Shulou(Shulou.com)06/03 Report--

Abstract

This technical paper clarifies the continuous cooling requirements under the background of Tier Standard:Topology of Uptime Institute. Tier IV is the only level that requires continuous cooling. However, Uptime Institute recommends continuous cooling when the density of each rack exceeds 4 kilowatts (kW), regardless of the level.

The effect of continuous cooling

With the increasing power density of data centers, the need for continuous cooling becomes more urgent. The risk of losing cooling in an uninterruptible power supply (UPS) outage event and the associated impact can be catastrophic for the enterprise. The IT device may be malfunctioning or damaged.

Depending on the cooling or UPS technology deployed in the device, the requirements for continuous cooling may vary greatly. In this paper, the definition of continuous cooling is clarified, and the matters needing attention in the deployment of various cooling technologies are described in detail.

Continuous cooling refers to the ability to provide a stable thermal environment for critical IT equipment without any interruption. Continuous cooling requires a stable server entrance temperature and the time it takes for the mechanical system to restart after any cooling system power outage (including the time to transfer to the engine generator, if applicable). It also requires proper maintenance of the designed cold channel temperature.

Maintaining a stable thermal environment helps to reduce the sudden increase in UPS output load. The increase in load may be due to increased power consumption of server fans to help make up for temporary cooling losses. If the load is not managed properly, the increased load will make the UPS overload and affect the operation of the whole equipment.

The American Society of heating, Refrigeration, and Air conditioning Engineers (ASHRAE), in cooperation with major IT equipment manufacturers, has developed thermal guidelines for data processing environments, including recommended inlet air temperatures for computer equipment to ensure reliable operation of servers, storage, and network equipment. As of 2015, the ASHRAE guidelines (acceptance Global) recommend that the equipment inlet dry bulb temperature be maintained between 18-27 °C (66-81 °F), and the dew point temperature 9 °C to 15 °C (scale °F) and relative humidity below 60% to meet the manufacturer's standards. It is important to note that for the actual server entrance temperature, the requirement for continuous cooling is independent of the ASHRAE guidelines. The server entrance temperature following the ASHRAE guidelines is based on the owner's decision of the organization's individual needs.

However, the uptime Institute's continuous cooling requirements are linked to ASHRAE's allowable rate of change guidelines because it involves the definition of a stable thermal environment. The ASHRAE guidelines specify the maximum allowable changes in the inlet temperature of IT equipment. Tape storage data center (specialized) the maximum allowable temperature change for a typical IT device is limited to 5 °C per hour. All other devices are limited to a maximum allowable temperature change of 20 °C / h. With data centers and any type of equipment, ASHRAE limits the speed of this change to 5 °C in any 15 minutes. It should be noted that this is not a rate of change, but a discrete range of temperatures up and down.

A continuous cooling solution must be able to provide a stable thermal environment within the time required for the entire mechanical cooling system to restart.

These parameters have a lot of influence in practical application. To give an example:

Consider a data center with a cold channel seal, in which the chilled water computer room air treatment (CRAH) device uses the air supply temperature control method and the normal set point is 20 °C. If the control dead zone is set to 1 °C, this means that during normal operation, assuming there is no air mixing between the CRAH unit and the IT device, the device will provide an IT device inlet temperature of 19-21 °C.

If a failure or power loss occurs when the equipment supplies 19 °C air to the critical environment, and the failure causes the CRAH equipment to raise its temperature to 24 °C at its peak, the inlet temperature cannot deviate from 19-24 °C in 15 minutes. This means that if the equipment is restored within 15 minutes, the equipment should not be so cooled that the air supply temperature drops below 19 °C.

Maintaining this temperature range requires close attention to the control algorithms used to guide the cooling of CRAH cells. Using the same example, if the failure results in a peak air supply temperature of 22 °C before the recovery, the recovery can be cooled to 17 °C to remain within the limit of 5 °C for any 15 minutes.

The time required to restore mechanical cooling must also be taken into account. Although ASHRAE uses a 15-minute cycle to define the maximum allowable temperature change, the Standard: topology also requires a continuous cooling solution to provide a stable thermal environment for the entire time required to restart the mechanical cooling system after any cooling production interruption or power outage. The restart time of the mechanical system starts from the moment when the power is lost, until the engine generator (or other field power generation system) starts and shuts down to the critical load, when the mechanical system resumes power supply operation, and can provide rated cooling capacity under steady-state operating conditions. For example, chillers cannot be counted as formal operation until normal water supply and backwater conditions and flows are restarted and operated.

Although the manufacturer is reducing the device restart time, the interval between the power loss and the ability to restore the system to produce stable cooling needs to be used as a data point when determining the transit time. For example, if it takes 10 minutes to restore stable mechanical cooling after a power outage, then the cold storage tank (TES) must be able to provide 10 minutes of frozen water storage.

