Standard content for Members only
To continue reading this article, please login to your Utility Week account, Start 14 day trial or Become a member.
If your organisation already has a corporate membership and you haven’t activated it simply follow the register link below. Check here.
Utility Week speaks to Swedish entrepreneur Niclas Adler about his plans to combine high-performance computing with district heating through his latest venture, Synthetic Analytics.
Niclas Adler started his career as an academic but has also become a prolific entrepreneur, founding a long list of technology companies. Much of his work in recent years has focused on utilising high-performance computing, particularly within healthcare.
One of the things he came to see is that the current growth model for high-performance computing is “totally unsustainable.” Requirements were already increasing at a significant pace, but the coronavirus pandemic has accelerated this growth, as will the rollout of 5G telecoms networks.
“Today more than 2 per cent of all energy is used for high-performance computing and that’s bigger than the UK and more than France,” he remarks. By 2030, Adler says this figure could rise to as much as 10 per cent of global energy demand.
He says the energy characteristics of high-performance computing are different to those of data storage: “These are all super active processors that are run 24/7 so they generate an enormous amount of heat and of course need to be powered by an enormous amount of energy.”
Conventional air-cooling itself consumes a huge amount of energy: “That means for all of the energy you spend on processing you have to spend 30 per cent extra energy to cool down the processors to the right operating temperatures.”
The obvious wastefulness of this situation led Adler to create Synthesis Analytics. Founded in 2017, the company is seeking to turn one sectors’ trash into another’s treasure.
At its research and production facility located within a converted power station near the city of Norrkoping, the company has designed, built and tested a prototype computing system that can feed otherwise wasted energy from the process to district heating networks.
Instead of being air-cooled, the system uses immersive cooling, whereby the processing units are submerged within sealed tanks in a liquid that is a thermally conductive but electrically insulating. The tanks are installed on racks in a standard 40-foot shipping container, alongside an exchanger that is used to extract the waste heat from the cooling liquid.
Adler claims only 2 per cent of its overall energy consumption is required for the cooling. The vast majority – more than 95 per cent – can be recaptured and used to heat buildings.
He says the containerised, modular design will allow the systems to be installed quickly and scaled up to meet the needs of the local population. The company is initially focussing on its home market of Sweden, where around two thirds of households are already connected to district heating networks.
Working with Eon, the firm will shortly deploy their first commercial unit in Norrkoping and has projects in three further cities in the pipeline. The company recently launched the first phase of a £300 million fundraising programme, seeking £40 million to help finance these initial projects.
Adler says they will probably look to expand next into neighbouring Norway but also have their eye on the UK’s steadily growing market, noting the opportunity to rollout out smart city infrastructure alongside district heating.
Mirroring the trend in the energy sector, Adler says high-performance computing is being increasingly located next to demand in an effort to reduce latency: “Instead of transporting information and data to the other side of the planet and bring it back, you want to have your high-performance computing placed as close to the use as possible.”
“All of this traffic that you create with both energy and data is unnecessary,” he adds. “You get energy loss problems in the energy sector and in the computing world you get latency – unnecessary waiting time that creates a lot of problems for different services.
“5G networks are a perfect example. If you move the computing power away from the users you really get very little of the benefit”.
Please login or Register to leave a comment.