Tunneling for a Pumped Storage Power Plant

The Ritom pumped storage power plant in the Swiss canton of Ticino was originally put into operation in 1920 and acts as an electrical power buffer for both the Swiss Federal Railways and the Ticino electrical grid, with a capacity of 44 megwatts supplying power for both the Swiss railways at 16.7 Hz and the public power grid at 50 Hz. The plant operates by using surplus power to pump water more than 700 metres up to Lake Ritom, then supplying power on demand as the water descends and drives turbine-generators.

A project is presently underway, at a cost around CHF 250 million and scheduled for completion in mid-2023, to upgrade the Ritom plant to a capacity of 60 megawatts, with increase in water flow to obtain the increase in power. While the original plant used pipes on the surface, the upgrade was designed to route the water through a tunnel dug through the mountainside by a tunnel boring machine.

The terrain through which the tunnel was to be routed was very challenging, with fractured rock that allowed water flow, voids, and zones where cave-ins into the excavation were possible. Construction of the tunnel required a series of interventions to stabilise the rock as the boring process progressed.

Why do Swiss railways use 15 kV, 16.7 Hz AC power?


That seems like a lot of effort to add 16 MW.


My guess is that the main motivation for the project was replacing century-old infrastructure which had become difficult and expensive to maintain with modern equipment. The capacity of such an installation is fundamentally set by the head (altitude difference between reservoir and pumping/generation station) and the amount of water which is physically possible and permissible to add and remove from the reservoir. In the case of a natural lake in a mountain environment, these are fundamentally limited and not able to be expanded very much. I suspect the gain in power capacity was mostly due to greater efficiency in the new pumps and generators, with the increased diameter of the water channel allowing higher peak power than the previous pipes running along the slope.

One thing about Switzerland is that there is an understanding of the need to maintain infrastructure and keep it up to date before the costs and difficulty rises too high due to neglect.


If the 86% round trip efficiency of lithium battery storage systems that I found is correct, PSH is comparable or slightly less, and lasts way longer.

Something the US has either forgotten or ignores


It is truly mind-boggling/astounding that a state apparatus which acts in an omnipotent and omniscient manner in its efforts to control every aspect of behavior and thought of its law-abiding subjects (it does not concern itself with criminals any more), has neglected its most basic functions - like maintenance of infrastructure and fiscal responsibility. At least after all the bridges have collapsed and roadways crumbled, we will be able to burn worthless paper money for a modicum of warmth in winter. But they may just go ahead and ban winter instead; such is the state of the elite’s connection with reality.


There seems to be a belief among those who rule us that whatever they have now can never be taken away from them. Their focus is always on what new thing they can add. Sadly, it seems Our Betters have never heard of rust.


Fascinating , and truly amazing . Men of steel.


This film is a gem for folks like me who share a peculiar obsession with hard engineering solutions to problems with layers of hidden complexity.

Throughout the film I found myself repeatedly muttering “I can’t even…”

Is “gripping” too strong a word for a corporate documentary-style film?

The project is an amazing feat, but another peculiar obsession also compels me to ask an incorrect question: Is all of that cost and effort economically superior to the installation of e.g. a single GE 6F gas turbine powerplant?


Switzerland was a pioneer in pumped storage of electricity, with the first installation in 1907, and is today the world leader in deployment of the technology, with pumped storage accounting for (2017) 32.6% of total electrical generating capacity, far ahead of #2 Austria, with 18.7%. (In the U.S., by comparison, pumped storage is only 2.1% of total generating capacity.)

There are some excellent reasons for this. Electricity generation in Switzerland is dominated by hydroelectric power, accounting for 62% of all power, with nuclear power at 29%. Hydroelectric power is a base load source, able to provide a steady contribution to the grid except for unusual circumstances such as maintenance, mechanical failure, or drought. Electricity load, however, is variable, with greater demand in winter than summer and daily peaks and troughs.

There are two main ways to cope with a fixed supply and variable demand: peak generation plants and energy storage (traditionally, mostly pumped storage, although utility-scale battery technology is entering this market). For Switzerland, pumped storage has the great advantage of not requiring a fuel source, such as the natural gas used by most present-day peaking plants. Switzerland has essentially zero fossil fuel reserves, so all fuel for thermal peaking plants would have to be imported, with the risk of price shocks and supply disruptions (the latter remaining in the consciousness of a population whose children learn in school the consequences of having been entirely surrounded by hostile countries between 1940 and 1945). With a pumped storage plant, once you’ve financed the capital cost and relatively low maintenance, there is zero fuel cost and no worry about supply. Also, a natural gas peaking station requires infrastructure (pipeline, rail, or highway) to supply fuel, and this may be expensive in relatively isolated areas such as Switzerland south of the Alps.

Second, deploying a pumped storage plant requires terrain in which you have substantial elevation differences adjacent to the area in which electrical power is produced and consumed, and is particularly attractive when natural bodies of water exist at high altitude. Well, Switzerland doesn’t have any oil or gas, but what it does have in abundance is mountains, water, and mountain lakes, which means nature has provided most of the engineering prerequisites for pumped storage.

Deploying pumped storage in Kansas or Kazakhstan would be an entirely different matter.

There are thermal peaking stations in Switzerland—for example, a 43 megawatt plant in Cornaux, around 15 minutes’ drive from by house, able to run on either natural gas or heating oil. The plant was installed in 1973 and is able to come on line within 15 minutes. It has on-site oil storage of 27.5 million litres of heating oil, enough to run continuously for three months without supply from the adjacent oil refinery in Cressier, the only crude oil refinery in the country.