Although Tier IV is the only Tier level that requires continuous cooling, data centers with higher-than-average IT load density should consider using continuous cooling to mitigate significant temperature increases due to common power supply or component failures. As a reference, Uptime Institute conducted a demonstration of a 6 kilowatt / rack average power computer room. After the cooling failure, or even just after the loss of air flow, within 60 seconds, the intake temperature in the computer room exceeds the maximum value set by the computer room. In addition, the demonstration shows that a 1-minute cooling loss takes 20 minutes to recover.

Considering the power outage of the city, UPS continued to supply power to the IT equipment during this period, but the operation of the mechanical equipment was interrupted. Depending on the cooling technology deployed, this interruption may last for several minutes. During this period, any increase in temperature in the computer room may damage the IT equipment. Continuous cooling provides a bridge to keep the thermal environment stable until the mechanical or other cooling system is restored. A properly designed continuous cooling solution can prevent any increase in the average server intake temperature.

Continuous cooling provides a bridge to keep the thermal environment stable until the mechanical or other cooling system is restored.

Application of continuous cooling requirements

T4 is the only level that requires continuous cooling. Here are some examples of ways to achieve continuous cooling. Rating standards are not normative, but provide results-and performance-based standards, so there may be other ways to achieve the goal of continuous cooling on the basis described below.

Example 1:

Continuous cooling for chilled water systems is usually accomplished using TES capacity (also known as chilled water storage). The secondary pump and CRAH also need to be installed on the fault-tolerant UPS power supply. The power supply can be IT UPS or a separate, parallel maintainable and fault tolerant UPS system dedicated to mechanical systems. If the cooling system is in the primary pump system, the main pump needs to be installed on the UPS. In addition, cold storage tanks and how they are connected to chilled water distribution must be considered. Depending on the connection of the tank to the chilled water distribution system, the chilled water supplied and returned (and therefore warmer) can be mixed, which may reduce the time during which the chilled water tank can provide a stable thermal environment for the IT equipment.

Example 2:

The continuous cooling of the direct expansion (DX) system requires the computer room air conditioner (CRAC) and the external condenser to be on a UPS system that can be maintained and fault-tolerant at the same time. In addition, the DX system may require some additional engineering analysis to meet the continuous cooling requirements. During the normal operation of the CRAC, the compressor cycle can suspend the operation of the compressor for several minutes to protect the compressor. In an environment with higher than average density, these gaps may lose public power during normal operation and failure, causing large temperature fluctuations and making a stable thermal environment more difficult to achieve.

Example 3:

For continuous cooling of evaporation systems (such as direct or indirect evaporative heat exchangers), pumps (secondary or system circulation) and distribution fans are required on fault-tolerant UPS power supplies.

Example 4:

Continuous cooling of 100% external air systems that can provide cooling throughout the year requires fans (or systems that deliver air to the computer room) to be installed on a fault-tolerant UPS power supply.

When the runner UPS system is deployed as an IT UPS (the runner UPS should be on the diesel) and the cooling system is located on the uninterruptible bus, additional continuous cooling measures may not be required because the mechanical system will not be interrupted (or other on-site power) during the conversion from the utility to the engine-generator. If the cooling water system is deployed with a rotary UPS without batteries, the site must demonstrate that if TES is not installed, the average server inlet temperature will not exceed the previously specified limit. Each specific case should be reviewed to ensure that the requirements for a stable thermal environment are met during a power loss event.

In view of the possible damage to facilities and IT assets, continuous cooling is a reasonable protection for any facility with an average density of more than 4 kilowatts.

Conclusion

By providing thermal stability to the IT environment during any interruption of the cooling system, such as the transition from utility outages to engine-generator power, continuous cooling ensures that utilities do not cause high-cost thermal damage to IT hardware or critical equipment and do not increase the UPS output load to the overload point. Continuous cooling is only a requirement for Tier IV certification, but for possible damage to facilities and IT assets, continuous cooling is a reasonable guarantee for any facility with an average density of more than 4 kW.

Modify

ATD technical paper series: continuous cooling, version b. All updates to this version will take effect in October 2017.

Related publications

Topology level criteria:

Series of technical papers for certified designers

Research Institute of uptime

Uptime Institute is an impartial consulting organization dedicated to improving the performance, efficiency and reliability of critical business infrastructure through innovation, collaboration and independent certification.

Uptime Institute serves all stakeholders responsible for the availability of IT services through industry-leading standards, education, peer-to-peer networking, consulting, and incentive programs to enterprise organizations and third-party operators, manufacturers and suppliers. Uptime Institute is a globally recognized first-level standard and certification for data center design, construction and operation sustainability, and effective endorsement of the creation and management body review, FORCSS ®methodology.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.