1.3 terawatt hours – now Cern is also supposed to save energy

Not only private households and industry are affected by rising energy prices. Universities and research institutes also have to be at least heated in winter. And experimental research facilities sometimes have a large power requirement. For example, the famous Cern research center in Geneva, with its world’s largest particle accelerator, consumes around 1.3 terawatt hours of electrical energy every year. This corresponds to an average output of around 150 million watts (MW). The peak consumption is even 200 MW.

Scientists at the Karlsruhe Institute of Technology (KIT) are dealing with the question of how particle accelerators can be operated more efficiently and with less energy in the future. “Until now, accelerators have usually only been designed for maximum performance, but not for energy efficiency,” says Professor Anke-Susanne Müller, who heads the Institute for Accelerator Physics and Technology at KIT. She estimates the savings potential of accelerators at around 30 percent. In order to exploit this, many adjustment screws have to be turned and optimized.

The Karlsruhe researchers do not want to find out what can be done in detail by just thinking about it at their desk. Rather, new concepts for energy efficiency are to be tested and optimized in practice at the existing KARA accelerator facility. To this end, a new research facility called KITTEN was inaugurated in July 2022.

The acronym KITTEN stands for “KIT test field for energy efficiency and grid stability”. KITTEN essentially includes the accelerator and the new “Energy Lab 2.0”, as well as a photovoltaic system on the roof of the accelerator hall and access to a cold water layer in the ground.

“This is a research project that is unique in the world,” says Müller. “With the knowledge gained here, not only will particle accelerators be operated more efficiently and sustainably in the future, many other areas of experimental science will also benefit.” Accelerator required are also used in other research projects – for example cooling and vacuum technology, strong magnets, power supply lines or electrical transformers.

Accelerators are a collection of different technologies, each of which can be optimized in terms of energy efficiency. “You can see the accelerator as a model,” says Müller, “what we learn here can be transferred to many other systems. We think about that from the start. Universities will also benefit from this.”

At particle accelerators, a large part of the energy is required to operate the refrigeration technology, i.e. to generate low temperatures. Not only do many detectors and samples have to be brought to extremely low degrees Celsius, superconducting magnets also need liquid helium as a coolant.

If it is possible to operate the compressors in refrigeration technology more efficiently, a not inconsiderable amount of energy could be saved. “I can even imagine fundamentally different refrigeration cycles that also include natural resources,” says Müller. Well drilling to a depth of 38 meters each has been approved and executed for the KITTEN project. “There is a cold water layer relatively close to the surface that we can include in our heat and cold management,” explains Müller.

However, the most radical idea of ??the Karlsruhe researchers for saving energy in accelerators is the abandonment of the continuous operation of the facilities. “So far, almost all components have remained permanently switched on, even if some things are only needed temporarily,” says Müller. “Our goal is to switch off individual processes temporarily and save energy in this way. You can compare that to a really good stand-by function for electrical appliances in the home.”

This idea of ??savings should cause some accelerator physicists to be horrified. After all, these are highly complex and sensitive systems whose operating parameters must remain as stable as possible in order to guarantee the high quality of the measurement data. After switching components off and on again, it takes a long time before the original stability is regained.

“But we can risk it,” Müller is convinced. It relies on the use of artificial intelligence. “AI makes it possible to regulate the systems very quickly, so that they are ready for use again after a short time. Intelligently fluctuating rather than constant maximum energy consumption is the key to efficient accelerator facilities.”

So the KITTEN project is not just about improving hardware, but also very centrally about the development of intelligent software for AI-supported real-time optimization of the system.

An intriguing question is whether operation with fluctuating power consumption might not affect the measurement accuracy of the research facility and thus science. The KITTEN researchers want to determine how large these effects actually are in practice.

But it is already clear to Müller that there will probably be a reduction in the measurement accuracy from a certain point of saving energy. At this point, a whole new, rather unfamiliar question could arise for scientists: Does better measurement accuracy justify the associated higher energy costs?

When it comes to gaining new insights on the horizon of secured knowledge, there can be no compromises. There it is obviously not expedient to drive with the handbrake on and to prevent scientific knowledge that could actually be achieved with a system. Scientific progress has its price – also a price for energy.

You certainly won’t accept any loss of performance with the Cern. However, that does not mean that there is no interest in improving energy efficiency. As long as this is not at the expense of experimental possibilities, it is welcome. And so there is certainly cooperation between the researchers at KIT and CERN in matters of energy saving. The knowledge gained in Karlsruhe could be taken into account when building future, even larger accelerators.

However, some experiments are not about fundamental new insights at all, but rather about service measurements. In some accelerators, for example, the synchrotron X-rays generated by the particles whizzing in a circle are used to carry out material science investigations.

At this point one could certainly negotiate whether a slightly larger error bar with the cost savings that this could bring could be accepted. “That will lead to a change of perspective,” says Müller, “suddenly you’re not just a scientist, you’re also an energy consumer. This aspect will definitely gain in importance.”

Another aspect of KITTEN research is improving the resilience of research facilities. This means that the sensitive systems suffer little or no damage after an unplanned power failure. The researchers are also preparing for the possibility of a blackout, i.e. a complete power failure. Then greater self-sufficiency in the power supply can be helpful. After all, the solar cells on the roof of the Karlsruhe accelerator facility can provide an output of up to 240 kilowatts.

Even if one does not assume the worst case scenario of a large-scale collapse of the power supply, experts still expect that the stability of the grid frequency will deteriorate in the coming years and that there will be more frequent fluctuations in the frequency. “Accelerators live from a stable network,” says Müller, “fluctuations in voltage and frequency can have a negative impact on the measurements and research results.” The energy contained in the magnet system can also be used to stabilize the power supply.

The question of how quickly a research facility can be brought back into normal operation after a power failure is also of interest from another point of view. Finally, it is conceivable that individual research facilities might want to be shut down for a certain period of time – to save on energy costs or to protect the entire energy system from overloading in a crisis situation.

Experience has already been gained in this regard with the KARA accelerator. Because KARA is only used from Monday to Friday and is put into a “sleep state”, as Müller calls it, over the weekend. Switching off KARA completely has not been practical so far because it would take too much time to restart.

“KARA consumes at least 30 percent less electricity at the weekend than during the week,” says Müller, “so our research goal is that the entire system, even in normal operation, only needs as much energy as it currently does at the weekend.”

Differently deep sleep states are conceivable with accelerators. The KITTEN scientists hope their research will provide new insights into how low the energy consumption in “standby” can be reduced without making it impossible to start up again within an acceptable period of time. “We’ll still be able to learn a lot there,” says Müller.

Müller receives support from the President of KIT, Professor Holger Hanselka, who emphasizes the importance of accelerators as key research instruments: “Regardless of whether it is about future energy supply, climate protection or the mobility turnaround, material sciences or medicine: research needs high-performance infrastructures to meet these challenges. At the same time, we are faced with the task and assume responsibility for making the operation of these systems as energy-efficient as possible. With KITTEN we can further develop the accelerator technologies – which are fundamental for many disciplines – in terms of performance and sustainability”.