Parallel Capacity VS Parallel Redundancy

Selecting the Right UPS Architecture for Critical Loads

Executive Overview

Connecting multiple UPS units in parallel can serve two different purposes:

  • Parallel Capacity combines the output of several UPS units to support a load that exceeds the capacity of one unit.
  • Parallel Redundancy provides additional UPS capacity so the critical load can continue operating after one UPS unit is lost.

Parallel operation alone does not guarantee redundancy. The correct classification depends on the capacity that remains after a failure and on the design of the complete power path, including batteries, bypass systems, switchgear, cables, and load distribution.

Understanding N

N represents the minimum UPS capacity required to support the complete critical load.

For example, if a critical load requires:

  • 100 kW
  • 125 kVA

The UPS system must support both values without exceeding its kW, kVA, output-current, or power-factor limits.

N can be provided by:

  • One large UPS
  • Several UPS units operating in parallel
  • Several power modules inside a modular UPS

The number of UPS units does not determine redundancy by itself. The important question is whether sufficient capacity remains after one unit is unavailable.

Parallel Capacity

What Is Parallel Capacity?

In a Parallel Capacity system, two or more UPS units operate together to increase the total available power.

The combined capacity is required to support the load, with no complete spare unit available.

Example

Critical load:

100 kW

Installed UPS units:

  • UPS 1: 50 kW
  • UPS 2: 50 kW

Total installed capacity:

50 + 50 = 100 kW

This is a Parallel Capacity system and is classified as an N architecture.

If one UPS fails, only 50 kW remains available, which is not enough to support the complete 100 kW load.

Depending on the UPS design and operating conditions, the result may be:

  • Overload of the remaining UPS
  • Transfer to bypass
  • Load shedding
  • Complete load interruption

When Is Parallel Capacity Used?

Parallel Capacity may be selected when:

  • The load exceeds the rating of one UPS unit
  • Larger UPS ratings are unavailable or impractical
  • Future capacity expansion is expected
  • Load sharing between several units is preferred
  • Full redundancy is not required

Parallel Capacity increases available power but does not preserve full load capacity after one UPS unit is lost.

Parallel Redundancy

What Is Parallel Redundancy?

In a Parallel Redundant system, additional UPS capacity is installed above the capacity required by the load.

After one UPS unit becomes unavailable, the remaining units must still be capable of supporting the complete critical load.

This arrangement is commonly used to achieve N+1 redundancy.

Example

Critical load:

100 kW

Installed UPS units:

  • UPS 1: 50 kW
  • UPS 2: 50 kW
  • UPS 3: 50 kW

Two UPS units are required to support the load:

N = 2 × 50 = 100 kW

The third UPS provides the additional capacity:

N+1 = 150 kW installed capacity

If one UPS is lost, the remaining two units still provide 100 kW and continue supporting the complete load.

N+1 Requirement

For identical UPS units, N+1 is achieved when:

Capacity remaining after one unit is lost ≥ Critical load

Or:

(Number of units − 1) × Unit capacity ≥ Critical load

If the UPS units have different ratings, the calculation should be based on the loss of the largest unit.

The assessment must consider:

  • kW capacity
  • kVA capacity
  • Output current
  • Load power factor
  • Static bypass capacity
  • Overload capability

Parallel Capacity VS Parallel Redundancy

Parallel Capacity

  • Main objective: increase total power
  • Installed capacity is approximately equal to the load requirement
  • No complete spare capacity is available
  • Loss of one unit reduces capacity below the load requirement
  • Normally classified as N
  • Lower initial cost

Parallel Redundancy

  • Main objective: improve availability
  • Additional capacity is installed above the load requirement
  • The load remains supported after one unit is lost
  • Normally classified as N+1
  • Can support some maintenance activities without losing protected power
  • Requires additional equipment, space, and cost

Core Principle

Parallel describes how the UPS units are connected. Redundancy describes the capacity remaining after a failure.

Redundancy Depends on the Actual Load

The redundancy level can change as the load increases.

Example

Four UPS units are installed, each rated at 40 kW.

Total installed capacity:

4 × 40 = 160 kW

At a 120 kW Load

After one unit is lost:

3 × 40 = 120 kW

The system still supports the load and operates as N+1.

At a 140 kW Load

After one unit is lost, only 120 kW remains available.

The system is no longer N+1 at this load, even though the number of UPS units has not changed.

Redundancy must always be evaluated against the actual and future critical load.

Load Sharing in Parallel UPS Systems

Parallel UPS units must share the load evenly and operate in synchronization.

This function is called Load Sharing.

For example, if three identical UPS units support a 90 kW load, each unit should carry approximately 30 kW, within the manufacturer’s permitted tolerance.