Electricity generation in Switzerland is dominated by hydroelectric power, accounting for 62%

This makes the choice of pumped hydro even more puzzling to me: we’re taking losses to pump water from one reservoir to another (only?) to make up for the fact that the overall system generation capacity is insufficient to meet peak demand? Here I sat, ignorantly thinking that pumped hydro made sense only given truly and uncontrollably variable energy sources such as solar and wind.

Pumped storage is an existence proof that hydro power can be throttled below some upper limit, which compels me to another question: why is pumping preferable to increasing the peak generation capacity of the non-pumped hydropower system? My pea brain wonders: why not add e.g. another 60 MW unit to an existing hydro plant rather than building and maintaining a completely separate pumped system with its attendant round-trip losses? I gather it’s not possible (or economically feasible?) to add generation capacity to existing hydropower plants? Or to build more non-pumped dams? It’s a curious and fascinating situation to have an excess of energy supply but insufficient supply density.


Assume you have a single hydroelectric dam that produces enough power (or a modest surplus) on a steady basis. You can easily reduce the power generation, but the maximum is set by the turbine and generator capacity, which in turn is scaled to the steady-state water flow and head of the reservoir.

Now, connect this power supply to a variable demand which, say, has 25% more power consumption during peak periods than trough periods. With the pumped storage plant, you can bank energy during the trough periods and release it during peaks as required. This requires only capacity to meet the peak demand above the base provided by the main hydro dam. Attempting to meet every need from the main dam would require scaling it up to peak requirement, which would mean bigger water channels, turbines, and generators, that, on the scale of a large hydroelectric facility, would cost much more than the pumped storage plant.

This is precisely why utilities almost universally have a mix of base load and peaking plants on their grids. A pumped storage plant is simply a peaking plant that replaces fuel by storing excess base load capacity at times of low demand.


But then the main use of the addition is as a peaker, not as to pump up the reservoir.

When operating as a pump, why not just reduce flow to the main turbine?

Let’s be clear. Went you said “reservoir” you are referring to the same one used for the pumped hydro rather than a different one that is under some constraint on reservoir level.


The pumped storage plant almost always has its own dedicated reservoir which feed only it, not a main grid generator dam. (I was about to say “always”, but there may be some exception of which I’m unaware.) For a pumped storage plant, you typically have a relatively small amount of water and a large head (in the case of Ritom, around 800 metres). For a base load hydro dam, there’s usually a much smaller head but far larger volume of water, essentially set by the average flow of the river (or other source of water, such as snow melt) that’s feeding the dam.


Is this the idea:

Lake Neuchatel has a hydroelectric plant. The water level needs to be kept fairly constant to please diners at the three star shoreline restaurants. The plant averages 70% power, often due to low demand and not low supply of water. Thus, often, it is at less than 100% power but to prevent flooding of the lake they bypass some water around the dam.

So they build Lake Fourmilab as a pumped hydro reservoir. Then, instead of bypassing the Lake Neuchatel dam, they run the Neuchatel plant above demand and then use the extra power to pump up Lake Fourmilab. In the rare case demand for the Neuchatel plant hits 100%, they bring lake Fourmilab generation online as a peaker.


That’s exactly the application. In many cases, the pumped storage plant is placed (where possible) close to the primary load so you don’t use grid capacity and experience transmission losses when supplying the peak power. In the case of Ritom, it is just a few km from the Gotthard Tunnel, which is a major load on the Swiss Federal Railway system, and feeds civil power into the grid for the canton of Ticino, where it is located.


fd1 makes a good point. It would suggest that building a pumped storage plant would make economic sense only when 100% of the area’s hydroelectric generation capacity had already been built out.

The one thing we know as an absolute without even putting pencil to paper is that pumped storage (where possible) beats batteries. And pumped storage does not depend on children working in African mines.


Not necessarily. When building a large dam and power station intended for base load generation, it may cost more to up-size everything in order to increase peak generation capacity which would only be used at certain times, limited by the permissible draw-down from the reservoir. Having pumped storage on the grid allows the big dam to run at a steady state, banking excess capacity when load was low. I also don’t know how long it takes for a big dam to adjust its generation power. One of the statistics often cited about peaking plants and pumped storage is how quickly (usually on the order of 15 minutes) they can start producing power when needed.


As an aside- and I haven’t been able to find it- I wonder about the cost of electricity for trains in Switzerland. Any idea of cost per mile? I know train type & size varies. How many CHF for the run of a standard length train (not double-length for rush hour) from Zurich to Bern, for example?


Here’s how they do it in San Diego, not without controversy:


This page from the “Facts and Figures—Infrastructure” Web site cites these numbers for 2021:

  • Electricity produced and procured: 3,063 GWh
  • Electricity used for railway operations: 2,275 GWh
  • Proportion of traction current from renewable sources: 90.2%

The Wikipedia page “Transport in Switzerland” says passenger train services account for around 274,000 km/day. I haven’t found comparable figures for freight, but the Infrastructure site says there are are more freight trains per day, but perhaps fewer kilometres per run. I’ll just guess and double the number to 548,000 km/day for passenger plus freight.

If we then take electricity per railway operations and convert to daily use, we get 6.2 GWh/day. Dividing by 548,000 km/day, we get 11.4 kWh/km. Using the mean 2022 business electricity price of CHF 0.162/kWh, we get an electricity cost of CHF 1.85/km to run the trains.

Since the Swiss Federal Railroads buys electricity in bulk and operates eight of its own hydroelectric plants, it probably pays less for electricity than the typical business rate.