What is the critical load in the power grid

Who disturbs the network frequency

On January 10, 2019, the network frequency in Europe fell to 49.8 Hertz. This near blackout has many causes - but it is by no means the only net-shaking event in recent weeks.

But let's start at the beginning: Wednesday, January 10, 2019, 9 p.m. Europa made herself comfortable on the couch, the living room was pleasantly lit. Well-deserved evening mood. A completely different feeling, however, seized the employees of the European power grid monitoring centers: The Grid frequency dropped to the critical value of 49.8 Hertz across Europe in less than two minutes and threatened to fall further.

It quickly became apparent that the downward trend could not be stopped with primary control power (PRL) alone, Reserve power plants could not be approached fast enough. The French transmission system operator RTE reacted decisively: An emergency load shedding order for all 22 disconnectable large electricity consumers in France took 1,500 MW off the grid, the amount corresponds to the electricity consumption of the metropolitan region of Lyon. The grid frequency could be stabilized and climbed back into the normal range by 9:10 p.m., at 9:25 p.m. 50 hertz was reached again. The European living rooms remained bright and warm.

Possible causes for the frequency drop on the production side

The pan-European electricity system is a complex network on which unmistakable factors act with varying degrees of intensity. However, an apparently harmless event, in conjunction with other apparently harmless events, can destabilize the system and cause small, medium, and large problems. The following applies: The larger the individual, security-relevant parts of the system, the less the system as a whole can cope with an individual failure.

While the ENTSO-E has not yet carried out a comprehensive analysis of the events of January 10, 2019, some causes for the drop in frequency can be deduced from the public data available to us. We would like to list these chronologically for a better traceability of the events:

20:00 CETAt 20:00 CET, according to ENTSOE-E in Spanish Litoral coal-fired power station Almeria has 140 MW of available power from Unit 1 from the grid. However, it is not certain whether this will have an impact on the specific frequency drop around an hour later.
20:26 CETIn French Penly Nuclear Power Plant near Dieppe, according to ENTSOE-E, Block 2 will be canceled at 20:26 CET. 850 MW of available capacity of 1330 MW of installed capacity are taken from the grid.
21:00 CETAccording to information from the industry information service montelnews.com, which in its report denotes TenneT-Press spokesman Mathias Fischer quoted, there was a "short but significant consumption peak" at around 9:00 pm due to the start of pumps in pumped storage power plants, including the 1.1 GW pumped storage power plant Goldisthal in Thuringia.

Another fundamental cause can be observed every day on the hour: This is where the so-called Hour change instead, which regularly leads to frequency fluctuations of 0.1 Hz in the morning and evening hours. That doesn't sound like much - but it is at least 50 percent of the tolerance range of 0.2 Hz provided by the AC frequency band.

These hour changes are caused by the "Shift change" of large power plants and have their cause in the partly hourly electricity trading: Some power plants try to get out of the delivery as quickly as possible in order to meet the delivery deadline and not "overproduce". The other power plants try to start delivery as late as possible - after all, the electricity generated is only paid for from then on.

Problems now arise from the Ramp times: If the turbines and generators do not yet deliver their full output “on the ramp”, there will be frequency drops. Controllable renewable energies such as bioenergy, aggregated in virtual power plants, can deliver performance much faster - this is allowed as a side note.

But let's take a look at the frequency curve measured at our company location in Cologne-Ehrenfeld. Here, too, the noticeable network imbalance across Europe could be traced: At time 1, the Litoral coal-fired power station in Spain failed, which was already causing unrest in the network frequency. Time 2 marks the failure of the Penly 2 nuclear power plant block near Dieppe. However, this gap can still be filled.

At time 3, however, at 9 p.m., the pumps of the Goldisthal pumped storage power plant switch on. In an unstable grid situation, the grid frequency drops rapidly - until the French RTE activates the 1,500 MW interruptible loads. In Germany, too, the TSOs activate emergency reserves, such as the 153 MW of the Herdecke pumped storage plant near Dortmund. But why were the Goldisthal pumps switched on at all in this delicate network situation?

A data error as the cause of the frequency drop?

A data error at the transmission system operator TenneT may be responsible for this, as reported by “Der Standard” from Vienna, citing the Austrian TSO Austrian Power Grid (APG). Hypothesis: The control system there misinterpreted the network situation, the network regulator supplied incorrect data. The subsequent decisions based on this incorrect data, largely made fully automatically, then pushed the grid frequency down.

Even if more details from TenneT about this incident are still missing due to the ongoing investigation, our experts believe that a network regulator could also have contributed to the problem. These fully automatic control devices continuously measure the network frequency at defined points, for example border coupling points. If this deviates upwards or downwards, the network regulator automatically sends control commands to electricity producers and consumers who provide control energy. If such a controller fails or if it supplies imprecise data, which is extremely rare thanks to up to fourfold redundancy architecture, the power grid frequency can deviate to a greater or lesser extent.

