What Is a Sequence of Operation in a BMS?

The recipe behind the building

Imagine handing a new kitchen assistant a shelf of ingredients and a row of appliances, but no recipe. They have everything they need to cook dinner, yet nothing useful will happen. They do not know what goes first, how long to wait, or when the dish is done.

A building full of chillers, pumps, fans and valves is in exactly this position. In the earlier parts of this course we met the individual control loops — a sensor reads a value, a controller compares it to a target, and an actuator nudges the system back toward that target. Each loop is competent on its own. But a chilled-water plant is not one loop; it is dozens, and left to act independently they would trip each other, waste energy, and occasionally destroy equipment.

The sequence of operation is the recipe. It is the document — written first in plain English, then translated into controller code — that says what runs when, in what order, and under what conditions. It is where a pile of capable loops becomes a plant that behaves as one system.

This is Part 6 of our Building Automation & BMS Fundamentals course. If loops are the vocabulary, the sequence of operation is the sentence.

"When X, do Y": the plain-English recipe

At its heart, a sequence of operation is a list of conditional statements: when X happens, do Y. Written well, a non-programmer can read it and understand exactly how the building will behave.

A fragment might read:

> When the building enters occupied mode and the return chilled-water temperature rises above setpoint for more than five minutes, enable the lead chiller. Before starting the chiller, prove that its condenser-water pump and cooling-tower fan are running.

Notice what that single paragraph contains. There is a trigger (occupied mode, temperature above target). There is a time condition (for more than five minutes). There is an order of operations (pump and fan before chiller). And there is an implied set of numbers — what is "setpoint," how long is the delay — that will be listed elsewhere in the document.

The reason this is written before any code is that the sequence is a specification, not a program. The building owner (or their consultant) writes down the intended behaviour; the automation contractor implements it. The document is the shared understanding between the two. When the plant does something surprising two years later, the sequence of operation is the reference everyone returns to: does the plant do the wrong thing, or the thing it was told to do?

Interlocks and safeties: order matters

Some parts of the recipe are not about comfort or efficiency at all. They exist to stop the plant hurting itself. These are interlocks — hard rules that permit or block an action depending on the state of other equipment.

The classic example, and one worth memorising, involves the chiller and its heat-rejection loop. A chiller works by absorbing heat from the chilled water that cools the building, then dumping that heat into a separate condenser-water loop, which carries it out to a cooling tower where it is released to the outside air. (We cover how chillers and cooling towers actually move heat in a later part; for now the direction of heat flow is what matters.)

Here is the danger. The moment a chiller starts, it begins rejecting heat into the condenser-water loop. If that water is not already circulating, and if the cooling-tower fan is not already throwing off heat, the pressure and temperature on the condenser side climb almost immediately. Within seconds the machine can trip on a high-pressure safety — or, if a safety fails, suffer real damage.

So the sequence enforces an order:

- First, start the condenser-water pump and prove that water is actually flowing (a flow switch or a pressure reading confirms it — a running pump motor is not proof that water moves).
- Then, start the cooling-tower fan.
- Only then, once both are proven, permit the chiller to start.

That word proven is important. An interlock does not trust that a command was obeyed; it waits for evidence. Commanding a pump to run and confirming the pump is running are two different things, and a good sequence never confuses them.

The same logic runs in reverse on shutdown. You do not stop the condenser-water pump the instant the chiller stops, because the machine is still full of hot refrigerant that needs somewhere to go. The sequence keeps the pump and fan running for a short period afterward to let the chiller settle. Order matters going down as well as coming up.

Staging: adding and shedding capacity

A building's cooling demand is never constant. A tropical office is nearly empty at 7 a.m., packed and sun-loaded by 2 p.m., and winding down by 7 p.m. Sizing a single chiller big enough for the 2 p.m. peak means that same machine spends most of the day running at a fraction of its capacity — and chillers, like most machines, are inefficient when badly loaded.

The answer is staging: installing several smaller chillers and bringing them online one at a time as load rises, then dropping them as load falls. The sequence of operation defines the rules.

