Chiller plant control: sequencing, setpoint reset and low delta-T

Chillers are often the biggest single electrical load in a commercial building, so how the BMS runs them shows up directly on the electricity bill. The good news is that most of the waste comes from a handful of well-known problems.

Let the chiller look after itself

As with boilers, there’s a clear split. The chiller’s own controller runs the refrigeration side: compressors, capacity control, expansion, condenser fans on air-cooled units, and all the safeties such as high and low pressure, freeze protection and loss of flow. The BMS decides when cooling’s needed, which chillers and pumps run, and what chilled water temperature to ask for.

The interface usually comes down to:

a hardwired enable, plus run and fault status back
a setpoint reset signal, either 0–10 V or over comms
a BACnet or Modbus gateway for everything else: temperatures, alarms, run hours

When you’re reading chiller data over Modbus, double-check the scaling and data types. A lot of supposed chiller faults turn out to be mapping errors. There’s a guide on Modbus register offsets and word order if you need it.

Flow and return temperatures

Comfort cooling has traditionally been designed around something like 6 °C flow and 12 °C return. Plenty of newer systems run warmer, 10/16 °C or more for chilled beams, which also helps chiller efficiency. Whatever the design, keep an eye on the gap between flow and return, the delta-T. It matters more than most people realise, as you’ll see below.

Starting and stopping

A typical sequence:

There’s a demand for cooling. A cooling valve or AHU calls for it, or it’s warm enough outside during occupied hours.
The chilled water pump starts, and flow is proven.
On water-cooled plant, the condenser water pump and the cooling tower or dry cooler fans start as needed.
The lead chiller is enabled, and it controls its own capacity to hit the setpoint.

Stopping is the reverse, with a pump run-on so the chiller can unload safely.

Chillers have anti-recycle timers that limit how often the compressors can start. The BMS should respect minimum run and off times too, rather than flicking the enable on and off with every change in load.

More than one chiller

You’d normally bring on another chiller when the chilled water flow temperature has sat above setpoint by a margin for a while with the running chillers close to full load, or when the calculated load (flow × delta-T) is more than they can handle. Drop one off when the running chillers are lightly loaded and one fewer could cope. Rotate the lead by run hours, and skip any chiller that’s in fault.

Setpoint reset

Asking for warmer chilled water when the load allows makes chillers more efficient, because the compressors have less work to do. The usual rule of thumb is a few per cent improvement for every degree, although the real figure depends on the machine.

There are two common ways to do it. Outside air reset simply raises the setpoint in milder weather. Valve position reset keeps raising the setpoint until the most-open cooling valve in the building is almost fully open. That second method is usually the better one, because it’s based on what the building actually needs.

Be careful anywhere humidity matters, though. Warmer chilled water takes less moisture out of the air.

Primary, secondary and the decoupler

A lot of plants have primary pumps that keep a constant flow through the chillers and secondary pumps that serve the building. The two circuits are joined by a short pipe called the decoupler, or common pipe.

If the primary flow is bigger than the secondary flow, the extra chilled water goes back through the decoupler to the chillers. The chillers then see cooler return water than the building actually sent back. If it’s the other way round, warm return water comes up the decoupler into the supply, and the building gets warmer chilled water than the chillers are making.

Watching which way the decoupler is flowing, or the temperatures either side of it, helps the BMS judge when to bring on or drop another chiller and pump.

Low delta-T syndrome

This is the classic chilled water headache. The return water comes back much cooler than the design says it should, maybe 9 °C instead of 12 °C. The chillers can’t load up properly, so you end up running more chillers and more pumps than the real load needs, each one at poor efficiency.

The usual causes:

3-port valves or bypasses on terminal units, sending unused chilled water straight back
oversized or badly controlled 2-port valves letting too much water through the coils
fouled coils or dirty filters that can’t transfer heat properly
space setpoints so low that valves sit wide open all day
units with no control valve at all

The BMS is the best tool you’ve got for finding it. Trend the plant delta-T alongside the cooling valve positions across the building, and look for units sitting wide open with hardly any temperature drop across them.

Free cooling

When it’s cold enough outside, some plants can make chilled water without running compressors, using dry coolers or a free cooling coil built into the chiller. The BMS generally enables free cooling once the outside temperature is a few degrees below the chilled water return temperature, and blends it with mechanical cooling as conditions change.

The usual faults

chillers short-cycling because the BMS toggles the enable on small load changes
two chillers running at low load when one would do
setpoint reset switched off, so chilled water is always at its design minimum
flow switch trips from blocked strainers or pumps starting late
an AHU heating and cooling at the same time in winter
mis-scaled gateway points, so the BMS makes decisions based on nonsense temperatures

The chillers are only half the story. The air side matters just as much, and that’s covered in how an AHU control strategy works.