Integrating intermittent and random energies, i.e. variable renewable energy sources (vRES) from wind or photovoltaic origin in an electrical network clocked strictly at 50 Hz, is not simple at all. They depend on meteorological risks: wind velocity, clouds, fog, etc. and if they are certainly more and more predictable thanks to the progress of weather forecasts, they have to be endured in all cases. In contrast, conventional energy sources, fuel, gas or coal-fired power stations, nuclear power plants, hydroelectric power stations (excluding “run-of-the-river” units) as well as units supplying electricity from biomass are said to be controllable, that is to say that their level of production is controllable, plannable and adjustable, depending on the electricity demand.
Technical studies, in particular one by EDF R&D, conclude that an admissible proportion of wind and photovoltaic electricity is around 40% of annual average power production (varying from 25% at low network load to 70% at very high load). Exceeding this proportion increases the risk of destabilizing the network, which could even lead to black-out (collapse of the network).
In addition, this is only possible at the cost of making more complex networks (smart grids) and investments in controllable resources to be kept as backup/emergency solutions.
Going above this proportion in the energy mix will only be possible when the problem of massive industrial storage and the return of electricity on demand have been solved.
An article written in collaboration with Georges Sapy.
A little history
The beginnings of electrification were characterised by a confrontation between two historical figures: Nikola Tesla, the alternating current supporter with Georges Westinghouse and Thomas Edison, who championed direct current. As we know, alternative current won.
The three-phase alternator (1)(2), invented by Tesla, makes it possible to produce high powers under high voltage. It becomes possible to transport current over long distances.
Indeed, at equal power P, increasing the voltage U, makes it possible to decrease accordingly the intensity I, and thus to minimize the losses by Joule effect, expressed by RI², where R is the resistance of the line.
Transformers make it easy to raise the voltage for long distance transport and then lower it to deliver power to the low voltage network and the end customer.
Public networks are a way of connecting producers to consumers. They make it possible to pool and optimize the means of production thanks to the proliferation of consumption resulting from the fact that at all times, not all consumers use all the maximum power they have subscribed. It is estimated that, at the scale of a country, a network can divide by approximately five the necessary means of production.
There are national transmission networks, operating up to 400,000 V three-phase (they are operated by RTE in France, Terna in Italy, Tennet in the Netherlands, etc.), interconnected in Europe within the ENTSO-E European organisation and distribution networks, from 20,000 V three-phase to 230 V single-phase to the end customer (mainly operated by Enedis in France).
At any moment, the generation of electricity injected into a network must be equal to the consumption of users. The transmission network operators manage the balance, either by letting the primary and secondary control of the production plants coupled to the a network acting automatically, or by implementing additional scaled means (tertiary or load follow control), or by activating on imports, or even, in extreme cases, causing deferred consumption or controlled load shedding.
The frequency of the current is indicative of the balance of the network. It is monitored in order to maintain its nominal value of 50 Hz. It can for instance be seen online at the Swissgrid website.
During the “diplomatic” incident between Kosovo-Serbia and ENTSO-E in March 2018, with these countries failing to meet their supplying obligations, the frequency of the European interconnected network fell to 49.996 for a long time, causing a delay in clocks based on synchronous electric motors to make them work, resulting in a commotion in the population relayed by the media.
In Europe, electrical networks are clocked at 50 Hz. This frequency is determined by the rotating speed of electricity generators (alternators). The stability also greatly depends on the mechanical inertia of all the rotating generators coupled to the network and working synchronously together.
The insertion of non-synchronous, wind and photovoltaic power generation machines without natural inherent mechanical inertia decreases the stability of the network, even if the addition of so-called synthetic inertia systems makes it possible, in the case of wind turbines, to recover the inertia of their rotor.
Various studies have focussed on evaluating the impact of photovoltaic and wind energy on networks, such as:
- The ADEME study, due in 2050, in France, foreseeing 100% of renewable energy sources, based on a meteorological history of 6 to 7 years.
- The Agora Energiewende study by German Fraunhofer-IWES taking into account 7 countries (France, Germany, Benelux, Switzerland and Austria), due in 2030, with a meteorological base reduced to 2011.
- The “Technical and economic analysis of the European electricity system with 60% RES” study by A. Burtin and Vera Sylva of EDF R&D, 2015. This takes into account all 34 European countries, federated within the ENTSO-E, meteorological base over 31 years, due in 2030, taking into account 60% of renewable energies sources (RES), of which 20% controllable, hydraulic and biomass energies, and 40% intermittent, from wind and photovoltaic energies (vRES).