Nevertheless, the massive drop in frequency can only be understood as a causal chain from the most varied of causes, which we now summarize again:

  • Large power plant failures in France and Spain are destabilizing the European network.
  • A Data error in the network regulator of the transmission network operator TenneT lets the pumped storage power plant Goldisthal start up the pumps - even though the electricity network cannot supply the pump power.
  • The usual one Hour change when replacing large power plants, it regularly causes frequency fluctuations of 0.1 Hertz.
  • A Underfunding at the infeed, which may be due to the data error.

Why were the reserves of the electricity market insufficient?

The primary reserve power (PRL) is provided for in the electricity market design for fluctuations in the European high-voltage network. It intervenes in a matter of seconds if the network frequency, measured directly on the system providing the PRL, fluctuates upwards or downwards within a certain tolerance range (dead band). We also provide PRL in our virtual power plant and were therefore able to see very quickly that the entire existing potential had been exhausted.

But why wasn't it enough? After all, besides us, who are still a relatively small light in the pan-European PRL pool, many large-scale power plants also provide PRL. A predominantly theoretical analysis of the web portal netzfrequenzmessung.de suggests two possible reasons, quote:

“The error in the power balance of the network was greater than the primary control power activated at the time (over 2.6 GW). [...] The service for providing primary control power was not complied with by many power plants during this period. "

However, the portal does not provide a reason why the power plants did not fulfill their obligation to provide PRL. From this incident it can be deduced, however, that the provision of the primary control power required to stabilize the power grid might run counter to other interests of the power plant operators, or in other words: the provision of control power within seconds was not attractively priced enough.

Why is the power grid not coming to rest?

In the last few months, those involved in the electricity market have found it difficult to avoid the impression that the otherwise stable European electricity grid is in disarray. Three serious and many smaller events attracted a lot of attention.In addition to the already described power frequency drop on January 10, 2019, the following events also occurred:

Activation of interruptible loads according to AblaV in December 2018

On December 14, 2018 there was an imbalance in the German power grid due to a false prognosis for PV feed-in, which was countered by switching off industrial electricity consumers in accordance with AblaV. As we have documented and commented on in our blog post “The Green Scapegoat”, this event raised a lot of emotions, especially in heavy industry, and provoked a heated FAZ article.

Dangerous increase in grid frequency on January 24, 2019

On January 24, 2019, at 6 a.m., the grid frequency pendulum swings in the other direction: across Europe, the frequency meters show 50.2 Hertz, and locally the grid frequency even exceeds this upper limit for safe grid operation. Observers and the trade press speak of an extraordinary phenomenon - but have not yet provided a really reliable explanation. What is certain, however, is that the 50.2 Hertz outlier occurred again at the change of hour.

Renewable energies apparently played little or no part in this event, our trading experts suspect the cause in electricity trading. A surplus of electricity in the network was noticed the day before, and the hour change at 6 a.m. with the start of the morning shift in the large power plants was also involved. There was also a decrease in the use of negative secondary control power, which also brought more electricity into the grid and consequently drove up the frequency.

Attempted explanations and hypotheses of the cause

We also do not know exactly how the extraordinary network events of the last few weeks came about. Nevertheless, hypotheses can be made that lead to real problems not only theoretically, but in parts already very practically.

In terms of hardware, one thing is clear: The European power grid is a highly complex system in which very small and very large units are supposed to jointly ensure grid frequency stability. With all safety precautions and redundancies, it is still possible that individual units of this complex structure fail - and the effects of this failure depend on the size of the unit.

In other words: if a biogas power plant with one MW fails, the power grid does not fluctuate - with 1,300 MW from Penly Block 2, things look different. Decentralization in power generation creates security in purely system theory terms - in a system swarm, failed systems can be replaced at any time. The highly centralized French power grid shows almost every winter how the lack of flexibility and the cluster risk of predominantly nuclear power generation affect load fluctuations.

The way network reserves are dealt with has changed

The "smoking gun" for the numerous network disruptions has not yet emerged - but one fact catches the eye: The last serious pan-European network disruption occurred before October 2018 on November 4, 2006 - eight years ago, however, there were wrong reactions to a planned shutdown a 380 kV extra-high voltage line in East Frisia is the cause, and therefore a physically identifiable reason.

The current malfunctions, on the other hand, are far more difficult to assess and can only be incompletely explained with individual events such as power plant failures or network regulator malfunctions. There must therefore also be causes for the fluctuations outside of the physical power grid - one possible cause hypothesis is the introduction of the mixed price method for the balancing energy market. Consequences such as the rapid depletion of the electricity market reserves and declining balancing group loyalty destabilize the overall system.

But one thing is very clear: This time it was not renewables, even if the green scapegoat is again heavily criticized all over the country: Neither wind nor sun, nor biogas or hydropower were significant deviations from the forecasts.

How helpful was this article?
[Number of ratings: 22 Average: 4.8]

by Christian Sperling

Christian Sperling is the online editor at Next Kraftwerke and responsible for the content of our website. As an online editor and blog author, he is interested in all aspects of the energy transition; as a studied historian, he does detailed research and gets to the bottom of things.