December 19, 2024 A Power Management Playbook for the AI Era

How manufacturers can optimize their backup power strategy to align with rapid technology advancement.

By Rick Conneely, Growth Segment product management, Eaton

The noise around AI in the industrial sector is no longer just a whisper – it’s a roar. According to a recent report from Fluke Reliability, manufacturers are allocating a whopping 44% of their technology budgets to AI in 2024. According to the survey, 61% of manufacturers are expecting to achieve their AI goals in just 11 months.

As AI use grows in the industry 5.0 era and more manufacturers pilot use cases for machine learning applications, they must evolve their approach to power management in alignment with business and operational priorities. With a strategic approach, manufacturers can benefit from AI and other emerging technology solutions while ensuring uptime, supporting sustainability objectives and advancing innovation.

Call for backup power

Manufacturers may find themselves trying to strike a delicate balance between operational priorities when assessing the power requirements for AI. Many have begun leveraging advancements in automation and machine learning to streamline processes across their facilities, yet reliability continues to be a paramount consideration for this ongoing transition to be successful. Downtime is already a major concern for the industry, costing the manufacturing market $50 billion per year according to recent estimates.

To help ensure uptime and safeguard critical operations, manufacturing IT leaders must have an integrated backup power strategy that incorporates uninterruptible power supplies (UPSs) serving as the important bridge to generator power in the event of an outage. Manufacturers can integrate these systems with a variety of innovative solutions to optimize their strategies in parallel with their objectives for AI across operations.  

Deploying UPSs with lithium-ion battery technology can help significantly improve service life, provide faster recharge capabilities and a reduced footprint for these solutions in comparison to traditional valve-regulated lead acid (VRLA) batteries. Keeping up to date with regularly scheduled maintenance on the UPSs, their batteries and generators will provide IT managers with the confidence that the systems will remain in position to perform well in the case of an unplanned outage.

Another key advantage of lithium-ion batteries, which will likely become more popular as sustainability demands increase, is their inherent energy storage capabilities with highly precise State-of-Charge (SoC) monitoring. Data analytics and sophisticated controls can be utilized to turn a UPS into a distributed energy resource (DER), creating a bi-directional flow of energy to and from the grid. Utilizing this approach with technologies like Eaton’s EnergyAware UPS, manufacturers can offset energy costs while helping integrate more renewable energy to advance sustainability goals.

Embrace digital transformation

In addition to piloting machine learning applications, many manufacturers have embraced some level of digitalization to accommodate these solutions and maximize productivity. Based on the “The Intersection of digital transformation and the energy transition” report,  a research survey commissioned by Eaton and developed by S&P Global Market Intelligence, around half of companies in process and discrete manufacturing report that they are actively executing a digital strategy.

Software tools, such as Eaton’s Brightlayer Data Centers Suite, have emerged amid this transition to help manufacturers effectively monitor and manage technology assets and critical infrastructure. Power monitoring software, for example, can provide IT teams with real-time power and environmental system insights at a single facility or multiple locations, helping to reduce power consumption costs and avoid unplanned downtime due to system failure.

Solutions such as Distributed IT Performance Management (DITPM) software (DITPM) also help IT teams maintain control over power systems across distributed networks, even where technicians aren’t onsite. Aligning these solutions with power devices via integrated network cards allows manufacturers to be proactive with remote management – making timely firmware updates to stay on the forefront of cybersecurity demands. The software also allows IT teams to gracefully shut down or re-boot a ‘hung’ server or other equipment from a remote location in the instance of a power event.

Manufacturers also have the ability to leverage advanced predictive analytics services to monitor and analyze power trends over time. This allows IT teams to make smarter decisions about their equipment and proactively fix or replace devices rather than retroactively responding after an event causes downtime.

Hardware essentials for AI

While software strategies might be the most impacted by AI’s emergence, hardware demands are also evolving. AI servers are often larger, heavier and hotter than other industrial servers and require more power. They may not fit in standard IT racks. Heavier servers also make it impossible to fully populate a standard rack without exceeding the rack’s weight capacity.

Manufactures should consider new heavy-duty, floor-standing enclosures as a future-proof solution to accommodate the changing specifications of IT equipment. Advanced enclosures have weight capacities up to 5,000 pounds (static) to support a new generation of servers. Extra deep 54-inch enclosures can house large AI servers, which are typically 45 inches deep, and provide space for cables and cable managers. Wider options also have side channels for accessory mounting, cable routing and airflow.

New heavy-duty enclosures are also available in powder-coated white, which makes it easier to locate specific cables or ports on dark equipment. Better visibility helps facilitate equipment installation and maintenance and helps reduce accidental disconnections. White racks are becoming more popular for IT teams with the growth of AI and other technological advances.  

Combining the best in both digital and physical power management solutions will help to ensure a well-rounded strategy. To that end, IT managers should take measures to incorporate industrial-grade hardware devices, such as surge protectors, power strips and rack enclosures to cover their bases on power protection.

Balance progress with poise

The emergence of AI is demonstrating a positive impact in manufacturing, helping address operational challenges ranging from workforce skills gaps to production efficiency and sustainability. However, as companies embark on ambitious projects in the industry 5.0 era, it is crucial to ensure their power solutions keep pace. By implementing an integrated power management strategy that incorporates the latest physical and digital innovations, manufacturers can ensure resiliency of critical infrastructure exploring the burgeoning possibilities of an AI-enabled future.

About the Author
Rick Conneely leads the Growth Segments Team for Distributed IT at Eaton, where he is responsible for developing growth in markets and applications adjacent to traditional IT infrastructure such as Healthcare, Industrial and Energy Transition. He has 25+ years of successful results in the electrical, IT and industrial markets driving growth through new products, strategic partnerships and M&A.

Subscribe to Industry Today