A Solution to Intermittent Renewables Using Pumped Hydropower

Integrating large quantities of renewable generation with low-carbon technology will require the development of large flexible carbon-free generation and storage assets. Over the last 40 years, numerous large capacity pumped-storage hydropower plants (PSPs) have been built worldwide, many of which could be retrofitted to provide grid-balancing services that the grid of the future will need.

 le Cheylas pumped storage plant scheme. Credit: Alstom.

Overflow basin of LeChelyas Pumped Storage Hydropower Plant. Credit: Alstom.

Almost all PSPs are equipped with fixed rotating speed units. This technology imposes a fixed power capacity per unit in pumping mode: the pump is either stopped or operating at full capacity. The result of this is that other types of flexible generation plants must be kept online when the pumps are in operation. In a country like Austria where hydro generation represents a large fraction of the energy mix, keeping some PSP units in generation mode and some units in pumping mode provides a portion of the frequency regulation needed.  However, operating this way has a negative impact on the cycle efficiency, literally resulting in the intermittent generation heating up the water in the upper reservoir. The fixed operating speed also limits the ability of the hydraulic unit, when operating in turbine mode, to keep a high efficiency over a large head and flow range.

New variable speed technology for PSPs was recently introduced in Europe.  A key benefit of this new technology is that is gives PSPs the ability to regulate power both in pumping and generation modes instead of just in generation mode, like conventional units. Because this new technology allows for additional flexibility for PSPs, once equipped with them, these PSPs are now among the best solutions to provide necessary ancillary services for the grid.

While there are a few variable speed PSPs already commissioned or under construction in Europe (the 1000-MW Linthal plant; the 942-MW Nant de Drance and the 800-MW Venda Nova plant), the industry has no experience upgrading a conventional plant into variable speed.

Upgrades are much more complex than developing new plants because they require accommodating the existing embedded parts of a pump turbine and civil work. Variable speed motor generator volume is generally larger than conventional and necessitates the installation of large power electronic equipment close to the units in generally very constrained cavern areas.

The eStorage Project

In 2013 a consortium of energy stakeholders from across the EU joined together to create the eStorage project. The consortium consists of Elia, EDF, the Imperial College, Kema, and Alstom.

The aim of the eStorage project is to develop cost-effective solutions for the widespread deployment of GWh-scale variable speed pumped hydro storage plants (PSP) across the EU and to enhance grid management systems in order to advance the cost-effective integration of large amounts of renewable generation. The consortium set out to do the following:

  • Demonstrate the technical and economic feasibility of upgrading the existing Le Cheylas PSP to variable speed technology, enabling a new regulation capacity in pumping mode, and increasing its overall efficiency in both the turbine and pumping modes.
  • Enhance and validate the functionality of IT systems to demonstrate new balancing capabilities with energy storage by looking at closer to real-time market systems.
  • Quantify the benefits of an EU-wide rollout of variable speed PSP under alternative future development scenarios.
  • Propose changes to the market and regulatory frameworks in order to create a viable business model for energy storage and enable a wide deployment of flexible storage technologies in the EU.
  • Develop and examine technology solutions that would allow for the upgrade of 75 percent of European PSP (totally 30 GW of capacity) to variable speed to obtain additional capacity for flexible balancing in pumping mode.

This article is a discussion of the first goal of the eStorage project, the conversion of le Cheylas PSP to variable speed.

The Demonstration Site: Le Cheylas PSP

The PSP le Cheylas is located in the French Alps and has been operated by EDF since commissioning in 1979. The facilities generate electricity using the river Arc inflow conveyed from Longefan to Le Flumet reservoir by a 20 km long tunnel, with an annual average generation of 550 GWh and store it on a daily cycle, by pumping at night from le Cheylas reservoir into le Flumet.

There is only one powerhouse, which is equipped with two identical 270 MVA reversible units, each located in a shaft. A water conveyance system common to the 2 units connects the powerhouse to the 2 reservoirs. The eStorage project consists of converting unit 2 to variable speed.

Hydraulic Design

In order to take advantage of power regulation in pumping mode, the hydraulic design would need to allow increased water flow variation, which meant re-engineering the hydraulic design.

