All the World is Green

These days, all the world seems to be #green. We’re exploring for #green #minerals in order to enable the #green #transition so that we can move away from fossil fuels to #green #energy. Of course, the greenest and most sustainable energy is the energy that you don’t use in the first place and in this article, we will give you a glimpse into three of the #green #solutions we have in our portfolio here at NTNU Technology Transfer AS that will help you with that. While they have varying degrees of technological complexity, maturity, and time to market, the one thing they have in common is a vast potential to significantly reduce global power consumption, electricity grid load and greenhouse gas emissions.

According to the IEA, 23% of the global final energy consumption is due to the need for space and water heating in buildings. An additional 10% is consumed due to the need for cooling, and this figure is expected to soar over the coming three decades.

This is where our first project, named DEP Joint, can make a significant impact. While it is the least technologically complex project, it is also the one that is closest to market. Invented at NTNU Department of Civil and Environmental Engineering by Habibollah Sadeghi, CGDIT? and Rao Martand Singh , their simple yet ingenious modification of traditional joints for pre-cast driven concrete piles will transform building foundations from passive structures to massive ground heat exchangers – without compromising on bearing capacity. Our initial calculations indicate that 60-70% of a typical building’s baseload for heating and cooling can be covered by energy harvested from the pile foundations. Thanks to support from Norges forskningsr?d (the Research Council of Norway), we are now running a verification project in collaboration with leading industry actors Leimet and Sandnes & J?rbetong. We are currently producing the first prototypes of the joints and will soon start casting a test batch of energy piles. The next step is to perform structural and hydraulic integrity tests on the joints before we run a full-scale field test on the energy piles. We hope to be able to offer the DEP Joint to the market in the second half of 2023.

Use of the DEP Joint also comes with the additional bonus of significantly reducing the load on the electricity grid. IEA states that electricity accounts for 20% of the global energy consumption and that 1% of the electricity demand is due to large data centres.

While this might not sound like much on a global scale, it has a large impact in the local scale, as demonstrated in for instance West London and Ireland. As to the former, the Greater London Authority has recently warned developers that they may not be able to start new housing projects in the area until 2035, as there is no available capacity in the electricity grid. Part of the reason for the grid overload is the large number of data centers that have been located there to take advantage of the fibre optic cable corridor along the M4. For the latter, both Microsoft and Amazon have been prohibited from hooking new data centres onto the Irish electricity grid due to capacity issues.

Even the most efficient data centres use around 10% of their power consumption for cooling. While the DEP Joint could help reduce this figure by 70%, our next project can help with the remaining 90% by making the software running in the data centres more efficiently.

This project is called TIP (short for Time-Proportional Instruction Profiler), and it is a hardware tool that will help make software more efficient. Profiling is a way to evaluate how the software running on your computer performs. When you know where your software spends most of its CPU time, you can inspect the code and uncover inefficiencies. Unfortunately, this is not always enough, as additional inefficiencies are introduced when your human readable code is translated into machine code that the computer can understand. With TIP, the inventors Magnus Jahre and Bj?rn Gottschall from the NTNU Fakultet for informasjonsteknologi og elektroteknikk (Faculty for Information Technology and Electronics) and Lieven Eeckhout from Universiteit Gent have created a novel hardware mechanism that provides an extremely accurate profile of where the translated machine code spends time so this can be optimized even after compiling. In one industry benchmark, TIP quickly identified instructions that didn’t actually do anything but were still executed – when these were removed, the performance increased by 93%. For performance critical operations like cloud services, such performance increases amounts to significant reductions in power consumption and the need for new hardware. As a comparison, removing similar, seemingly innocuous, instructions from the operating system of your mobile phone would amount to almost doubling its battery life. If we take a look at underlying numbers and consequences of software inefficiencies outside just power consumption, The National Institute of Standards and Technology states that there is an average of 25 errors for every 1,000 lines of code. The total cost of such poor-quality software is estimated by the Consortium for Information & Software Quality to exceed two trillion dollars in the United States alone. But while every cell phone manufacturer, major cloud provider and institution running compute-intensive operations will benefit from TIP, they need companies like 英特尔 , AMD or Arm to license and implement TIP in their next-generations CPUs before the benefits can be reaped.

Finally, no matter how efficient the software is, energy will still be needed to store data. Several studies show that the need for storage is rising exponentially – and unless the technology for storage changes fundamentally, so will the need for electricity powering the storage devices. AKCP estimates that almost 8% of the power in data centres in the United States are consumed by storage devices – and furthermore, that the lifespan of the devices are only around 4.4 years.

This is where our last project, SkyAxxell, can help. Aside from the technological potential, the project has already led to groundbreaking research results. Invented through an impressive collaboration effort between Erik Lysne, Dennis Meier, Erik Roede (from the NTNU Faculty of Natural Sciences and NTNU’s Center for Quantum Spintronics) and Markus Altthaler and Istvan Kezsmarki (from the Universit?t Augsburg ), SkyAxxel provides an innovative pathway towards more energy efficient data processing and storage. However, in order to get the technology ready for the market and produce the first proof-of-concept devices, additional development is required. For future information storage, magnetic skyrmions are believed to play a key role. Skyrmions are magnetic pseudo-particles that arise in certain materials on the nanoscale, and they can be controlled in an energy-efficient way using electrical currents. SkyAxxel is a novel space- and energy-saving driving mechanism for skyrmions, with the potential to finally make low-power skyrmion-based storage devices a practical reality. SkyAxxel should be a very interesting opportunity for companies like 希捷科技 , IBM , Western Digital and 三星电子 .

To sum up: While very different on the surface, all these three projects share the common attribute that they can help reduce the power consumption of data centres, each in their own way. DEP Joint will reduce the power consumption needed for cooling significantly. TIP will help software developers produce more efficient code, reducing the need for hardware and consequently also reducing power consumption and heat production. And finally, SkyAxxel will reduce the power required for storing the data.