Now has entered the era of highly developed information, and information centers have become an important part of people's daily lives. The scale of the data center has become larger and larger, and the requirements for reliability and life have become higher and higher. Only in this way can the data center maintain a strong support for the network. The power supply system as the heart of the data center is the first element that supports the data center and the network. Its working state directly affects the reliability of the data center and the network.

Energy saving is the primary indicator of data centers. Energy saving is to reduce power consumption. Only by reducing power consumption can equipment be maintained at low temperature, which can improve reliability and extend life. This is the purpose of reducing PUE. According to Arrhenius's law, every 10 ° C rise in electronic equipment (including batteries) halves its life. In other words, increasing the temperature by a gradient of 10 ° C will shorten the life of the equipment by 1 / 2n. Currently , a large number of electrolytic capacitors and liquid lead-acid batteries are used in UPS systems.They are the major factors that seriously affect the life of equipment and systems. Locally, there is an urgent need to change this status quo. The emergence of fuel cells once gave people hope, but due to some objective reasons, it has not yet been applied to UPS.

The emergence of lithium-ion batteries has also opened the way for the reliability of UPS power supply systems, and this battery has also been successfully used in electric vehicles: the battery's power supply range is 48 to 500V, covering the entire UPS application range. The discharge current is larger than the UPS, and the environment is worse than the UPS, because the car is not always in a constant temperature and vibration-free environment like the UPS. In principle, it can completely replace liquid lead-acid batteries, but for some reasons, it is prohibitive for users. Experiments are only being done in individual places.

① The advent of spiral-wound lead-acid batteries rejuvenates lead-acid batteries

Previously, it was thought that lead-acid batteries could only have a liquid flow structure, and coupled with environmental pollution during production and use, there were restrictions on the manufacture and use of such equipment. The appearance of spiral wound lead-acid batteries has broadened people's horizons and opened the door to hope for improving the reliability of UPS power supply systems. Because according to relevant data statistics, more than 70% of the failure of the UPS power supply system comes from the battery. The actual service life of lead-acid batteries with a design life of five years is two to two and a half years on average. The spirally wound structure is completely solid, as shown in Figure 2. Its 100% charge and discharge reaches 350 times, which is more than twice the flow battery's 150 times; 30% charge and discharge reaches 4,000 times, which is more than 3 times the flow battery's 1200 times. The operating temperature is very wide, -55 ～ + 70 ℃; Of course, it can also be placed in any position, which is very practical in military and field unattended situations; its leakage current is very small, but the discharge current is very large, which can reach 3 to 5C; it can reach 95% of its capacity after 40 minutes of charging .

The current capacity of this battery is 75Ah. And it is easy to assemble.

② Film capacitors replace electrolytic capacitors

The development of film capacitor technology has made its capacity as large as that of electrolytic capacitors, but because it is dry, it has functions that electrolytic capacitors do not have. The voltage range of film capacitors is much larger than that of electrolytic capacitors, the internal resistance is also much smaller, the stray inductance is an order of magnitude smaller than that of electrolytic capacitors, and the life is much longer than electrolytic capacitors at the same price, which greatly expands film capacitors. Application areas. Spiral-wound lead-acid batteries and film capacitors provide a means to improve the reliability and extend the life of UPS power systems.

(2) The current mode of power supply and distribution is characterized by large concentration and large dispersion

The power supply mode of the UPS system is determined by the scale structure of the data center. In large data centers, due to the requirements of large data concentration, there are too many devices. At least 200,000 servers are working in the data center. The power supply system in the case must also be centralized. The advent of micro-module data centers and container data centers has led to the decentralized configuration of the original mode of centralized power supply.

(3) Build concentration with decentralization

The scale of contemporary data centers is getting larger and larger, from a few hundred square meters to tens of thousands of square meters. But no matter how big, most of the equipment in a data center with a floor area of 200 square meters or more can not be in one step. Generally, about 150 racks with a capacity of 400 to 500kVA are used as a "module" to install and debug first. This module proved to be correct during operation, and the rest of the computer room can be copied and installed according to this module, which inspired the idea of micro-modules. In fact, this is only one aspect, and the integration of data center infrastructure and physical facilities is also the second reason for the micro-module structure. The advent of micro-modules and container-type data centers has greatly increased the construction speed of the computer room. The former can be invested while building, and the latter eliminates the need to build buildings and machine room decoration. The idea of micro-module data centers and container data centers has led to the progress of modular power supply systems. The data center composed of micro-modules and container groups is an example of decentralized composition and concentration.

2 Types of current UPS power supply systems and development of power distribution systems

(1) Two types of UPS power supply system

The development direction of AC output UPS is high frequency. At present, it is in the era of high frequency UPS replacing industrial frequency UPS. At present, the main types of UPS power supply systems: rotary generator flywheel energy storage type and static conversion type. The increase in flywheel energy storage flywheel speed and backup time has made it available for trial in some data centers. However, due to the relatively large capacity of such UPSs, most of which are above 150kVA and the volume is also large, and more than 70% of the current more than 500,000 data center computer rooms are below 200 square meters, which also limits its popularity. Static conversion UPS has been widely used in data centers. This type of UPS has evolved from fully thyristorization in the 1960s to full IGBT now, and the efficiency has also increased from less than 90% of full load in the past to more than 95% of half load now. Taking 100kVA as an example, the high-frequency UPS can save more than 50,000kWh of electricity each year than the industrial frequency UPS. Economic and social benefits are quite significant, providing favorable conditions for achieving low PUE in data centers. The capacity of the static conversion UPS is not limited in the small capacity range, and can be as small as tens of watts. In the large capacity range, a single machine can currently achieve 1600kVA. Compared with industrial frequency machine UPS, it is more energy efficient and material saving.For example, an imported 300kVA industrial frequency machine UPS weighs 2200kg, while the domestic 300kVA high frequency tower machine weighs only about 360kg, which reduces 1.8 tons! The N + X modular structure is more efficient for data centers Construction and operation provide convenient conditions.

