A brief educational program for beginners about why modular UPSs are cooler and how it happened.
Uninterruptible power supplies for data centers are divided into two large groups according to the construction architecture: monoblock and modular. The former belong to the traditional type of UPS, the latter are relatively new and more advanced.
What is the difference between monoblock and modular UPS
In monoblock uninterruptible power supplies, the output power is provided by one power unit. In modular UPS, the main components are made in the form of separate modules that are placed in unified cabinets and work together. Each of these modules is equipped with a control processor, a charger, an inverter, a rectifier and is a full-fledged power part of the UPS.
Let's explain this with a simple example. If we take two uninterruptible power supplies - monoblock and modular - with a capacity of 40 kVA, then the first will have one power module with a capacity of 40 kVA, and the second will consist, for example, of four power modules with a capacity of 10 kVA each.
Zoom options
When using monoblock UPSs with an increase in power demand, it is necessary to connect another full-fledged unit of the same power in parallel to the existing one. This is a rather complicated process.
Modular solutions are characterized by greater design flexibility. In this case, one or more modules can be connected to an already functioning unit. This is a fairly simple procedure that can be completed in a short time.
Possibilities of smooth increase in power
A smooth increase in capacity is important at the initial stage of data center operation. It is quite logical that in the first months it will be loaded by 30-40%. It is more practical and economical to use uninterruptible power supplies designed specifically for this power. As the customer base grows, the data center load will increase, and with it, the need for additional power supply will increase.
It is convenient to increase the power of the UPS in stages along with the technical infrastructure. When using monoblock uninterruptible power supplies, a smooth increase in power is impossible in principle. With modular UPSs, it is easy to implement.
UPS Reliability
Speaking of reliability, we will operate with two concepts: the mean time between failures (MTBF) and the mean time to restore the system (MTTR).
MTBF is a probabilistic value. The value of mean time between failures is based on the following postulate: the reliability of a system decreases with an increase in the number of its components.
According to this parameter, monoblock UPSs have an advantage. The reason is simple: Modular UPSs have more components and plug-in connections, each considered as a potential point of failure. Accordingly, theoretically, the possibility of failure is higher here.
However, for uninterruptible power supplies used in the data center, it is not the failure itself that matters, but how long the UPS will remain inoperative. This setting is determined by the Mean System Recovery Time (MTTR).
Here the advantage is already on the side of modular blocks. They have a low MTTR because any module can be quickly replaced without power interruption. This requires that this module be in stock, and its dismantling and installation can be performed by one specialist. In fact, it takes no more than 30 minutes.
With monoblock uninterruptible power supplies, the situation is much more complicated. They cannot be repaired so quickly. This may take from several hours to several days.
To determine the fault tolerance of the system, one more parameter can be used - availability or otherwise operability. This indicator is higher, the longer the mean time between failures (MTBF) and the lower the mean time to restore the system (MTTR). The corresponding formula looks like this:
average availability (uptime) =
With regard to modular UPSs, the situation is as follows: their MTBF value is less than that of monoblock ones, but at the same time they have a significantly lower MTTR indicator. As a result, the performance of modular uninterruptible power supplies is higher.
Energy consumption
A monoblock system requires significantly more energy because it is redundant. Let us explain this with an example for the N+1 redundancy scheme. N is the amount of load required for the operation of data center equipment. In our case, we will take it equal to 90 kVA. The N+1 scheme means that 1 redundant element remains unused in the system before the failure.
When using a monoblock 90 kVA uninterruptible power supply, to implement the N + 1 scheme, you will need to use another of the same unit. This results in a total system redundancy of 90 kVA.
When using 30 kVA modular UPSs, the situation is different. With the same load, to implement the N + 1 circuit, one more of the same module will be needed. As a result, the total redundancy of the system will no longer be 90 kVA, but only 30 kVA.
Hence the conclusion: the use of modular power supplies can reduce the power consumption of the data center as a whole.
Economy
If you take two uninterruptible power supplies of the same power, then a monoblock one is cheaper than a modular one. For this reason, monobloc UPSs remain popular. However, increasing the output power will double the cost of the system, because another of the same unit will have to be added to the existing one. In addition, there will be a need to install patch panels and switchboards, as well as laying new cable lines.
When using modular uninterruptible power supplies, the power of the system can be increased smoothly. This means that you will have to spend money on purchasing such a number of modules, which are enough to meet the existing need for power supply. No unnecessary stock.
Conclusion
Monoblock uninterruptible power supplies are inexpensive, easy to set up and operate. At the same time, they increase the energy consumption of the data center and are difficult to scale. Such systems are convenient and effective where small capacities are required and their expansion is not expected.
Modular UPSs are characterized by easy scalability, low recovery time, high reliability and availability. Such systems are optimal for increasing the capacity of the data center to any limits at minimal cost.
Source: habr.com