above the hydrostatic 875 m head to accelerate the water, they need to pump 90 GJ / 0.5 MPa = 180,000 m^3 of water to supply the kinetic energy. At a total cross section of 380 m^2, that’s ~473 m of movement. Call it 0.5 km. If the acceleration is steady from standstill to 3 m/s, the average speed during that time will be 1.5 m/s, so it will take 500/1.5 = 333 s, ~5.5 minutes to start up. This seems unrealistic for something so large, so maybe it’s only 1 bar of overpressure available, for a 30 minute startup. Doesn’t look inherently unworkable on these grounds.

I am currently trying to better understand the economics of pumped storage, albeit I am at a very early stage of that process.

* I am trying get ‘ball park’ costs in $/MW for pumped storage projects (say in the 100-300 MW range), but just for the key generation kit (i.e. presumably using reversible Francis turbines/pumps and the generators). Do you know where I might be able to find such information or who I should talk to in relation to getting access to such information? I have found it surprisingly difficult to find any details. At this stage, I am interested in indicative costs only.

* I also noted that the commentary suggests that the cost of pumped storage would typically increase significantly with smaller capacity. I suppose the question I am really asking is: are the costs for the key generation kit broadly scalable?

My reply:

Thank you for your question. I do not have recent cost figures.

What I can tell you is the cost of pumped hydro is very site specific. By “very” I mean order of magnitude. So realistically, unless you are dealing with a specific site, any cost estimates would be next to useless.

You asked for ‘ball park’ cost in $/MW. That is one of the unit costs. But the other you would need, if your interested is in energy storage for renewable energy, is $/MWh of energy storage capacity. Because, with renewable energy, the important parameter is how much energy you can store, not how much power you can generate.

This site gives a broad range of costs (both $/kW and $/kWh) for pumped hydro in the USA:
http://www.electricitystorage.org/technology/storage_technologies/technology_comparison

Here are some tools that can help with cost estimating, but you would need an engineer with a great deal of experience in hydro to assist with the data input.
http://hydrohelp.ca/eng/home.htm

Lastly, in this paper posted on 10 February 2012
http://bravenewclimate.com/2012/02/09/100-renewable-electricity-for-australia-the-cost/, I made a crude cost estimate for the “100% renewable electric NEM” system analysed by Elliston et al. (2011). My critique of the Elliston et al. paper provides some comments on the Australian pumped hydro storage capacity and the rate that energy can be stored.

Few people realise that pumped hydro has its place when matched with low cost base load power supply, but is most unlikely to be economic for storing renewable energy.

I understand that the costs are going to vary substantially on site specific features, but I suppose I was trying to a handle on those that presumably do not (i.e. the turbines, pumps and generator, but I note you don’t have recent information in this regard). That way, you can work backwards and work out how low your site specific costs (storages, pipes, electricity connection etc.) would need to be in order to make it work, if indeed it can, which seems highly unlikely based on all the numbers I have seen.

I do not have recent costs for the mechanical and electrical equipment. However, I’d caution against the approach where you start by estimating the cost of items that are a small component of the total cost and then look at the sensitivity to the high cost items. Other than adding a small pumped hydro facility to an existing dam, the costs of the civil works will dominate the costs as Table 1 in the Tantangara-Blowring post shows (the pumps, turbines, generators and transformers comprise about 5% of the total cost in that estimate).

I’d tend to start with the item that is the highest cost as the first step. There is no “general” price for pumped hydro. The cost depends on the hydraulic head (the elevation difference between the minimum operating level in the upper reservoir and the full supply level in the lower reservoir), the distance between the outlet from the upper reservoir and the lower reservoir, the length of tunnels or surface pipes, the length and thickness of steel pipes ( a very high cost item), and the minimum amount of storage available in smaller reservoir.

By the way, please note that the Tantangara-Blowering reservoir was purely and exercise to reveal the costs, and what is involved in the costs. The capital cost figure I derived is too low by at least a factor of two, as the two reviewers’ comments at the end of the post show. I expect the cost would be in the order of $15 billion. However, there is no way that the Tantangara Blowering pumped hydro scheme could be viable. It is not viable physically nor financially. It is not viable physically because of the time and power that would be required to stop and start the flow of 20 million tonnes of water moving at 3 m/s in the tunnels each time pumping had to stop and start. Think about the power and energy needed to accelerate 20 million tonnes to 3 m/s. That would be equivalent of having one tug boat at the head of a line of 100 fully loaded 200,000 tonne ships and trying to accelerate them to 11 km/h

Presumably with the carbon tax, the case for pumped storage can only get harder in most Australian jurisdictions, as off-peak prices will increase by much more than on peak prices reducing the opportunity for arbitrage?

Yes. I suspect you are correct on that point.