Poor load sharing may result from:

  • Unequal output-cable lengths
  • Different conductor sizes
  • Unequal cable impedance
  • Incorrect parallel settings
  • Communication failure
  • Firmware mismatch
  • Incorrect installation
  • Faults in current-measurement circuits

Unequal load sharing may overload one UPS and reduce or eliminate the intended redundancy.

Requirements for Parallel Operation

UPS units must be specifically designed and approved for parallel operation.

The system should support:

  • Output synchronization
  • Automatic load sharing
  • Parallel communication
  • Fault isolation
  • Coordinated bypass operation
  • Approved unit isolation and reconnection procedures

The design must also verify:

  • Maximum permitted number of parallel units
  • Model and rating compatibility
  • Firmware compatibility
  • Power-cable requirements
  • Parallel communication-cable requirements
  • Unit addressing
  • Start-up and shutdown sequence
  • Manufacturer-approved parallel settings

UPS units must not be connected in parallel based only on matching voltage and frequency ratings.

N+1 Does Not Mean Full System Redundancy

A UPS system may have N+1 capacity while still containing shared components that can interrupt the entire load.

These components are known as Single Points of Failure.

Examples include:

  • One main input breaker
  • A common input switchboard
  • A common output bus
  • One output cable
  • A common static bypass
  • A common maintenance bypass
  • A shared battery bank
  • One central controller
  • A common distribution board
  • A shared cooling or communication system

The complete power path must therefore be evaluated, not only the UPS power modules.

Capacity Redundancy and Path Redundancy

Capacity Redundancy

Capacity Redundancy means that additional UPS capacity is available to compensate for the loss of one unit.

Example:

Parallel Redundant N+1 system

Path Redundancy

Path Redundancy means that more than one independent power path is available to the critical load.

Example:

2N architecture

A system can provide Capacity Redundancy without providing Path Redundancy.

Understanding 2N Architecture

A 2N architecture provides two independent power systems, and each system is capable of supporting 100% of the critical load.

Example

Critical load:

100 kW

Power Path A

  • UPS System A: 100 kW
  • Battery System A
  • Input and output switchgear A
  • Bypass arrangement A
  • Distribution path A

Power Path B

  • UPS System B: 100 kW
  • Battery System B
  • Input and output switchgear B
  • Bypass arrangement B
  • Distribution path B

Each path independently provides the full N capacity.

The total installed UPS capacity is therefore:

2 × N

The purpose is not to supply twice the load. The purpose is to allow one complete path to be isolated or lost while the other path continues supporting the load.

2N Is Not Simply Two UPS Units

Installing two UPS units does not automatically create a 2N system.

For a true 2N design, common components should be minimized or eliminated wherever practical.

Shared components may include:

  • Input supplies
  • Battery systems
  • Static bypass circuits
  • Maintenance bypass circuits
  • Output switchboards
  • Distribution cables
  • Control systems
  • Load-distribution equipment

If both UPS units depend on one shared output bus, battery system, or bypass path, the design may not provide complete 2N independence.

Parallel Redundancy VS 2N

Parallel Redundancy

  • Multiple UPS units operate within one power path
  • Additional capacity compensates for the loss of one unit
  • Normally provides N+1 capacity
  • May include shared batteries, bypass, or distribution equipment
  • Protects mainly against loss of UPS capacity

2N Architecture

  • Two independent power paths are provided
  • Each path supports 100% of the critical load
  • One complete path can be isolated or lost
  • Shared components are minimized
  • Protects against the loss of an entire power path

N+1 provides spare capacity within a system.
2N provides a second complete power path.

Dual-Cord and Single-Cord Loads

Dual-Cord Loads

Many critical servers and network devices include two independent power inputs.

  • Input A connects to Power Path A
  • Input B connects to Power Path B

If one path fails, the equipment continues operating from the other path.

The equipment manufacturer’s requirements must be reviewed to confirm that either input can support the required load.

Single-Cord Loads

Equipment with one power input may require a:

  • Static Transfer Switch
  • Automatic Transfer Switch
  • Redundant Power Distribution Unit

However, the transfer device may itself become a Single Point of Failure and must be included in the reliability assessment.

Batteries and Redundancy

Shared Battery Bank

A shared battery bank may reduce:

  • Cost
  • Space
  • Number of breakers
  • Number of cables

However, it may create a common failure point.

A fault in the shared battery breaker, bus, cable, or protection system may affect all connected UPS units.

Separate Battery Banks

Separate batteries provide better isolation between UPS units or power paths.

They allow one battery system to be maintained without directly affecting the other.

Separate battery systems are generally more suitable for a true 2N design, but they require greater cost, space, monitoring, and maintenance.

