Editorial - VGB PowerTech Journal 4/2014
Renewables versus grid stability
The electrical energy supply system is a closed-loop control system – yet mainly backed by conventional thermal power plants. Control variable is the grid frequency, the actuating variables are the instantaneous power reserve (“spinning reserve”) as well as the power plants’ primary frequency control capacities. The closed-loop control process for the grid frequency is being provided by the inertia of the rotating masses of the turbo generators. The process control stabilisers are the primary controllers of the power plants themselves.
Wind power and photovoltaic conversion plants are no power plants. They are machines that harvest the latent energy provided by wind and sun and convert these into electrical power and feed it into the electrical grid – not on demand but on supply. In terms of system stability they are a disturbance variable like any consumer and do not provide any contribution to grid control. Apart from that, these expensive renewables – with their feed-in privilege – also reduce the full load hours of the cheaper conventional power plants, squeesing them out of the market. As such, not only the number of grid controllers but also the rotating masses supporting the grid frequency control are reduced.
This fact was also recognised by dena (the German Energy Agency/Deutsche Energie-Agentur) and published in its latest study “System services 2030” entitled “Security and reliability of a power supply system with a high share of renewable energies”. According to dena, the problem is to be solved by altering their legal specification, demanding also from these distributed energy systems contributions for frequency control. This applies in particular to large wind power plants. It is expected that their current grid start-up time constant of 10 seconds is reduced to 3.5 seconds. Alternative suppliers of frequency control services might be control power pools including bio gas plants, emergency diesels and large battery sets as well as demand-side management of energy-intensive industries. However, the future will reveal whether these suggestions will be sufficient and what they will cost.
What is going to happen if these measures are insufficient can be seen in Ireland: 10 GW of installed conventional power plant capacity are confronted with 2 GW provided by wind power. Summer low load can amount to 2.2 GW and high wind in-feed to 1.9 GW, i.e. at certain times, more than 50 % of power is already be generated by wind energy. Higher proportions of wind power are rejected by the grid operator as precautionary measure. In case of power plant grid disconnections such a generation portfolio can already lead to frequency gradients of more than –1 Hz/sec, which in turn leads to protective further shutdowns of remaining power plants. Especially components like couplings, shafts, turbine blades, seals and generator windings are being jeopardised. It is obvious that such frequency gradients in combination with power plant shutdowns cannot only result in largescale user power interruptions but can also bring the system to the edge of a black-out.
Now what is the situation in continental Europe, do we have to expect a similar situation as in Ireland? The Rostock University conducted jointly with VGB and member companies of VGB an investigation in order to identify which frequency gradients and deviations are to be expected in Germany and Europe by the year 2023 if new renewables are extended as in the past, reaching capacity shares of 70 % (Germany) and 50 % (Europe). It was assumed that frequency reserve and primary control will solely be provided by still available conventional power plants and that the grid start-up constant TN (representing the inertia of the rotating masses) will drop from ten seconds to two seconds in Germany and to five seconds in Europe. Besides, it was assumed that a large part of primary control capacities will be provided by Alpine high-pressure hydro power plants which, however, display insecure behaviour with this kind of control due to the so-called “output counteraction”.
Calculations have shown that in case of a 3 GW design disturbance for primary control in Germany a frequency gradient of -1 Hz/sec can be achieved and the maximum permissible dynamic frequency deviation of -800 mHz can be exceeded. Additionally the tendency of oscillation of primary frequency control is strongly increased for about one minute after the disturbance. Apart from the jeopardies that have already occurred in Ireland, this higher variation of frequency generally increases wear of the power plants employed in primary control.
Secure grid operation would already be impossible in Germany without our supporting neighbours who supply the missing grid services from their fossil and nuclear power plants via the interconnected grid automatically and free of charge, especially spinning reserve – thereby stabilising the system. It needs not to be discussed that any “Energiewende” at the expense of our neighbours does not make any sense at all.
In addition to the vague suggestions made by dena, we will have to keep a sufficient number of real power plants on the grid in order to provide the required grid services. However, in future it will not be possible to provide these services for free as it is currently.