Chilled water plants must reconcile two competing needs. Chillers prefer stable flow through their evaporators, within manufacturer limits. Buildings need variable flow, because coils use less water at part load. The pumping arrangement is how the design reconciles these needs.
The two most common arrangements in modern plants are primary-secondary (also called decoupled) and variable primary flow. Both can perform well. This guide explains how each works, its advantages and pitfalls, and how to choose.
Primary-secondary systems
In a primary-secondary system, there are two sets of pumps:
- Primary pumps circulate a constant (or stepped) flow through each operating chiller. Usually one primary pump is dedicated to each chiller.
- Secondary pumps are variable speed pumps that distribute chilled water to the building coils, modulating to maintain differential pressure.
The two loops connect through a decoupler, a short, low-resistance common pipe. When primary flow exceeds secondary flow, the excess flows through the decoupler from supply to return. When secondary flow exceeds primary flow, warm return water flows backward through the decoupler and mixes with chiller supply water, raising supply temperature to the building. That is a sign that another chiller should be started.
Advantages
- Simple, robust chiller flow control: each chiller sees constant design flow.
- Chiller and distribution pumps can be selected independently, which is helpful when chillers are far from loads or in campus systems with tertiary loops.
- Well understood by operators and designers.
- Tolerant of controls problems: chillers are protected even if distribution controls misbehave.
Disadvantages
- More pumps, more space and more first cost.
- Primary pumps run at constant speed, consuming energy even at part load.
- Low delta T in the building causes secondary flow to exceed primary flow, forcing more chillers and primary pumps to run than the load requires. See low delta T syndrome.
- Reverse flow in the decoupler raises supply temperature, reducing dehumidification.
Some designs include a check valve in the decoupler to prevent reverse flow, but this effectively puts primary and secondary pumps in series and can cause other problems. It should only be done with care.
Variable primary flow systems
In a variable primary flow (VPF) system, one set of variable speed pumps circulates water through both the chillers and the building. Flow through the chillers varies with building demand, within limits set by the chiller manufacturer.
Key components:
- Variable speed primary pumps, typically headered so any pump can serve any chiller.
- Isolation valves on each chiller, often motorized, to open and close as chillers are staged.
- Flow measurement at each chiller, or at the plant, to control minimum flow.
- Minimum flow bypass, with a modulating control valve, that opens to maintain minimum chiller flow when building flow drops below it.
Advantages
- Fewer pumps, less space and lower first cost.
- Lower pump energy, since all pumping is variable speed and the separate constant primary pump energy is eliminated.
- Chillers can be loaded more fully, because more flow can be pushed through operating chillers when building delta T is low, within the chiller's maximum flow limit.
- Higher return water temperature to chillers in many operating conditions.
Disadvantages
- More complex controls, particularly during chiller staging.
- Chillers must tolerate variable flow and the rate of flow change. Manufacturers specify minimum and maximum flow and acceptable rates of change, often expressed as a percentage of design flow per minute.
- Accurate flow measurement is essential; poor measurement leads to nuisance trips.
- Staging a chiller on suddenly diverts flow from operating chillers, which can cause trips if not managed with slow-opening isolation valves and coordinated pump control.
- Operators need training to understand the system.
Side-by-side comparison
| Feature | Primary-secondary | Variable primary flow |
|---|---|---|
| Number of pump sets | Two | One |
| Chiller flow | Constant or stepped | Variable within limits |
| First cost | Higher | Lower |
| Pump energy | Higher | Lower |
| Controls complexity | Lower | Higher |
| Sensitivity to low delta T | High: extra chillers run | Moderate: can overflow chillers within limits |
| Chiller compatibility | Any chiller | Chillers rated for variable flow |
| Operator familiarity | High | Moderate |
Minimum flow bypass in VPF
The bypass protects chillers when building flow is very low. Design points:
- Locate the bypass in the plant, close to the chillers and flow meters, where it can respond quickly to changes in chiller flow. This is the most common arrangement.
- Size the bypass valve for the minimum flow of the largest chiller.
- Control the valve from measured chiller flow, opening only when flow approaches the minimum.
- Keep the bypass closed during normal operation. A bypass that opens frequently suggests a setpoint or measurement problem.
Bypassing chilled water directly to return lowers return temperature and wastes pump energy, so minimizing bypass operation matters.
Chiller staging differences
In primary-secondary systems, chillers are typically staged on when decoupler flow reverses or when supply temperature rises. In VPF systems, staging is commonly based on chiller load (percent of rated capacity or current), supply temperature and flow. Because flow through a chiller can exceed design when delta T is low, VPF can sometimes defer starting another chiller, saving energy. See chiller plant sequencing and optimization.
Other arrangements
- Primary-only constant flow with three-way valves: an older approach that wastes pumping energy.
- Primary-secondary-tertiary: used in campuses and district systems, where buildings have their own pumps connected through a bridge or decoupler.
- Distributed pumping: small pumps at each load rather than central secondary pumps, used in some campus and data centre systems.
How to choose
Primary-secondary may be the better choice when:
- Existing chillers are not rated for variable flow.
- Operating staff prefer a simpler, robust system.
- The plant serves multiple buildings or distant loads with very different pressure needs.
Variable primary flow is often the better choice when:
- New chillers rated for variable flow are being installed.
- Pump energy and first cost are priorities.
- Controls design, commissioning and operator training will be done well.
Converting existing primary-secondary plants to VPF is possible and can save energy, but it requires confirming chiller flow limits, adding flow measurement and isolation valves, and rewriting sequences. Thorough commissioning and water balancing follow.
Frequently asked questions
Is variable primary flow always more efficient?
It usually reduces pump energy, but overall plant savings depend on controls quality. A poorly controlled VPF plant with frequent bypass or chiller trips can perform worse than a well-run primary-secondary plant.
Can heating hot water systems use variable primary flow?
Yes. Many modern condensing boilers allow variable flow within limits, and the same principles apply.
Key takeaways
- Primary-secondary uses constant chiller flow and variable distribution flow, decoupled by a common pipe.
- Variable primary flow uses one set of variable speed pumps through chillers and building.
- VPF saves pump energy and first cost but needs good flow measurement, chiller compatibility and careful controls.
- Low delta T hurts primary-secondary plants more by forcing extra chillers online.
- The right choice depends on chiller type, operator capability and the system served.
Talk to an engineer
Have a project or a problem this guide touches on? Tell us about the building and we’ll be in touch within one hour.