Static and Maintenance Bypass

Static Bypass

The Static Bypass transfers the load to an alternative AC source during conditions such as:

  • UPS overload
  • Internal inverter fault
  • Overtemperature
  • Other manufacturer-defined operating conditions

The design must confirm:

  • Full-load bypass capacity
  • Overload withstand capability
  • Bypass-source availability
  • Whether the bypass is common or independent
  • Whether it creates a Single Point of Failure

Maintenance Bypass

A Maintenance Bypass allows the UPS equipment to be isolated for service while the load remains energized from another source.

However, while the load is on Maintenance Bypass, it may no longer receive the power-conditioning and backup protection normally provided by the UPS.

A Maintenance Bypass improves maintainability but does not automatically provide redundancy.

Modular UPS Redundancy

A Modular UPS simplifies N+1 design because the total capacity is divided between replaceable power modules.

Example

Critical load:

120 kW

Power-module rating:

40 kW

Required modules:

  • 3 modules = N
  • 4 modules = N+1

If one module fails, the remaining three modules continue supporting the 120 kW load.

However, other internal components must also be reviewed, including:

  • System controller
  • Static bypass
  • Input and output busbars
  • Cooling fans
  • Communication system
  • Battery charger
  • Frame capacity

A spare power module cannot compensate for the failure of a non-redundant central component.

Future Load Growth

The design must consider both present and future loads.

Example

Three UPS units are installed, each rated at 50 kW.

Total installed capacity:

150 kW

At a 100 kW load:

  • Two units support the load
  • One unit provides redundancy
  • The system operates as N+1

If the load increases to 130 kW:

  • After one unit is lost, only 100 kW remains
  • The system no longer provides N+1 redundancy

The maximum load that preserves the required redundancy must be defined during design.

Selecting the Correct Architecture

Select Parallel Capacity When

  • The main objective is to increase power capacity
  • Full operation after one unit failure is not required
  • A short interruption or bypass transfer is acceptable
  • Project cost is a major constraint
  • Future modular expansion is required

Select Parallel Redundancy When

  • The load must remain operational after one UPS unit fails
  • A balance between cost and availability is required
  • The UPS supports approved parallel redundant operation
  • Some shared system components are acceptable
  • Maintenance must be completed with minimal interruption risk

Select 2N When

  • Interruption of the critical load is unacceptable
  • The system must tolerate loss of one complete power path
  • Planned maintenance must not expose the load to one remaining shared path
  • The critical equipment supports dual power inputs
  • The financial or operational cost of downtime justifies the additional investment

Common Design Mistakes

  • Assuming every parallel UPS system is redundant
  • Confusing Parallel Capacity with Parallel Redundancy
  • Calculating total capacity without calculating remaining capacity after a failure
  • Checking kW but ignoring kVA
  • Ignoring future load growth
  • Failing to verify load sharing
  • Ignoring Static Bypass capacity
  • Using a shared battery without evaluating the risk
  • Calling two UPS units a 2N system despite shared components
  • Ignoring downstream switchgear and distribution
  • Failing to identify Single Points of Failure
  • Failing to test the loss of one UPS unit during commissioning

Final Engineering Checklist

Before approving the system design, confirm:

  • The critical load has been measured in both kW and kVA.
  • The required N capacity has been defined.
  • The purpose of parallel operation is clearly identified.
  • Remaining capacity after loss of the largest unit has been calculated.
  • UPS overload limits have been reviewed.
  • Load-sharing performance has been verified.
  • UPS models and firmware versions are compatible.
  • Parallel communication and control settings are correct.
  • Static Bypass capacity is sufficient.
  • Maintenance Bypass operation is understood.
  • Shared and independent battery options have been evaluated.
  • All Single Points of Failure have been identified.
  • Input and output switchgear has been reviewed.
  • Downstream power distribution has been included in the assessment.
  • Future load growth has been considered.
  • Loss of one UPS unit has been tested.
  • Transfer to bypass has been tested.
  • Unit isolation and return-to-service procedures have been tested.
  • All settings and commissioning results have been documented.

Technical Conclusion

Parallel Capacity and Parallel Redundancy use similar parallel UPS connections, but they serve different objectives.

Parallel Capacity combines several UPS units to support a load that exceeds the rating of one unit. If one unit is lost, the remaining capacity may be insufficient.

Parallel Redundancy adds spare capacity so the remaining UPS units can continue supporting the complete critical load after one unit becomes unavailable.

A 2N architecture goes further by providing two independent power paths, each capable of supporting the full critical load.

The final reliability level depends not only on UPS capacity, but also on:

  • Batteries
  • Static and Maintenance Bypass systems
  • Switchgear
  • Cables
  • Control systems
  • Load distribution
  • Common failure points
  • Maintenance procedures

Parallel Capacity increases available power.
Parallel Redundancy preserves the required power after a UPS unit is lost.
A 2N architecture preserves the load after an entire power path is lost.

Related Posts