Despite what many of us may think in the UK on this cold and snowy December day, there is one thing we can guarantee about our climate and the global climate in general. It is not getting any cooler and the demand for air conditioning systems continues to grow worldwide. This growth is also fueled not just by climate change but the growing tendency for IT based operations (server rooms and datacentres) and the general growth of industry and commercialisation across our planet. For many people, air conditioning could be considered a utility-type service without which they simply cannot function.
Most air conditioners provide two ambient environment related functions: cooling and heating. When an air conditioner fails to operate at the required thermostatically controlled temperature the implications can range from general discomfort to service interruption and even a critical system failure. In terms of server rooms and datacentres operators need to ensure that the environment they operate their IT systems within is within an ambient temperature of 20-25°C. This is important to protect temperature-sensitive systems within the environment and also to provide comfortable working conditioners for the people who have to entire the ‘white space’ area. Whilst most IT and electronic systems will work up to 40°C this can have a limiting effect on operational life leading to earlier than planned for failures and potential service interruption. Some devices such as uninterruptible power supplies may also have temperature sensitive assemblies (e.g. value regulated lead acid batteries) which can age more quickly in environments above 25°C.
Air conditioning systems installed within more critical applications tend to be monitored remotely via a building management system (BMS) or data centre infrastructure management (DCIM) suite. Should a fault condition arise the cooling system will generate an alarm and this can be transmitted along a suitable cable connection (signal contact or network IP address) to a management console interface. The air conditioner may also show a red warning light on its front panel. Once a fault is indicated, it is then important to identify the actual fault, cause, impact and rectification process. Simple faults may be cleared on-site but often it will be necessary to call-out an HVAC (heating ventilation and air conditioner) engineer. The call-out may be under warranty or outside warranty (chargeable) or covered by an on-site air conditioning system maintenance contract. Typical cooling system faults include:
A complete failure is unlikely but can occur. More likely the thermostat temperature setting has been altered either manually or remotely via the BMS or DCIM suite. A quick visual and manual check can suffice here. For thermostats and air conditioners connected to an IP network, there is also the potential for a breach in cybersecurity.
When designing cooling systems for server rooms or datacentres it is best practice to model the environment and install some form of hot-isle, cold-isle containment and air flow planning. Air vents can become blocked through the placing of equipment or packing in the wrong place; most commonly during works inside the white space area. Alternative, air flow could be disrupted by the moving of racks or removal of blanking plates from a server rack. Air conditioners also have air filters which can become blocked over time. A maintenance contract will normally include preventative maintenance visits during which consumables like filters are either cleaned or replaced. If your air conditioner is not under maintenance, we always recommend an assessment.
Air conditioners and cooling or chiller systems are powered from AC electricity sources and normally the mains power supply to the building. Whilst it is normal practice to not put air conditioners on the output of uninterruptible power supplies (to prevent over sizing of the UPS system and its batteries), air conditioners may be classed as essential loads and supported by a standby power generator. If the air conditioner is not working, check the electricity connection to both system components; air conditioner system within the building and the external heat-exchanger or condenser. The issue can be as simple as a tripped breaker either within the air conditioning system or a standby power generator or main or sub-distribution boards.
The internal element of your cooling system is the air conditioner unit we most commonly see hung from a ceiling or fitted to a wall. Larger systems will be floor mounted. The role of the unit is to absorb heat from the surrounding air using a set of coils. The external heat-exchange unit or condenser also has a set of coils which radiate the heat externally. If either set of coils is damaged or dirty then the air conditioning system efficiency will reduce and the system will fail to provide the thermostatically controlled temperature.
The cools contain a refrigerant which flows carrying cool fluid into the air conditioner and heated fluid to the heat-exchanger or condenser. If a coil cracks or there is a poor joint connection, fluid can leak leading to system inefficiency and incorrect temperatures.
When a datacentre building is designed appropriate space is provided as part of the site plan to provide provision for the heat-exchangers and condensers as well as other critical components. The condensers may be roof-top mounted, attached to the side of the building or located near to it. The point being that it is harder to block the condenser element of the air conditioning systems installed for cooling such buildings. For smaller installations and/or those fitted as retrofit space may be more of a premium. Condenser units can become blocked by a variety of materials and obstacles. Any performance related issue should be investigated to ensure heat can be released into the external environment.
Using a certified air conditioning/HVAC design, installation and maintenance company is one way to ensure you have ‘hassle free’ cooling for your server room or datacentre facility. Whilst HVAC companies tend to have a preferred list of brands and manufacturers they work with, their certified engineers may be able to provide service and support on third-party installed AC systems.
As with any critical system within a server room or datacentre environment maintenance cannot be ignored. Regular inspections as part of annual or bi-annual preventative maintenance visit will help to ensure that your systems continue to operate 24/7. A maintenance contract should also mean that manufacturer approved spares are used when consumables or other assembles require replacement. Certified engineers will also have access to the necessary documentation and software tools required including service checklists and maintenance procedures.
Should an alarm or failure occur there are some steps that on-site IT and facilities or engineering managers can take to prevent prolonged downtime and possible failure. Where a long-term fix cannot be applied, the initial option may well be to consider a complete unit or system replacement. This bring other advantages. Not only should this guarantee reliability but a new air conditioner system should be more energy efficient, more compact and reliable. The unit may also provide a greater scope of remote monitoring as many device manufacturers provide more enhanced and intelligent systems to connect to Cloud based reporting services and become managed via Internet of Things (IoT) technologies. New systems also tend to have new manufacturer warranties and this can give the ultimate peace of mind and assurance.
For more information on server room and datacentre ambient temperature guidelines visit ASHRAE: https://www.ashrae.org/.
The statement that our world is becoming more inter-connected is an understatement when you consider ‘The Internet of Things’ (IoT). In the world of HVAC and cooling systems it is easy to see that air conditioning systems are becoming more intelligent with a greater range of remote communications options which should lead to improvements in operating costs and overall energy efficiency.
Whether you are running an in-house enterprise server room or colocation datacentre, critical power and cooling are key environmental factors that must be managed to prevent IT system operating errors, erroneous alarm conditions, data processing errors, system malfunctions and unexpected shutdowns.