June 2011, F. Lempérière
Potential, Costs, Impacts
1. Energy generation
1.1. Present hydropower
The installed worldwide capacity is close to 1000 GW, supplying up to 3.500 TWh/year. A part of the energy may be stored seasonally (20%?) and a large part daily. This production may be adapted to needs or associated with intermittent renewable energies. In many countries the full capacity generation may be permanent during the rainy season (when solar energy is reduced) and mainly used for peak hours during the long dry season.
Dams may be used over one century and mechanical equipment may be updated after 50 years. The production and flexibility of operation may be reduced by siltation of reservoirs. However over 80% of the production by existing plants is likely to be kept in 2100.
The total storage of dams is 7.000 billion m3 of which 6.000 totally or essentially for Hydropower. 3 or 4.000 are dead storage, partly filled by silt. The area of hydropower reservoirs is over 300.000 km², i.e. 100 km² per TWh/year or 300 km² per GW. But these ratios have been progressively reduced from 1950 to 2000 and the likely ratios for the future may be 20 or 30% of the figures above because the largest reservoir sites have been used and for reducing the problems of environment and resettlement. The seasonal energy storage will thus be lower for future hydropower but the daily storage will remain very important.
The cost varies considerably and is much reduced after 30 or 40 years because investment is paid for and O and M costs are low. The average investment is in the range of 1000 US$/KW but the yearly operation may vary between 1500 and 8.000 hours. The total world expenses are presently about 100 billion $ for supplying over 3.000 TWh, i.e. 3 cents/KWh.
Small hydro (plants under some dozens of MW), generates about 10% of the hydropower and suggestions from antidams ONG to use only small hydro would thus mean quite the end of hydropower investments. Most power is from rather huge schemes: 20 schemes supply together 500 TWh/year and Inga (Congo) could supply 300 TWh/year for a cost under 2 cents/KWh.
1.2. Future onshore hydropower
The total technical potential is 15.000 TWh/year, the economical potential about 10.000. The generation may well raise from 3.500 TWh/now to 7.000 mid century (with 2000 GW) and up to 9.000 end of the century. 5 countries (China, Russia, Brazil, Canada, Congo) have 40% of the world potential (and of the world area).
The cost per KWh may be in the range of 2 or 3 cents of $ for some future large schemes and justifies a generation very far from utilization with 1 or 2 cents for transport. For many schemes the cost will be in the range of 5 cents during first decades. A cost in the range of 10 cents for such high quality power may be acceptable mid century.
Presently industrialized countries are supplying 40% of existing hydropower but will probably generate only 5% of the additional capacity.
Half of existing dams were designed when ecology was quite overlooked: not surprisingly many old dams may thus been critized but most dams built recently take due care of environment and reduce resettlement. The total lakes area for future hydropower dans will be under 200.000 km². In most developing countries, criticizisms against dams do not prevent their implementation but may often cause delays especially where foreign financing is necessary (Africa).
1.3. Tidal energy
The theoritical potential is high, in the range of 20.000 TWh/year but a small part only will used, probably in the range of 1000 TWh/year, because a reduced part only of the tidal potential available some where may be used, the cost is too high where the average tidal range is under 4 or 5 m and a large part of the potential is in very remote places.
There are favourable sites in 10 or 15 countries, the most favoured being Canada, Argentina, Russia, China and France. Two solutions are possible: tidal plants operating between large basins and the sea or turbines in tidal streams.
The conditions of foundation of tidal plants and relevant dykes are usually rather favourable (sand, gravel or rock at 10 to 30 m depth) the cost of dykes exposed to high waves is however high and the length of dykes should be low as compared with the supplied energy which is proportional to H² S, H being the average tidal range and S the basin area. This favours narrow bays or estuaries (as Fundy in Canada, Severn in U.K. or Kutch in India) or large sites along shore such as in France, Russia or China.
The technical feasibility is not questioned but the optimization of the solutions for dykes and turbines is needed before implementing the largest schemes.
The environmental impacts seem more favourable than for onshore hydropower; the area of basins (where conditions will be rather improved) will be about 30 km² per TWh/year, and the production is well known in advance. However a high supply week is followed by a low supply week and the supply varies along each tide, a storage is thus necessary and may be associated with storage for wind and solar energy.
