By eliminating traditional air conditioning and harnessing 100% of the heat generated by its electrical components, all without having any impact on the landscape, D4 stands as one of the world’s most sustainable and efficient data centers.

Infomaniak aims to make this innovation widely accessible. Rather than retaining exclusive rights, they are committed to making their heat recovery technology open-source and easily replicable across the industry. This initiative encourages other data center operators to adopt similar ethical and sustainable practices for collective environmental benefit.

This page provides a detailed guide to replicating a data center like D4. We will begin by conceptually describing the heat recovery system that serves as the primary operating mode of the data center, along with the emergency mode design, which builds on Infomaniak's experience with their previous data center, D3. Next, we will present detailed diagrams of D4, including precise measurements and key architectural characteristics. Additionally, we will outline the technical specifications of D4’s components, accompanied by their respective technical sheets. Finally, we will discuss critical considerations for selecting a suitable site for constructing a similar data center.

Heat Recovery System

Infomaniak's D4 heat recovery system is a pioneering solution designed to repurpose 100% of the heat generated as a byproduct by its electrical components to heat up homes in Geneva, with the projected capacity to heat up to 6,000 Minergie A-certified households annually. At full capacity, D4 will consume 14 GWh of electricity annually, generating an equivalent 14 GWh of heat per year. In Switzerland, the primary sources of residential heating are heating oil, heat pumps, natural gas, and wood. If D4’s heat were to replace heating that would otherwise rely on natural gas, heating oil or wood pellets, it could prevent the release of approximately 3,388 tonnes, 4,312 tonnes or 6,440 tonnes of CO₂ emissions, respectively [1]. This is equivalent to the CO₂ emissions generated by approximately 1,694, 2,156, and 3,220 round-trip flights between London and New York, respectively [2].

The following section provides a detailed overview of D4’s heat recovery system, designed for seamless implementation in other facilities.

Concept

The following diagram represents a vertical cut of D4 showcasing the design of the data center.

The D4 heat recovery system operates as a closed system, meaning it is self-contained and does not exchange matter with its surroundings. It consists of three interconnected closed loops, which are systems where the working fluid continuously circulates within a sealed pathway, ensuring no external input or loss of the fluid during operation, as illustrated in the diagram below.

1st closed-loop system: from the servers to the air-to-water exchanger

Electricity powers the data center, enabling the servers to function. During operation, the servers generate heat, producing warm air at around 45°C, which is expelled into hot aisles. The warm air naturally rises toward the ceiling, and at one end of the closed-off server room, ventilators placed behind an air-to-water exchanger guide this warm ceiling air toward the exchanger. An air-to-water exchanger is a system designed to capture heat from the air and transfer it to water through a heat-conducting surface or medium. The exchanger is composed of coils within which room temperature (approximately 28°C) water flows. As the ventilators direct airflow across the room, the warm air passes through the exchanger, where the heat is absorbed into the circulating water, cooling the air to roughly 28°C in the process.

Following this heat transfer from the air to the water, the cooled air is pushed by the ventilators under the raised floor that supports the servers. The airflow creates a low-pressure area, encouraging the cooler air to flow up through perforations in the floor, accessible only in the cold aisles. This cool air goes through the intake of the servers and cools them down, absorbing heat once again, and the cycle for this 1st closed-loop system repeats.

The heat generated by the other electrical components of the data center such as UPS systems, fans, security cameras, electrical panels, etc., is also collected and directed towards the air-to-water exchanger, resulting in D4 truly reusing all the heat it generates.

2nd closed-loop system: from the air-to-water exchanger to the heat pump

Unlike the 1st closed-loop system which was based on the movement of air, this 2nd closed-loop system is based on the movement of water.

After exiting the air-to-water exchanger, the water isn’t yet hot enough for transmission to the district heating systems of Services Industriels de Genève (SIG). In order to be suitable for SIG’s district heating installations the water must be at 67°C in winter and 82°C in summer. Therefore, after being warmed by the air-to-water exchanger, the water enters the heat pumps where the gas in the heat pumps expands by absorbing energy from the water, which cools the water to 28°C. The now room-temperature water is directed to the air/water exchanger and the cycle for this 2nd closed-loop system repeats.