The key design target was to allow a power variation in pump mode of 80 MW under the full head range (i.e. constant power variation close to 30 percent of the unit’s nominal power, leading to a quite extended speed variation of +/- 7 percent). Both the pump and the turbine design represented an ambitious challenge, at the crossroad between variable speed technology and refurbishment (see table 1, below).

Table 1: Alstom variable speed hydraulic designs for on-going projects.

Table 1: Alstom variable speed hydraulic designs for on-going projects.

The pump design is the key part of a variable speed upgrade. When compared to a typical pump-turbine refurbishment two new hydraulic challenges constrain the design. The lower pump power limit is set by an instability area where it is not possible to operate (called the humped zone) while the high power limit is defined by cavitation phenomena.

Since the entire generator needed to be changed in this case, engineers could select whatever rotation speed they desired. The choice of the new reference speed — from a hydraulic point of view — is therefore a balance between cavitation and hump criteria. For design optimization, the decision was made to keep the existing reference speed as the appropriate solution to maximize the pump power variation.

Pit turbine 1 of LeChelyas Pumped Storage Hydropower Plant. Credit: Alstom.

Pit turbine 1 of LeChelyas Pumped Storage Hydropower Plant. Credit: Alstom.

For obvious economic and logistical reasons it was decided to modify only non embedded parts: the runner and the wicket gates. These components have higher water velocities and they bring the best efficiency / cost trade-off. The existing draft tube, stay ring and spiral casing were left unchanged.

An additional constraint was that the turbine must still match the existing transients performances. This was even more challenging since Unit 1 would not be upgraded.

Electrical Design

Installing a new double fed induction motor-generator (DFIM) inside an existing pit requires special attention to existing limitations such as structural design, auxiliary interfaces as well as the capacity of the crane to handle the increased rotor dimensions and weight. The integration of a new frequency converter has a direct impact on space requirements, the cooling system and the power supply and is therefore one of the key elements to be checked in the early stage of the project.

The DFIM uses the exchange between the wound rotor and the frequency converter to provide the speed variation, per figure 2. As a consequence, the stator needs to be oversized in sub-synchronous mode, due to the additional power transiting from the rotor to the stator, which could lead to power limitation due to the given civil works.

Keeping the existing stator could have been considered if the reactive power supply could have been reduced. The reactive power is partially provided by the frequency converter. However, in such a case, the stator winding needs to be compatible with the rotor winding. Due to a requirement for a power factor decrease from 0.91 to 0.85 in generating mode, the replacement of the existing stator was compulsory.

The 47.8-m gate operating rig at LeChelyas Pumped Storage Hydropower Plant. Credit: Alstom.

The 47.8-m gate operating rig at LeChelyas Pumped Storage Hydropower Plant. Credit: Alstom.

Another constraint on the DFIM design was to fit the stator and rotor within the motor-generator pit. The pit dimension is a factor limiting the DFIM’s maximum output, especially as the rating of the stator is bigger than its synchronous counterpart for the same power at the main transformer.

The DFIM wound rotor is also 23 percent heavier and has more volume than a salient pole synchronous rotor, which impacts the shaft line behavior. Due to the cylindrical rotor and to the bigger volume, the weight is greater, as the utilization factor is lower than for a synchronous machine.

Beyond the generator motor, the whole unit electrical equipment had to be reengineered: several parts of equipment could be reused while others had to be replaced. Of course new ones must fit in the powerhouse premises. For example, synchronous rotor excitation devices must be dismantled while stator MV gears could be reused.

Most of the new equipment that had to be installed in the powerhouse was for the DFIM rotor feeding. The equipment includes: Heavy duty power tapping on the MV side of the unit’s power transformer; Short circuit current limiting reactors; a MV breaker; a special VSI transformer; harmonic filters; a VSI; segregated phase bus ducts from the VSI to the rotor ring cubicle; a rotor over-current and over-voltage protection cubicle; and non-conventional current transformers and voltage transformers for rotor current and voltage measurement at very low frequency.

It has to be noted that the largest pieces of equipment required for rotor excitation (VSI transformer and VSI) represent roughly 150 m² of ground space, which might be difficult to find in some underground power stations. Hence, large pieces of equipment such as the tapping transformer or VSI module would need to be placed outside of the power station.