The cost per KWh will probably be in the range of 10 cents/KWh and probably less for largest schemes, i.e. very acceptable but the implementation may be slow for two reasons:
– Many countries with tidal potential have also a low cost onshore hydropower potential which may be developed first (China, India, Canada, Argentina, …)
– The lowest costs are linked with huge investments for sites of 5 to 15 GW and smaller sites should be first developed with higher costs par KWh.
It is thus likely that the total plants capacity will not be higher than 200 TWh/year in 2030, and 500 in 2050. It is thus limited worldwide but could be, for instance in France, 15% of the electric supply in 2050.
– Tidal current energy may be developed in small or medium schemes with power units in the range of few MW, i.e. much lower than for tidal plants. It is thus possible to get subsidies for developing various turbines designs. Some designs appear clever but the cost per KWh, presently well above 20 cents, may not be competitive even within 20 or 30 years and the realistic potential appears much lower than for tidal plants, 100 TWh/year in 2050.
1.4. Waves energy
The theoretical potential is very important and many studies and tests by small schemes have obtained much subsidies; various tentatives failed and it seems difficult to get a significant potential under 20 cents/KWh because the possible large waves require costly protection or structure.
Breakwater for harbours or tidal plants may be used as basis for waves generation but with a limited overall potential.
The realistic potential at an acceptable cost, mainly in countries with quite permanent waves (U.K., Namibia …) is probably limited to few hundred TWh/year in 2050.
The possibility of a more cost effective device should not be excluded but it should be very optimistic to rely on waves for a significant share of the world energy mid century and no country seems include it in their energy schedule.
1.5. Overall hydropower generation
A generation of 6 to 8.000 TWh/year in 2050 appears likely with possibly up to 10.000 end of century. It will be essentially from large hydropower onshore schemes. Two qualities of such power should be underlined:
– Its flexibility for adjustment to needs, to electric frequency control, to the grid security, and a possible association with intermittent energies thanks to daily storage.
– Its low long term cost. The present relevant total world expenses (including investment) are yearly about 3 cents per KWh and will remain about the same along the century, thanks to the very long life of dams and plants; it is much lower than any other energy generation.
2. Energy storage (Emerald lakes)
The name of Emerald lakes for P.S.P. (chosen for their colour, and the fact of being precious and favouring green energies) may be used because their purpose and design may be different from traditional hydropower lakes.
The most promising storage solution seems the Pump storage. The 400 Pump Storage Plants (P.S.P) existing or under construction worldwide total 150 GW, i.e. over 5% of the average electricity supply close to 2.500 GW. They are mainly used for peaking capacity but some recent ones (mainly in Europe) are backing the wind energy. They are usually operating between two dams reservoirs linked by tunnels. The differential head is between 80 and 1000 m, most often between 200 and 500 m.
The average site capacity is 3 or 400 MW but some plants are close to 2.000 MW. Each site has usually several power units.
The storage is usually 5 to 20 hours at full capacity. The average investment appears in the range of 1500 US$/KW of which about half for dams and tunnels. Japan has already a much higher storage capacity with 40 P.S.P. totalling 30 GW i.e. 25% of an average electricity supply of 120 GW. They are in a mountains area of 300.000 km² (1 GW for 10.000 km²).
It is foreseen that the world energy utilization will about double between 2010 and 2050 and that the percentage of electric energy will at least double, i.e. an electric supply over 80.000 TWh /year (close to 10.000 GW as average) instead of 20.000 TWh/year now. With the same conditions as now, the need of P.S.P. would thus increase from 150 GW to over 500 GW but the huge utilization of solar and wind will require much more storage. It is thus useful to evaluate the range of needs and the corresponding Pumping Storage Potential, cost and impacts.
Evaluations are difficult and should be made separately for each world part. A very rough global evaluation is however presented below.
In 2050 the electricity will probably be under 20.000 TWh/year from hydraulics, Nuclear and biomass and should be reduced to 10.000 TWh/year from fossil fuels; wind and solar should thus supply over 50.000 TWh/year, for instance with equal shares for wind, PV and CSP. The supply from intermittent sources will then be 6.000 GW as average but may be very low part time, especially at night with a huge gap at the peak of needs at 6-10 p.m. The theoretical need of storage should thus be close to 6.000 GW but in fact may be reduced for several reasons:
– CSP may store energy at a cost which may be less than the cost of P.S.P. storage.