EL asked:

I don’t see where the authors have widely missed the mark on hydro storage capacity and requirements. 2.2 GW capacity at 20 GWh DOES reflect an upper storage limit …

I’d like to expand on the comment here http://www.oz-energy-analysis.org/feed/show_me.php?comm=OzEA_DG0002 which addressed the question:

What is the total energy storage capacity (in GWh) of Australia’s existing Pumped-Hydro facilities?

The short answer is roughly 5 GWh can be stored per day and 20 GWh total.

EDM-2011 cited this comment on OzEA as the basis of their pumped hydro energy storage capacity, but they did not realise there is a constraint on the rate at which energy can be stored.

In reply to EL’s comment about the rate at which energy can be stored in the three existing pumped hydro plants:

“20 GWh DOES reflect an upper storage limit”. This statement is correct.

“2.2 GW capacity”. No. That is not correct. Wivenhoe is the only pure pumped hydro scheme in Australia. The other two PHES plants have some pumped hydro capacity within a plant that is mainly a hydro plant. EDM-2011 attributed the full generating capacity of these plants to pumped hydro.

Only 0.9 GW of generating capacity can be attributed to pumped hydro. The remainder of the 2.2 GW EDM-2011 assumed should be attributed to hydro. Furthermore, there are energy losses in pumping and generating, so only about 75% to 80% the energy can be recovered.

Wivenhoe pumped hydro energy storage facility:

All the water Wivenhoe PHES plant uses to generate power must be pumped up from Wivenhoe reservoir. It can store energy (that is energy that can be recovered, after losses) at the rate of about 328 MWh per hour. It pumps when demand is low and electricity is cheap – between about midnight and 6 am (longer on weekends). It is on standby (spinning and ready to go to full power within less than 1 minute) for about 12 h per day and generating to meet peak and intermediate demand for about 7 hours per day (at variable power output). It generates a little during standby to balance power and frequency fluctuations in the grid. It cannot pump while it is generating or on standby. Because it is needed to be generating and on standby during the day, and because power is cheap at night, the pumping is done at night when the plant is unlikely to be needed for generation.

It takes a lot of energy to start pumping the water. The weight (~23,000 tonnes) of water must be lifted up 100 m against gravity and accelerated from 0 to 3.6 m/s velocity each time pumping starts. Therefore, it is not economic to repeatedly stop and start the pumping. The pumps are not variable speed so once started they run at their constant pumping rate of 207 m3/s.

Therefore, the pumps should only be started if they can be assured they will pump for at least several hours (e.g. 5 hours, but I said let’s assume a minimum of 4 hours to be economically viable).

If we assume an average of 5 hours of pumping time per day and energy storage at the rate of 328 MWh per h, then Wivenhoe can store about 1,640 MWh per day.

Tumut 3 has 1,500 MW of generating capacity in six generating units but only three have pumps and the power of the pumps is only little more than half the power of the generators. So it is not appropriate to assign the 1,500 MW of Tumut 3 generating capacity to PHES.

Bendeela & Kangaroo Valley – similar comment to Tumut 3.

See more details here:
http://www.oz-energy-analysis.org/feed/show_me.php?comm=OzEA_DG0002”

You need to estmate the cost of your proposal, and provide the basis of your estimate so we can understand it, or it is measningless. Look at the lead article on this thread and the twoo hundred comments on the thread to get some background as to how to approach that.

Without costs, your idea has about as much credibility as a proposal to pipe hydrogen from the Sun.

There have been many comments on this thread about using Australia’s pumped hydro facilities to help make wind and solar power more viable.

A recent paper “100% renewable electricity for Australia’s NEM” by Elliston, Diesendorf and MacGilll http://www.ies.unsw.edu.au/docs/Solar2011-100percent.pdf demonstrates clearly that Australia’s pumped hydro plants could not help renewable energy much at the time it is most needed. (As explained in a number of comments up-thread there are also a number of reasons why the existing hydro facilities’ role and function in the NEM could not be diverted for the use envisaged by the wind and solar power advocates).

The objective of the paper is to show that a 100% renewable electricity system could reliably meet all of the NEM’s demand. In fact, it demonstrates the opposite.

Look at slide 12 in this slide presentation of the simulation study http://www.ceem.unsw.edu.au/content/userDocs/Solar2011-slides.pdf . There is no pumped hydro generation on 1, 2, 5 and 6 July, 2010. Those are the days where Wind + CST + PV could not provide sufficient energy to store energy at the CST plants, let alone in the pumped hydro plants.

This slide reveals much more too. For the period 1 to 6 July, the simulation would need gas generating capacity about equal to the winter peak demand. Ignore the hydo; it is wrong. Contrary to assumptions in the simulation, Australia’s total hydro capacity cannot be run at full power for days and weeks at a time (outside sun hours).

http://physics.ucsd.edu/do-the-math/2011/11/pump-up-the-storage/#comment-1655

Contrary to your statement that Kops II has a vertical shaft and the underground power station is located below the upper reservoir, this is not correct. Kops II is conventional layout [1]. The power station is located near the down stream end. There is a high level pressure tunnel, sloping pressure shaft, surge chambers, underground power station and tailrace tunnel. This is usually the least cost option and would probably be the least cost option for Tantangara-Blowering (if such a scheme with 53 km of headrace tunnel could ever be viable).