(2) Cancel the transformer and column cabinet

In the power distribution system, some transformers are currently installed in the original cabinets, in order to "isolate interference". In fact, this kind of transformer does not isolate the interference at all, but has brought many adverse consequences:

① Two more points of failure in series. The input switch of the transformer and the transformer itself are fault points.

② Increased power consumption, reduced reliability, and degraded PUE. The addition of a transformer reduces the system efficiency by at least two percentage points.

③ Increased weight, wasted materials, and increased the burden on the floor. Originally, the floor load can meet the requirements, but the floor must be reinforced because the transformer does not meet the requirements. The project volume and investment have been increased and the construction period has been delayed.

④ Increased investment. No doubt spending money on failure.

The current new technology is not only eliminating the transformer, but also eliminating the head cabinet, adopting the rail busway power distribution method, and has had successful experience in many places at home and abroad. Figure 12 shows a comparison between the power distribution method of the headstock and the rail-type busway power distribution method. The realization of this change will undoubtedly bring benefits to all forms of data centers. Now some building data centers and container data centers have adopted this scheme and achieved good results. It will gradually spread in the future.

3 Analysis of the misunderstanding of the data center's design of the power supply and distribution system

(1) It is considered that general power transformers have the ability to isolate interference, so the power transformer in the equipment such as UPS is not added to the column cabinet. The reason is that the transformer is in a linear state during normal operation, and the linear state does not distort the waveform. In other words, the basic requirement for the transformer is that the input and output waveforms are not distorted, that is, the input waveforms must look exactly the same as the output. This has been discussed many times and practical examples in related articles, and I will not repeat them here.

(2) Think that the generator and UPS can give the power calculated by the power factor under any load

This misunderstanding is more common, so that leaks occur in many places, machine trips and damaged devices occur repeatedly. In fact, the reason is very simple, as long as the output capacity of the power supply and the load do not match, it must be derated. As for how much it is down, it does not match the degree. For example, a power supply with a load power factor of 0.8 can only provide 53% of the rated value (that is, active power) when the load has an input power factor of 1. The generation source is already fully loaded. For example, a data center of a government agency purchased a UPS with a load power factor of 0.8, which repeatedly burned inverter power tubes at a load rate of 70%. Generators are no exception. For example, when a generator of a rural commercial bank in a certain province has a linear load inspection machine, 70% of the load motor trips. An industrial park in a certain province is also a 70% load motor overload alarm when the generator is inspected with a linear load. Therefore, when the backup generator is equipped with a high-frequency UPS, its power ratio is not 1: 1, but 2: 1. If the overload capacity and charging power are considered, it must be at least 2.5: 1.

(3) equip a battery pack with 32 12V batteries with AC and DC circuit breakers

The backup battery pack of the UPS must be equipped with a short circuit or fuse, but must be equipped with a DC circuit breaker. However, DC circuit breakers are much more loaded than AC circuit breakers in terms of materials and mechanical structure, so they are more expensive. So the battery pack is equipped with so-called "AC and DC circuit breakers" in many places. Because DC circuit breakers are rare, although arc extinguishing is easier at 48V DC voltage, it is very difficult at 240V high voltage. Therefore, the price of DC circuit breakers is several times that of AC circuit breakers of the same capacity, so AC circuit breakers are mostly used in many places. The application of AC circuit breakers or DC circuit breakers has very different effects.

Can DC circuit breakers be replaced by AC circuit breakers? It depends on the current level. It can certainly be used in the case of several amps to a dozen amps. When the current reaches several hundred amperes, the arc between the fractures will only extinguish when the current crosses zero, and there is no zero crossing of the DC current, so the arc cannot be extinguished. At this time, a special DC circuit breaker must be used. If the arc lasts more than 20ms, it can be called an explosion.

What are the characteristics of a DC circuit breaker?

It is relatively difficult for an AC circuit breaker to interrupt a DC short-circuit current. Because the condition of arc breaking is that the breaking arc voltage is greater than the power supply voltage. A major difference between DC and AC is that there is no natural zero crossing of the voltage, so DC arc breaking is more difficult. Therefore, DC circuit breakers require special arc blowing coils or use permanent magnet arc blowing technology to force the DC arc into the quenching chamber. The arc is cut, stretched, the arc voltage rises and cools down quickly, and it has been assumed that the polarity of the DC line is strictly paid attention to. In fact, AC circuit breakers have not completely adopted similar technical measures in design. The reliability, durability and breaking capacity when used on the road are far inferior to DC circuit breakers.

AC circuit breakers have a limited DC voltage value that can be disconnected by a single pole. AC circuit breakers with a nominal voltage of 230V AC may not necessarily be directly used in DC 220V lines.In GB10963-2, the rated voltage is 230V. When a single-pole miniature AC circuit breaker is used in a DC system, the DC power supply voltage should generally not exceed 220V, and 2-pole series should be considered for use from a safety perspective, and a lot of engineering practice experience should be considered.

In the above proposal, when the DC voltage is 250V, 4 contacts are required in series. In view of this, in some cases, it is more economical to continue to use AC circuit breakers than to choose DC circuit breakers directly. But the so-called AC and DC circuit breakers currently used are actually AC circuit breakers. Because in the AC use, who would spend more money to buy AC and DC products! What is more terrible is that in high-power UPS generally use 32 batteries, the rated voltage is nearly 400V, still use the two contacts of the AC circuit breaker in series There is a danger that it will not be able to be disconnected at that time.

LastThe inside story of the UPS poNextGlobal uninterruptible power s