Smart power conditioning

Smart power conditioning: need of the hour

With exponentially rising data volumes, processing power and the correspondingly colossal power requirements, the need for power conditioning has never been quite so crucial for data centers, writes Varun Aggarwal

We have heard quite a lot about greening of data centers, reducing power consumption, optimizing cooling etc and discussions on this topic have been there for a long time. However, what we haven’t talked about and perhaps overlooked is how ‘power quality’ has become a critical issue for Data Centers (DCs). In today’s economy, we depend on technology and other electric equipment more than ever. At the same time, this technology is more threatened than ever by increasing power problems and fluctuations.

While not limited to safeguarding any particular type of equipment, a UPS is typically used to protect computers, telecommunication equipment or other electrical equipment where an unexpected power disruption could cause injuries, fatalities, serious business disruption or worst data loss. UPS units come in sizes ranging from units which will back up a single computer without monitor (around 200 VA) to units which will power an entire DC or buildings (several megawatts).

A UPS is one of the most critical components in a DC. The servers that are hosted in a DC handle many online applications such as an ERP system, ATMs, online trading, mail servers, Web sites, etc. To handle such criticality, a company cannot depend on the standard electrical power supplied by power generating companies. Servers cannot work without power for more than 20 milliseconds. Once a server goes off, it takes time for it to start functioning again. During this period, the connection between the client and the organization is lost further affecting the business of a company. Today, due to an extremely competitive market environment, availability has become a key factor in the success of any organization. A reliable UPS keeps a DC up and running continuously even when there is a power failure so that the customers stay connected. In case of mains power failure, they have battery back-up to support the servers till a diesel generator kicks-in.

Subodh Tagare, Marketing Director, APC-MGE said, “It is important to know how to avoid potential power pitfalls that can have a serious impact on the function of your DC and the integrity of your data. A common misconception is that utility power is consistent and steady. It isn’t. No matter where you are, spikes, surges, brownouts and other power problems pose potential hazards to your network.”

He continued, “With any power fluctuation, you have the potential for data damage, data loss, file corruption and damage to hardware, which is not always immediately noticeable.”

As high technology products are upgraded, you grow more dependent on AC power to run them. Utility power is not enough.

UPS for the DC

DCs require high uptime and thus the technologies deployed in one ought to be different from those used in other environments. The on-line or on-line double conversion UPS systems with multi-level redundancy capabilities are best suited for DC environments. Redundant design is the key. In addition to providing protection against complete failure of utility supply, it provides protection against all common power problems.

Suhas Joshi, Country Manager-UPS Product Management, Emerson Network Power (India) Pvt Ltd., said, “It is an international practice to have a mirror image of the complete power system to run a DC in order to avoid shut-downs in case of power failure or for maintenance purposes. DCs go in for two redundant power supplies from different power companies to enable a consistent flow and keep the DC up and running. Online maintenance is critical since for the smooth functioning of the DC, shutdowns are an absolute no-no. Even planned shutdowns (for e.g. for maintenance) cannot be considered. Availability along with maintainability has to be ensured.”

So what is the way forward? The answer would lay in an intelligent UPS that can send alerts and automatically adjust to cope with some degree of fluctuations and outages in the power supply.

Having an intelligent UPS power distribution unit (PDU) with Isolation transformers is a topology that is gaining popularity. This Intelligent PDU helps in tracking the power consumption of each and every rack present in a UPS. It also gives a proactive alarm in case of excess power consumption by a rack. These features help an organization to take corrective measures and avoid unplanned downtime. They are also useful for a DC which hosts the servers of its customers. These customers are unaware of the type of server and power consumed in a DC. The k-rated transformers used in these PDUs solve the problem of lowest earth to neutral voltage requirements of server vendors. They also handle harmonic issues.

Under normal operation an online UPS is always running off the battery. This design means that there is no transfer time in the event of a power failure—if the power goes out, the inverter (and its load) keeps chugging along and only the battery charger fails.

Power availability is one of the most important challenges facing DCs today. In the past, DC floor space loomed as the primary issue. Nowadays, DCs run out of power availability before they run out of floor space. In addition, cooling requirements for dense servers are driving power demand and taxing normal DC operational procedures.

For years, electrical power usage was not considered as a key design criteria for DCs. Nor was electrical consumption effectively managed as an expense. In fact, many DC managers were unaware of what their monthly energy bill was. “This is true despite the fact that the electrical energy costs over the life of a DC may exceed the costs of the electrical power system including the UPS, or even exceed the cost of the IT equipment itself. If the DC were 100% efficient, all power supplied would reach the IT load. This would represent Power Usage Effectiveness (PUE) of 1.0,” said Tagare.

There is another key advantage to having equipment running off the battery most of the time: the double-conversion process totally isolates the output power from the input power. Any nasty surprises coming from the wall will affect only the battery charger, and not the output loads.

Static and dynamic UPS

Static UPS is a widely used technology in the world. The power electronic components used in these systems are reliable. It provides long battery back-up and an option of using different types of batteries depending upon the application. These compact systems also help in avoiding major noise pollution. No mechanical parts are used in the system, hence no wear and tear takes place. An organization can run these applications 24x365 without any concern.

A static UPS system usually consists of a battery to provide continuous source of electrical power; a rectifier/charger to maintain battery charge and to provide input to an inverter when utility power is available; an inverter to provide power to load during normal operation; a static switch to transfer load automatically and without disturbance between inverter and utility power; a manual switch to bypass the static switch for maintenance; input and output isolation transformers and filters to provide appropriate isolation and disturbance attenuation; and monitors, sensors, and control circuits.