Just for interest, I’ve compared Kops II [1], [2]] and Tantangar-Blowering:

Capital cost; $ m;$545;$12,000
Power; MW; 450; 9,000
Capital cost; $/W; $1.21; $1.33
Generation p.a.; MWh; 614,000;
Capacity factory;; 16%;

Pressure tunnel:
length m 5552 53,000
diameter m 4.9 3 x 12.7
Water weight t 104,696 20,141,622

Note: In changing from generating to pumping mode, Kop II has to stop and reverse the flow of 100,000 tonnes and Tantangara-Blowering 20 million tonnes of water.

References:

[1] http://www.kopswerk2.at/downloads/Folder_061006_englisch.pdf
[2] http://www.hydroworld.com/index/display/article-display/0159389604/articles/hrhrw/hydroindustrynews/pumpedstoragehydro/austria_s-450-mw_kopswerk.html

You are correct that the power station has to be sited underground. This was discussed up thread. I explained than that my original analysis was intended as a simple analysis making use of the Tumut 3 costs and engineering features which I had access to and could scale up. I did not have any way to do a rough calculation of the cost of the underground siting.

The cost of high pressure surface pipes is far too expensive. Also, surface pipes are more vulnerable to environmental threats, such as land slides. At higher elevations there is the threat of freezing in winter if there is a period when the water is not flowing in the pipes.

You suggest the shaft should be located at the upstream end rather than near the downstream end. The choices are:

1. Long, high level pressure tunnel to shaft above the power station and short low pressure tailrace tunnel. In this case the tunnel would begin a little belo Tantangara’s minimum supply level and decline at about 2% slope to the top of the shaft. A surge tower would be above the shaft.

2. Short high pressure tunnel, shaft to deep underground power station, and long tailrace tunnel.

3. sloping tunnel from Tantangara to the underground power station and then a short tailrace tunnel. The power station would be as close to Blowering as is practicable. It must at about 300m depth below surface so there is sufficient weight of rock above to contain the 925 m of static water pressure plus dynamic pressure.

Which option is selected would be based on engineering design and costs of the various options. (the surface pipe option would also be considered in early options analysis, but I suspect would be ruled out at the prefeasibility stage).

Here are some issues that would be considered in the options analyses:

1. Geology, rock conditions, tunnel stability, tunnel support and leakage along the various possible routes.

2. Location and cost of the surge chambers

3. Tunnel length

4. Length of access tunnels to the tunnels (for construction and removal of the excavated rock). The longer and steeper the access tunnels the higher the cost of the project.

5. Length and gradient of the access tunnels to the underground power station. (these tunnels have be used to get the huge turbines, generators, transformers and other large equipment into the power station and for operation and maintenance for 60 to 100 years.

6. Hydraulic head loss in the low pressure tailrace tunnels. The tailrace tunnel must slope down from the Minimum Supply Level of Blowering Reservoir. So the power station would have to be located deeper if it is located far from the downstream end. This means the shaft must be longer. Long tailrace tunnels have to be larger diameter or more of them.

I would expect the cost difference between the highest and lowest cost underground option would be no more than 20%. Surface option would be probably 5 to 10 times higher than the underground options.

The best option will not be decided until well into the design stage. However, the difference between the various underground options will be small compared with big issues that you, Neil, are concerned about. The cost difference between the various underground options doesn’t make the slightest difference to the fact that a large pumped hydro scheme like this is totally uneconomic now, even for storing energy from cheap baseload power stations. When we get to the point that nuclear generates at least 50% of our baseload, it is cheap electricity, and is available every night during low demand periods, then a project like this may start becoming attractive. It will only be attractive if the cost of electricity with pumped hydro is cheaper than from load following nuclear. Pumped hydro may become a viable option sometime after 2030 at the earliest, IMO.

One thing for sure, as I’ve made clear many times before, pumped hydro will not be viable as back up for intermittent renewable energy in Australia. Wind and pumped hydro will be some 30 times more expensive than nuclear (ref: http://bravenewclimate.com/2010/04/05/pumped-hydro-system-cost/#comment-86108 )

If you want to argue that wind power can be a fully dispatchable electricity supplier, like fossil fuels, hydro and nuclear, then I suggest you cost wind generators with energy storage at site. That should be your starting point. Then you can get a good estimate of the true cost of wind generation.

http://www.hydroworld.com/index/display/article-display/0159389604/articles/hrhrw/hydroindustrynews/pumpedstoragehydro/austria_s-450-mw_kopswerk.html