Figure 1: le Cheylas pumped storage plant scheme. Credit: Alstom.

Figure 1: le Cheylas pumped storage plant scheme. Credit: Alstom.

On the stator side, more traditional pieces of equipment need to be placed including: isolated phase bus ducts (part of which may be reused from existing synchronous unit); starting/braking short circuit breaker used for the DFIM launching in motor mode and for the re-generative braking sequence; and — depending on their condition, age and rating — a generator circuit breaker and phase reversal disconnectors. Last but not least, the unit power transformer would need to be replaced in order to accommodate a higher MVA rating. This replacement provides an opportunity to increase stator voltage and optimize DFIM design.

Figure 2: The rotor/converter exchange in the pumped storage plant. Credit: Alstom.


Figure 2: The rotor/converter exchange in the pumped storage plant. Credit: Alstom.

Some important features of the synchronous unit — such as black start operation, isolated network feeding or line charging capacity — must also be undertaken by the variable speed unit. Black start operation without tapping energy for rotor excitation is obtained thanks to a low-power feeder that energizes the VSI enough to create voltage on the stator side and from there build up stator voltage.

Figure 3: Planned unit layout for the pumped storage plant upgrade. Credit: Alstom.

Figure 3: Planned unit layout for the pumped storage plant upgrade. Credit: Alstom.

Mechanical design modifications were also necessary especially where the machine shaft line and bearing were impacted. The most critical feature for the shaft line was the natural bending frequency, which was overcome by the re-arrangement of the unit layout.  The thrust bearing needed to be redesigned with a higher capacity to support the supplement of the unit axial load that comes mainly from the DFIM rotor. In addition, the new rotor overweight required changes to the powerhouse lifting equipment.

Economic Analysis

In order to perform an economic analysis the consortium modeled the operation of le Cheylas PSP considering the possibility (or not) of offering primary and secondary control reserve using operational data from 2010 and 2012. Le Cheylas PSP, in its present configuration, is already able to provide flexibility in turbine mode, and the upgrade of unit 2 won’t modify this characteristic. In pumping mode, the upgrade would allow it to offer +/- 40 MW of frequency control.

The main results concerning frequency control supply are the following:

  • Le Cheylas supplies much more frequency control: sum of primary and secondary reserves represents an increase of +57 percent. Primary reserves produced an average of 41 GWh/year and secondary 62GWh /year.
  • Logically, this increase comes from unit 2 pumping periods.
  • But it is also coming from the new generation periods: the more energy is pumped, the more energy is then generated. Frequency control is supplied when generating this additional energy.
  • Unit 2 is almost always supplying frequency control while pumping, and unit 2 is much more used than Unit 1 for pumping.

The current price that RTE uses in France to compensate the electricity producer delivering capacity for frequency control is around 17 €/MWh That price is contractually defined for 3 year-periods. RTE organizes the secondary frequency control through mandatory bilateral contracts with producers that must reserve an asked volume for every half an hour.

Considering this fixed price, the annual value of additional flexibility turned out to be around €2 million per year at today’s prices and it is expected the price paid for frequency reserves will increase in the future to reflect the greater need for ancillary services.


Now that the technical issues have been identified, assessed and solutions have been put forward demonstrating the feasibility of the upgrade, the next step is the construction phase. Not only will the corresponding unit benefit from an improved hydraulic design and achieve greater dynamic and energetic performance, but the unit will also undergo a complete rebuild extending its operation lifetime. 

Finally, its important to note that this type of upgrade is about 10 times less expensive than the cost of building of new power plant and can be implemented in much shorter time (about 3 times faster than a new project). Moreover, by developing storage capacity and flexibility in pumping mode (corresponding to periods with the lowest availability of flexible capacities), this conversion can be considered as part of the solution for the integration of an important share of intermittent generation in the European power mix.

Nathalie Lefebvre works at EDF in France, Olivier Teller is a Product Director with Alstom Hydro, and Marie Tabarin, who has left the company, was a product manager with Alstom Hydro when this paper was written.

Previous articleBrazil-based Renova’s Partnership with SunEdison Shows Shift in Strategy
Next articleCalifornia Environmental Permitting Still Favors Natural Gas

No posts to display