In the world of data centers, reliability often seems to come at the expense of sustainability. However, this trade-off is a thing of the past, thanks to machine learning and AI advancements. This blog debunks common myths, showcasing how AI-driven optimization can enhance reliability and sustainability in data centers while boosting team productivity.
The Myth of the Trade-Off: Reliability vs. Sustainability
An article I recently read ranked uptime as the top three priorities for liquid cooling in data centers, with infrastructure costs following closely. Sustainability barely made it into the top ten. Data centers are undoubtedly in the uptime business, and everything else is secondary.
This highlights a persistent but outdated belief: reliability and sustainability are mutually exclusive. It’s time to challenge this notion using the appropriate digital tools. As an added bonus, this same digital tooling will make your teams more productive.
Data centers are growing at an unprecedented rate but face significant challenges that are hindering the ability to deliver the next generation of digital infrastructure. These including a lack of energy, skilled personnel, denser workload requirements, and increased regulatory pressures. To meet these challenges, data centers must embrace a digital transformation and leverage AI to drive efficiency and sustainability without compromising reliability.
But how?
At Coolgradient, we have optimized over 28 data centers across 3 continents, both new and legacy sites. Our track record speaks for itself – an optimized data center is not just more sustainable but also more reliable. With our solutions, data centers can confidently navigate the challenges of the future, ensuring energy efficiency, operational stability, and team productivity.
Hybrid cooling optimization
Data centers are now using hybrid cooling units to meet challenging PUE (Power Usage Effectiveness) targets and make use of free cooling. However, these units often do not operate as efficiently as intended. Our AI-powered SaaS tool is designed to pinpoint which hybrid cooling units are not optimized and to quantify the energy difference between their current operation mode and an optimized state. We then provide automated recommendations on how to best optimize the system.
By implementing these recommendations, your team can reduce energy and water usage and minimize chiller runtimes, thereby enhancing the reliability of your facility assets. This helps in extending the lifetime of your assets and reducing operational risks. With the assistance of AI, your team can increase reliability and improve sustainability in less time than before. We’re proud to say that, so far, we’ve saved our customers more than 10 GWh in energy by optimizing hybrid cooling systems.
Facilities harmonization
Just keep in mind that these chiller assets are not standalone; they are part of a larger facilities ecosystem that needs to adapt to ever-changing environments. Utilizing AI optimization ensures that your entire system operates in harmony. For example, if your chillers are not functioning properly, it could be causing your pumps to work overtime to compensate. So, what happens if you make changes upstream? Will there be problems downstream? This pump speed example is another frequent occurance where run times are higher than they need to be, causing considerable energy waste and decreased reliability if not detected and addressed well.
Floor Pressure and Fan Speeds
The next domain where we’ve achieved significant energy savings and increased reliability is the optimization of air pressure and fan speeds in rooms with CRAC/CRAH units. Often, these systems are overworked due to (IT load) imbalances or inefficiencies elsewhere in your cooling system. Coolgradient’s solutions have saved customers 11 GWh of energy by addressing these issues, also reducing runtimes, and improving overall efficiency.
Prescriptive Maintenance
Before optimizing your facility’s settings, it’s important to address any operational issues first. Coolgradient’s AI Optimizer can identify even the unknown issues in your systems by continuously analyzing hidden patterns in the data. These often show up as energy or water wastage on our platform. For instance, clogged filters are a relatively common finding, even in new data centers and outside of scheduled maintenance windows. They can be detected by unexpected chillers behaviour, often resulting in electricity and water waste. Our machine learning tools provide specific advice on the root causes and how to address them. Just a few examples of the valuable insights our platform can provide and as data centers become increasingly complex, it’s essential to equip your teams with the right tools to effectively manage them.
Productivity Gains
There’s yet another benefit from the power of AI: it makes your teams more productive.
The vast amount of data generated by data centers can overwhelm teams. Coolgradient’s machine learning engines, trained on billions of data points, identify specific optimizations and their impacts automatically. This information is all your people need to use and empowers any teams to prioritize actions with the most impact.
Improve Reliability AND Increase Sustainability while saving time
So far, we have helped our customers deliver over 21GWh in energy savings, mainly by reducing unnecessary run times for their facility’s equipment. When data centers operate as designed, it reduces risk and improves reliability, keeping uptime as the main goal. With the help of AI, you can achieve both uptime and improved sustainability.
You will rest better. Your equipment is not working too hard. Nor is your mission critical team.
Be Cool, Stay Up!
Read more use cases and how we optimize any facility here.
