You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
More InformationHow NTT DATA and WAGO save energy in data centers.
Digital applications, cloud computing and artificial intelligence are growing around the world at breathtaking rates. Global Data Centers (GDC) is a single entity that operates 45 data centers in the EMEA region. In order to control and reduce their power consumption in a targeted way, GDC relies on the interactionbetween artificial intelligence and automation components from WAGO. The following article shows how this data is collected and analyzed to verify efficiency potentials.
In the fall of 2024, news readers in Germany saw an astonishing headline: while Germany is currently in the final stages of ending the use of both nuclear energy and coal-fired generation, operators of large data centers in the USA are considering the installation of their own miniature nuclear power plants at their locations. According to the article, the rapid growth of applications using artificial intelligence are making computers and storage media run hotter and longer. The energyconsumption of data centers and smaller IT installations in Germany alone was almost 18 billion kWh in 2022 – sufficient to supply 5 million average German households with electricity for one year.
Using Energy Sensibly
In order to significantly reduce data center energy consumption and CO₂ emissions, the power usage effectiveness (PUE) value for existing buildings must be reduced by 2027, due in part to the EU Ecodesign Directive. It is true that the PUE values for data centers with more than 40 KW of IT connection power improved from 1.98 to 1.55 between 2010 and 2022. However, if, as some experts have prognosticated, this trend were to continue, then the energy consumption of data centers in Germany alone could increase to around 27 billion kWh per year by 2030.
With these figures, it is not surprising that the major players in the sector are deeply involved in discussions about efficiency. In Europe, for example, this is NTT DATA‘s Global Data Centers division. Still known to many as e-shelter, the NTT DATA network now covers more than 20 countries and markets worldwide with over 1,600 MW of IT load. Locations in Germany, Great Britain and other European countries play a central role in the company‘s global presence.
AI Pilot Project Started
The subsidiaries in these markets also want to assume a lead role in energy efficiency, “This has been an issue for us for many years. And we have sought to find sensible, individual solutions for the various locations,” says Steffen Benson. He is Senior BMS Manager at GDC and responsible for Europe and also the locations in the Middle East and Africa. “Since the different locations were established and have
grown over the course of many years, such concepts are,
unfortunately, not universally applicable. Therefore, we have
inaugurated a project to find an optimization solution for all
locations,” he says.
Ultimately, it was the impetus of a trade show conversation that led to this backing of artificial intelligence. It is intended to monitor energy consumption, cooling topologies, power distribution, sensors – in short, essential elements of the energy infrastructure – to make predictions about expected consumption and, from there, to set all aggregates to their optimal operating state every minute.
For this purpose, the mass flows are monitored from cooling generation up to their distribution, and adjusted as needed, so that the pumps operate at maximum efficiency. It goes without saying that the operational stability of the server park must not be affected. “Of course, this type of artificial intelligence is a powerful tool, and it can do all of this, but it also requires that we retrofit and renovate a large part of our operating technology,” explains Benson.
This has been an issue for us for many years. And we have sought to find sensible, individual solutions for the various locations.
WAGO Controls the Cooling System in the Data Center
For many years, automation components from WAGO have been used to monitor and control the building technology systems at GDC. The hardware and software products from Minden monitor heating, ventilation and air conditioning systems and control energy management in a wide variety of building types. “Back in 2005, under the company name eshelter, we tasked WAGO with controlling and monitoring our system technology. One of the most important arguments for us at that time was their industrial PLC, which supports Modbus® and SNMP protocols, and has HTTP capability. In addition, the controllers have a very compact and modular design. Thanks to these features, we are well prepared for the future and continue to build on our excellent collaboration with WAGO”, explains Benson.
The only difference between then and now is that the information from the countless sensors, sensors and HMI devices is no longer exclusively collected at the control level, where it is available to the building management. Instead, the data, now transmitted via the Modbus® protocol, is also provided to artificial intelligence for evaluation. This is necessary in order to be able to perform the complex calculations. “Since WAGO is a Linux®-based PLC system, this offers us the best opportunity for ensuring a secure interface to the AI,” says Benson.
Another plus point in terms of security: WAGO products have passed all of the penetration tests set by an independent external IT security company – a decisive factor, as data centers are considered critical infrastructure. WAGO supports the motto of “OPEN. To simplify security” as a practical means for implementing the interplay between security and openness.
Linux®, Modbus® and CODESYS in Interaction
Operating data is collected in real time from the sensors, actuators and system components, using the Modbus® protocol, which is then used by the AI software for analysis and optimization. The WAGO PFC200 controllers used in the project support programming languages according to IEC 61131-3 and function in a Linux®-based real-time environment using CODESYS V3.5.
The programmers created modules directly in CODESYS, which communicate via the Modbus® protocol. This allows easy integration of various systems and devices, while offering an open and flexible integration into the overall system due to WAGO technology.
“We are currently working on equipping 35 out of the 45 buildings in Europe with the AI-based optimization. This means that around 2,000 controllers will be affected, 1,300 of which belong to previous generations, meaning hardware that we have to replace.
»One essential aspect that occasionally drops off the map is the size of the installation. With WAGO’s modular approach, we have the option of connecting significantly more end devices or systems to one node. This can save several meters of control cabinet in some cases.
The remaining 700 devices only need to be reprogrammed,” Benson says, giving an insight into the scope of the project. “Of course, this is a mammoth project! Which is also primarily due to the fact that the systems will remain in operation, and customers will not experience any restrictions,” he clarifies.
Modular Design Advantages
This is the point at which the team around Benson will benefit from the structure of the WAGO hardware, “The great advantage of the modular system is that we only replace the smallest part. We can continue to use the input and output cards and do not need to make any installation changes. In addition, we were able to program the PLC in advance. We have decoded the previous project and, naturally, made a few changes.
In the commissioning phase, all we needed to do was switch our control cabinets to a safe state, and then replace them in the live system, sort of like plug and play components,” says Benson. The success of the proof-of-concept is already clear, even if Benson and the team are still in the data acquisition phase, “We expect to reduce the energy needed for cooling by 35 to 40 percent. At this rate, we will recoup the investment costs in less than a year of operation,” states Benson.
He is already looking forward, “The journey we are on will take several years. Due to the decision to continue implementing the project with WAGO, we immediately began using the new controllers and CODESYS as a programming environment. This ensures that we are future-proof and can convert one location after another to energy-saving operation,” says Benson with satisfaction.
POWER USAGE EFFECTIVENESS
Focus on Savings in the Auxiliary Systems
High computing power requires high energy consumption. This relationship cannot be solved at the moment; however, potentials still remain for optimizing the infrastructure units required for system cooling.
Therefore, the goal of energy and CO2 regulations is to reduce the energy used by data centers for cooling, lighting and other non-IT related infrastructure, in order to increase their overall efficiency. The Power Usage Effectiveness (PUE) value indicates the ratio between the total energy consumption by a data center and the energy consumption by the IT hardware.
In this case, a lower PUE value means better efficiency. Industry standards and guidelines for energy efficiency, such as those proposed by the European Code of Conduct for Data Centers, suggest that efforts should be undertaken to achieve PUE values of less than 1.5 for new data centers and approximately 1.2 for particularly energy-efficient facilities.
Benefits at a glance:
- Modular system architecture for flexible scalability
- Maximum availability with high energy efficiency and safety
- Safe monitoring and control solutions for data center infrastructure
Frank Jablonski, freelance journalist, mylk+honey, Würzburg
Text taken over from original – WAGO