On the other hand a dynamic or rotary UPS is used for applications requiring ride-through of short duration power system outages, voltage dips, etc. The dynamic UPS typically does not include batteries, and support times are usually on the order of a few seconds to a few minutes.

Tagare explained that some dynamic UPS systems consist of an AC Motor Generator (M-G) set with a flywheel, as well as a rectifier, storage batteries, inverter, static switch and solid state circuitry. Both the motor generator set and the rectifier/battery/inverter combination are supplied by the incoming utility service. They represent parallel supply paths and either path is capable of supplying the load. A static switch selects the path to be utilized to supply the load. During normal operation, the M-G set powers the loads, while the off-line static section is on ‘stand-by’ and charges the system batteries. Upon loss of the utility feed, the control circuitry will disconnect the M-G set from the utility by opening the static switch and closing the inverter-output circuit breakers, allowing the system batteries to power the M-G set through the inverter. Mechanical energy ‘stored’ in the flywheel allows the M-G set to continue to deliver its full-rated output for a minimum of 200 milliseconds. This provides sufficient time for the control system to sense a loss of utility supply and to transfer to the battery/inverter combination for supply of loads.

Data processing equipment rooms will typically overheat within a 15 to 30 minute period if the ventilation system is not working; making the generator set a near necessity for outages in excess of this time. Analysis of battery cost will often justify a generator set at lower cost than choosing a long battery support time with an accompanying restriction of eventually implementing an orderly critical load shutdown. A battery support time of as little as a few minutes may be specified with generator set backup; however, longer support times in the range of 15 minutes are more typical.

“All rotating machines need a high degree of maintenance and it is extremely expensive to maintain a dynamic system compared to the static UPS,” said Tagare. As an example, operational temperatures are a critical factor in maintaining integrity of the bearings and the lubricant. Lubricant life can be halved if the bearings are running at just 20 degree centigrade higher than the designed temperature. The rotary UPS with no built-in redundancy also has an exceptionally long MTTR and the replacement of bearings can take as long as 24-48 hours leaving no real protection for the load.

Joshi added, “These systems are prone to wear and tear as they have rotating parts built into the system. The systems are bigger than the static UPS systems and need to have trained manpower to manage mechanical wear and tear in the system. Noise pollution created by Rotary UPS is high when compared to static UPS. It requires a company to maintain the stock of diesel and follow all the safety standards for the storage of diesel. The efficiency of the rotary system is lower (85%) as compared to Static UPS (93-95%) as the former uses mechanical parts. An organization has to spend extra while buying the system as well as while using it.”

The initial investment in a rotary system is higher by 30-40% which is the main reason for them not being popular amongst the consumers. Since static UPS’ use fast switching (2 kHz/sec) PWM technology to generate sine wave, they have a high degree of flexibility and compatibility compared to the motor alternator using fixed winding to generate sine wave. In addition, a rotary system will generate a much higher noise level (82dBA 800KVA) compared to a static UPS which, at the same rating, will not exceed 72dBA.

Since the system efficiency is very low (around 84%), running costs can be exceptionally high when compared to similar static UPS. The output of static UPS’ generates a fairly clean sine wave voltage to support the non-linear load with a minimum of voltage distortion. Each UPS system should have an external static switch bypass to provide additional support in the event of internal fault or external downstream major short circuit. In order to provide the necessary redundancy, availability and resilience, parallel redundant configuration is recommended. In some DC, design has been set out for isolated redundant configuration.

Emerging UPS technologies

Newer technologies are amicable with the Indian power conditions which are very diverse in different pockets of the country. While the metros have a good power supply, the interiors reflect a different picture altogether. As one moves towards the interiors, one will realize that there are frequent power cuts on a daily basis, unclean power and other basic constraints like earthing, cabling and lack of a viable infrastructure that compound the problem. New technologies in the UPS segment are aimed at either increasing efficiency or decreasing cost or both at the same time.

For example, a UPS feature called Load Bus Synchronization (LBS) keeps the output of two or more independent systems in synchronization, even when they are operating from different power sources. Any UPS will typically synchronize to its bypass source. As long as the UPS systems are tied to the same input and bypass sources, they will automatically stay in synchronization. Take for example a Digital Signal Processor (DSP) based UPS. UPS were traditionally designed as analog circuitry but a DSP based UPS offers sophisticated control algorithms with highly flexible software, the ability to add user interfaces, reduce components, introduce testing procedures, and increase reliability. Hence a DSP based UPS is a digital UPS which offers higher performance for a lower cost.

Another upcoming technology is the Fuel cell UPS that has been developed in recent years using hydrogen and a fuel cell as a power source, potentially providing long run times in a small space. A fuel cell is a possible replacement for the batteries used in UPS designs.

Fuel cells are different from batteries in that they consume reactant, which must be replenished, whereas batteries store electrical energy chemically in a closed system. Additionally, while the electrodes within a battery react and change as a battery is charged or discharged, a fuel cell’s electrodes are catalytic and relatively stable.

The efficiency of a fuel cell is dependent on the amount of power drawn from it. Drawing more power means drawing more current, this increases the losses in the fuel cell. This however is still in the development stage and there is more to do, before the fuel cell replaces the battery in UPS.

varun.aggarwal@expressindia.com