In the latter, it appears that this RES sharing, although less stable than the current situation, remains sustainable at the cost of a high systemic additional cost due to the necessary implementation of sufficient controllable substitution generators, increased use of import/export, participation, albeit limited, of the intermittent means for frequency adjustment limited contribution, etc.
On the other hand, moving to a higher vRES injection rate will only be possible at the cost of developing storage techniques and restitution by destocking electricity that has become controllable trough production means using these stocks.
Storing electricity and on demand restoring facilities able to mitigate the lack of wind and photovoltaic energy at an acceptable cost, is the condition for the increase of vRES in the mix. This issue is the subject of intense research. The following four major focuses are being considered:
1- Pumping Energy Transfer Station (PETS) is the only commercially and economically viable operational large storage facility currently in operation, most of the time. It raises the water level in upper reservoirs by pumping during off-peak periods and it runs turbines in case of high demand for electricity. The storage capacity by PETS is high, around several GWh, with power capacity generally comprised between several hundred MW and more than one thousand MW. Its energy efficiency is also high, from 70 to 80%, but their number in France is insufficient and limited by the lack of additional geographical possibilities. The total storage capacity is only about one twentieth of one day consumption in cold winter conditions.
2- Compressed Air Energy Storage (CAES) consists of compressing air during periods of low demand, storing it and then restoring energy by relaxing it. This technique is not yet mature, only two units being in service in the world, and it is in competition with PETS and batteries.
3- Electrochemical batteries. They differ in the type of electrolyte and electrodes: lead, alkaline, sodium, nickel-cadmium, lithium… but all have high energy efficiency. However, they do not allow very large-scale and long-term storage. It has been calculated that all the batteries available in California, if solely dedicated to storage – which would immobilize vehicles – only represent 23 minutes of state consumption.
Today lithium-ion batteries can store 0.2 to 0.3 kWh per kilo, ten times more than lead batteries and allow 3,000 to 4,000 cycles. They have made enormous progress since the 1980s, due, among other things, to the development of the graphite anode by Rachid Yazami at CNRS in Grenoble in 1982.
Although they are mainly used for mobility, telephones, computers and electric cars, high-power industrial applications (several megawatts or tens of megawatts) are being developed, particularly for power networks.
EDF Renewable Energy, the US branch of EDF, commissioned 20 MW Li-Ion batteries for frequency adjustment of the PJM (North East USA) power grid. A new 49 MW project is under way in the UK.
4- “Power to Gas to Power” is the only solution available on the scale of massive and inter-seasonal needs, but it is still currently only theoretical. Two routes are possible:
The hydrogen route, by electrolysis: Electricity -> Hydrogen -> Electricity, global current efficiency being about 30%
- The methanation route: Electricity -> Hydrogen -> Methane -> Electricity, global current efficiency being about 20%
The investments needed for these two solutions are very high while the load factors are low. Profitability calculations (assumptions and calculations in appendix 3) lead, in order of magnitude, to €300/MWh for the hydrogen route, €500/MWh for the methanation route. These costs being far too high compared to electricity market prices!
v1 = V sin (ωt + φ)
v2 = V sin (ωt + φ + 2/3π)
v3 = V sin (ωt + φ + 4/3π)
In this case the system is said to be balanced and it is easy to check that: v1 + v2 + v3 = 0
In the case of single-phase current, the power is expressed as follows: P = vi = V sin (ωt) x I sin (ωt + φ) = VI cos φ – VI cos (2ωt + φ)
The second term of this expression is translated physically by a power absorbed by the transmission shafts of the rotating machines, in the form of torque oscillations which end up causing a fatigue fracture of the metal.
In the case of three-phase current, the term in cos (2ωt) disappears. It is this absence of an oscillating component of the torque that allows the three-phase alternators to reach extremely high powers without risking fatigue fractures, a considerable advantage.
The profitability calculation hypotheses are as follows:
- €70/MWh as the cost of the energy produced by dedicated vRES, nothing if we use the vRES surplus
- Investments estimated at €4.4 M/output MW for the hydrogen route, €7.7 M/output MW for methanation route (investments are counted by MW at the output of the routes)
- Load factors under 3,000 hours in case of dedicated vRES, under 900 hours for the surplus
- Amortization over 25 years based on a financing interest rate of 3%