We will go over the four main UPS (uninterruptible power supply) system design configurations commonly used to distribute power for the utility source to the critical loads. As with any other engineering design, the UPS configuration chosen has trade-offs between reliability and costs. The selection of a UPS system for a particular application is determined by the clients needs, risk tolerance and budget. We will explore the four main configurations used in the industry today.

4 Main UPS Configurations Used Today

Ordered from 1 being least reliable and least expensive, to 4 being most reliable and most expensive.

  1. Single (Capacity)
  2. Isolated Redundant
  3. Parallel Redundant
  4. Distributed Redundant

In the diagrams below, the following shorthand notations are used:

PDU (Power Distribution Unit): This is usually the panelboard to which your critical loads are fed from.
STS (Static Transfer Switch): This a switch that is able to choose between two incoming sources on its line side, to feed the load connected to it.

Single Capacity UPS Configuration

singleups

A single or capacity system is the most common type of UPS system used in the industry. It is commonly refereed to as an “N” system. This is a system where a single UPS unit feeds the critical load. All UPS units come with a maintenance bypass switch which is normally open. In the event that any maintenance needs to be done on the UPS, or there is a complete failure of the UPS unit, the UPS can be bypassed and the load can be fed from utility. In addition to this bypass, there is also what is known as the static bypass switch located within the UPS itself, providing an additional means of feeding utility to the load. The difference between the two switches is that the maintenance bypass is completed external to the UPS, thus can be used when full isolation of the UPS is required. The static switch is automatic switch that automatically transfer the load to utility power if there is any problem with the UPS output. This configuration is the cheapest possible, however the lack of redundancy limits the critical load’s protection against failure of the UPS itself.

Isolated Redundant UPS Configuration

isolatedredundant

An isolated redundant UPS system is also refereed to as an N+1 redundant system. Unlike a parallel redundant system, an isolated redundant system does not require a paralleling bus, nor does it require the UPS units have the same capacity, or even the same model for that matter. Under normal operation, the primary UPS is fed from the utility and feeds the load, and the secondary/catcher/isolating UPS remains unloaded. In the condition where the utility power is lost, the secondary UPS will feed the primary UPS. This creates the redundancy that if one ups fails, the second ups can take on the load.

Parallel Redundant UPS Configuration

parallelredundant

In a parallel redundant system, also known as an N+1 system, two or more UPS systems share the critical load. Thus in the diagram shown above, in normal operation, each ups will share the load. UPS A will take 50% of the critical load and UPS B will take the other 50% of the critical load. However upon failure of any one of the UPS units, the other UPS unit must be able to take on the full 100% critical load.  The more UPS units you add in parallel, the better reliability you can achieve. This setup normally requires a UPS paralleling cabinet, with data communication connections between the UPS units, so each UPS knows the state of all other UPS units.

Distributed Redundant UPS Configuration

distributedredundant

Distributed redundant system is the most costly configuration to install, but also provides the most reliability.  In all other previous configurations, you can notice that all though there were multiple UPS to provide redundancy, any point of failure on the critical load bus and distribution panel to the critical loads would end up dumping the loads. In a distributed redundant system, the output of each ups is sent to is sent to static transfer switches which feed multiple power distribution panels, which share all the critical loads. This reduces the likelihood of a single failure dumping the entire load.

NO COMMENTS

Leave a Reply