– Some flexibility is given by a remaining share of fossil fuels and by lakes hydropower.
– Price incentives and smart grids may reduce along few hours the needs for adjustments.
And the utilization of P.S.P. may vary with their cost: for instance the cost of PV + P.S.P. may be locally lower than the cost of CSP with storage and PV is easier to implement.
A capacity in 2050 between 2.000 and 4.000 GW appears thus a reasonable evaluation, i.e. 20 to 40% of the average total electricity supply then of 10.000 GW.
The need will be much lower in 2030 when solar and wind will supply 10.000 TWh/year associated with the flexibility of much fossil energy. It will probably be under 1000 GW.
The main investments for P.S.P. (as for solar energy) will then probably be between 2030 and 2060 at a rate close to 100 GW per year with a total capacity up to 5.000 GW reached in the second part of the century.
The relevant technology is well known and the industrial capacity may be easily reached; the problem is : are they possible sites for storing several thousands GW along about 30 hours at acceptable costs and impacts ?
The traditional solution between 2 dams may be extended as shown by the Japan example. However the advisable duration of storage will be probably closer to 20 or 40 hours than to 5 to 20 as presently and the capacity of traditional solutions probably under half of needs. And many world areas have no high mountains. It is thus likely that a large part of future P.S.P. will be with differential heads between basins under 100 m. One or both basins may be out of rivers and possibly at sea.
The low basin may be an hydropower or natural lake (they total million km²) or the sea. It is also possible to use two artificial onshore basins even in a same rather flat area and a simplified example is analysed below: a likely average plant capacity between 1 to 5 GW such as 2 GW with 30 hours storage.
For an area where the level is as average 10 m and at maximum 30 m over the level of the lowest place (referred to as level 0) it will be often possible to build a high basin operated between the levels 75 and 100 and a low basin operated between levels 5 and 15. The average head will be 77,5 and the necessary high basin area.
i.e a circular basin of 4 km diameter and a 12,5 km long dyke. The low basin, partly excavated for materials of the high basin dyke will have an area of 12,5 x 25/10 close to 30 km².
The main cost is the dyke of the high basin, about 90 m high requiring 15.000 m3 of materials per m, i.e. a total close to 200 millions m3 x 10 $/ m3 = 2 billion $ i.e. 1000 $/KW.
For such quantities the overall cost of 10 $/m3 is probably high and the cost of the low basin dyke is quite marginal.
The necessary worldwide earthmoving quantities will be very important, but well under those of open air mining quantities. Cost efficient equipment of open air mining may be used.
The cost of the plant itself (civil engineering and equipment) will be under 1000 $/KW and the total cost in the range of 2.000 $/KW.
This cost will be lower for more favourable topography or if using the sea as low basin especially if the high basin may be on a cliff.
Anyway the total investment for 3.000 GW will may be in the range of 5.000 billion $ for a yearly intermittent supply of 50.000 TWh, i.e. an investment of 10 cents per yearly KWh of intermittent energy and a cost per KWh close to 1 cent.
Should be added the loss of power, i.e. 20 or 25% of the stored energy which may be 30 to 50% of the intermittent energy, i.e. about 8 or 10% of a direct cost close to 5 cents in 2050, i.e. 0,5 cent.
It seems thus possible to get the necessary storage for intermittent energies for an average extra cost under 2 cents per KWh.
The environmental impact of P.S.P. is mainly linked with the area of basins. The example here above requires 20 km² par GW but the necessary area will be usually lower, especially if using seawater. The necessary area for 3.000 GW may thus be in the range of 30.000 km², most out of rivers, with a rather low impact on environment.
Summary and conclusions
– Hydropower generation seems worldwide the most attractive renewable energy by its cost of about 3 cents per KWh and its flexibility. But it will be limited under 10.000 TWh/year with a capacity increased to 3.000 GW and a lakes area increased to 500.000 km².
– Pump storage plants may be the best solution for the necessary storage associated with wind and power which will be used for a large part of the world energy needs. The corresponding extra cost for storage will be in the range of 2 cents per KWh of intermittent energy. The necessary P.S.P. capacity in 2050 may be 3.000 GW occupying 30.000 km², most out of rivers. As for generation, the P.S.P. may be used at least one century.
– Directly or indirectly, Hydropower will be an essential part of the future world energy.