3rd closed-loop system: from the heat pumps to SIG’s district heating

In this 3rd closed-loop system, the district heating water flows into the heat pumps where the heat pump gas, which has expanded by absorbing the heat from the warm 45°C water that exited the air-to-water exchanger, is now compressed releasing heat and warming the district heating water to 67°C in summer or 82°C in winter.

In a heat pump, the gas releases more heat than it absorbs because of the external work done during compression. When the gas is compressed, energy is added, increasing its temperature and total energy content. This allows the gas to release not just the heat it absorbed during expansion but also the additional energy from compression. As a result, the total heat released is the sum of the absorbed heat and the work done by the compressor. This process enables heat pumps to efficiently transfer heat from a lower temperature source to a higher temperature destination.

In order to create this external work and compress the gas, heat pumps require electricity. The appeal of heat pumps lies in their exceptional efficiency: they can deliver significantly more heat than the electrical energy they consume. This efficiency is quantified by the Coefficient of Performance (COP), a measure of the ratio of heat output to electrical energy input. For instance, a heat pump with a COP of 3 can produce 3 MWh of heat energy while consuming only 1 MWh of electricity. The higher the COP, the more efficient the system. The COP depends on factors such as the temperature difference between the heat source and the destination; smaller temperature differences yield higher COP values.

In D4, the heat pumps serve a dual purpose, they are used to heat up the water from SIG’s district heating and to cool down the servers. Indeed, it is the heat pumps that cool down the water from the 2nd closed-loop system which in turn cools the air in the 1st closed-loop system through the air-to-water exchanger, which then cools down the servers.

Site Selection for a New Data Center

Selecting the site for the data centre is a crucial step in the project. Designed as a model of sustainable urban infrastructure, the facility will operate without causing noise, emissions, or visual disruption by being built underground within the eco-district park. This approach preserves undeveloped land, optimizes urban space, and seamlessly integrates the data centre into a multifunctional environment that benefits both the community and the local ecosystem.

A key factor in site selection is the availability of an existing district heating system, which allows the data centre’s heat recovery capabilities to be integrated efficiently. By locating the facility near residential areas already connected to this network, the recovered heat can be distributed to homes and businesses, improving energy efficiency and reducing reliance on traditional heating methods. The heat generated by the data centre’s servers will be captured and distributed using technologies compatible with district heating systems, ensuring the network can absorb and redistribute the energy continuously. For instance, the system must be capable of handling the project’s full heat capacity, estimated at 1.7 MW, without interruption.

Energy availability is another critical consideration, as seamless integration with the network ensures efficient use of recovered heat throughout the year. The flexibility to adopt different energy transfer methods, such as water-cooling systems, further enhances adaptability and sustainability.

Additionally, evaluating the local energy landscape involves analyzing the current grid capacity and identifying opportunities to optimize energy use and revalorize resources, regardless of the energy source, to enhance the data centre’s sustainability. Incorporating renewable energy options not only reduces the data centre’s carbon footprint but also enhances the overall sustainability of the regional energy system. This approach ensures that the data centre’s power requirements can be met without straining the existing electrical supply. By choosing a site with energy integration potential, the project supports a circular energy economy, emphasizing the reuse and recovery of energy to enhance the area’s energy efficiency and environmental goals, regardless of the original energy source.

One approach is to place facilities underground, reducing environmental and visual impacts while preserving the aesthetics and multifunctionality of urban areas. For example, embedding infrastructure beneath parks can avoid converting natural spaces into single-purpose developments.

What Are the Benefits of This Approach?
By minimizing surface disruption, this method supports sustainable urban planning and promotes circular energy systems. Repurposing waste heat, for instance, can benefit local communities by feeding into district heating networks, improving energy efficiency, and reducing reliance on traditional energy sources. Environmental factors and biodiversity considerations are integral to the site selection process for the data centre. The goal is to minimize the ecological impact of the facility and ensure its sustainable integration into the surrounding environment. This involves conducting a thorough assessment of the site’s existing biodiversity, evaluating potential disruptions to local wildlife and ecosystems, and implementing measures to protect and enhance natural habitats. By selecting a site that already supports a sustainable urban development, such as an eco-district park, the project can avoid disturbing untouched greenfield areas and instead make efficient use of multifunctional urban spaces. The data centre’s design prioritizes sustainability by preserving natural aesthetics and promoting biodiversity. Its integration into the surrounding environment supports eco-friendly practices, including energy recovery systems that repurpose waste heat to benefit the community. These efforts align with broader environmental and sustainability goals, ensuring both functionality and minimal ecological impact.

Evaluating land costs helps ensure the project remains within budget while selecting a location with infrastructure capable of supporting future expansion and increased energy integration. Adhering to zoning regulations is essential to avoid legal complications and accelerate the approval process. Given the rarity of industrial activities in residential zones, collaborating with local authorities and stakeholders is critical to ensure compliance and community acceptance. The zoning must not only accommodate data centre operations but also support the integration of renewable energy infrastructure, such as solar panels. This often involves coordination with urban planners and local councils to align with sustainability goals while addressing regulatory constraints.

The impact of local development policies is a key factor in the site selection for the data centre. Local government policies and development plans can significantly influence the feasibility and future success of the project. Favorable policies that encourage sustainable infrastructure, renewable energy integration, and heat recovery can offer tax incentives, grants, or fast-tracked permitting processes, making it more attractive to locate the data centre in certain areas. Local policies promoting eco-friendly initiatives, such as district heating systems or green building certifications, can facilitate the integration of the data centre’s heat recovery system with the existing urban infrastructure. By aligning the project with local development goals, the data centre not only supports regional sustainability targets but also positions itself as a model for future urban planning and sustainable development initiatives. Additionally, such alignment can facilitate access to financial aid or subsidies, further enhancing the project’s feasibility. Understanding these policies ensures that the project can navigate regulatory requirements smoothly while contributing to the local community’s long-term environmental and economic objectives.

In conclusion, the site selection for the new data centre is a complex and multifaceted process that balances technical, environmental, and community considerations to ensure sustainable integration into the surrounding area. By prioritizing proximity to residential zones, the facility can efficiently link its heat recovery system with the district heating network, enhancing energy efficiency and reducing reliance on traditional heating methods. The commitment to using renewable energy sources and integrating with the local energy grid not only supports the region’s sustainability goals and minimizes the centre’s carbon footprint but also helps avoid the need for developing more polluting energy sources, such as natural gas or pellet power plants.

By integrating environmental and biodiversity considerations, the project safeguards the natural landscape and supports local ecosystems. Aligned with urban planning policies and zoning regulations, it demonstrates economic and regulatory sustainability. This data centre serves as a model of sustainable infrastructure, contributing to the community’s well-being and inspiring future developments.

VII) RSK Non-Return Valve

• Purpose: An RSK non-return valve in a data centre prevents backflow in drainage or cooling systems, ensuring unidirectional flow to protect equipment and maintain system efficiency.
• Type and Material:
  • Fully constructed from stainless steel (1.4301 or 1.4571).
  • Frame made from C-profile galvanized steel or aluminum, bolted at corners without any welding.
  • Suitable for all standard duct connections.
• Dimensions:
  • Connection flange widths: 25 mm, 36 mm, or 50 mm.
  • Body depth: 165 mm.
  • Blade width: 100 mm with EPDM gaskets.
• Design Features:
  • Rigid hollow-profile blades with interchangeable EPDM seals, resistant to temperatures up to 80°C.
  • Equipped with additional side seals.
  • Axes made of aluminum, securely inserted into hollow sections of the blade profile.
  • Supported on both sides with either:
    • Ball-bearing joints (steel).
    • Oil-impregnated sintered bronze plain bearings.
• Mounting Configurations:
  • Suitable for both horizontal and vertical airflow directions.
  • Horizontal airflow requires horizontal blade orientation.
  • Mounting direction must be specified during ordering.
• Optional Features:
  • Gas-tight external casing.
  • Interior mineral fiber insulation with a thermally separated secondary casing.
  • Axes supported by articulated ball bearings.
  • External linkage for blade operation.
  • Anodized or powder-coated finish.
  • Material certificates in compliance with EN 10204-3.1.
  • Seismic calculations and factory inspections available.
  • Additional accessories:
  • Grilles.
  • Limit switches.
  • Sealing class options: ~ CEN 1751 class 2.
  • Extended temperature resistance:
  • 80°C (standard).
  • 200°C or 600°C (optional).

Diagrams of the Data Centre