- As cooling demand climbs and one chiller can no longer hold the chilled-water temperature at setpoint, the sequence stages up — it starts the next chiller (with its own pump-and-fan interlock, of course).
- As demand falls and the running machines are lightly loaded, the sequence stages down — it stops a chiller so the remaining ones each carry a healthier share.

The payoff is that each machine spends more of its life near an efficient load point rather than one machine running badly loaded. Staging also builds in resilience: if one chiller faults, the plant limps on with the others rather than losing all cooling at once.

The same idea applies to pumps and cooling-tower fans, which are also often installed in banks. And it is why variable-speed drives appear all over modern plants: instead of switching whole machines on and off in coarse steps, a drive can modulate a pump or fan smoothly to match demand. The drive itself is a piece of power electronics — if you want to understand how it varies a motor's speed, the Electricity Fundamentals course explains it in how variable-speed drives and inverters work and how electric motors work.

A quick but important distinction while we are here. When the sequence "starts a pump," what it directly commands is often a small actuator or a drive — a low-voltage device that positions a valve, opens a damper, or tells a drive to spin up. The thing that actually moves the water is a three-phase motor running at 400/230 V, or on the largest centrifugal chillers a medium-voltage motor. The BMS commands; the motor does the muscle work. Keeping those two straight saves a lot of confusion on site.

Modes: the building changes character through the day

A plant does not behave the same way at 3 a.m. as it does at midday, and the sequence captures this through modes — named states, each with its own rules.

Occupied mode is the daytime norm: the plant holds tight comfort conditions because people are in the building. Chilled water is at full setpoint, ventilation is bringing in fresh air, and staging responds actively to load.

Unoccupied mode applies when the building is empty. There is no reason to spend energy holding a packed-building temperature in an empty one. So the sequence allows the cooling setpoint to drift upward — this is called unoccupied setback, and in a cooling climate like Malaysia's it means letting the space get a few degrees warmer than it would be during the day, not colder. Some equipment shuts down entirely. The plant idles.

Startup mode is the bridge between the two. If occupancy begins at 8 a.m., you do not want to start cooling a warm building at 8 a.m. and make everyone uncomfortable until 10. So the sequence uses optimum start: it calculates how long the plant needs to pull the building down to comfort and begins pre-cooling early — perhaps at 6:30 on a hot morning, later on a mild one. The building reaches setpoint just as the first people arrive, and not a moment sooner, so no cooling is wasted on an empty floor.

The reason modes matter is that the same sensor reading should produce different actions depending on context. A space at 26 degrees is a problem in occupied mode and completely fine in unoccupied mode. Without modes, you would need a different sequence for every time of day. With them, you write the rules once per mode and let the clock and the occupancy schedule switch between them.

A word on lighting and occupancy

Because we have been talking about "the whole building behaving as one system," it is worth being precise about scope. A BMS sequence of operation is overwhelmingly about the mechanical plant — chillers, pumps, fans, air handlers, the things that move heat and air. Lighting and occupancy control often live in a related but separate layer, sometimes a dedicated lighting-control system that the BMS merely coordinates with. An occupancy sensor might tell the BMS that a zone is empty so it can relax that zone's cooling, and the same signal might switch lights off — but the two systems are usually distinct products with their own sequences. This course stays focused on the mechanical side; just know that "building automation" in the fullest sense reaches wider than the chilled-water plant we keep using as our example.

The tunable knobs: setpoints, deadbands and delays

Inside every sequence sit the numbers that make it work — and these are deliberately adjustable, because no design is perfect on paper and every building behaves a little differently once it is real.

- Setpoints are the targets: chilled-water supply at a chosen temperature, a space at a chosen comfort level. Change the setpoint and you change what the whole loop aims for.
- Deadbands are the tolerance around a setpoint within which the system does nothing — the buffer that stops equipment chattering on and off around the target. We defined the deadband in Part 4; here the point is simply that its width is a knob the commissioning engineer tunes, not a fixed law.
- Time delays are the waits built into the recipe: the five minutes a temperature must stay high before staging up, the minutes a condenser pump keeps running after a chiller stops, the delay before an alarm is raised. Delays stop the plant reacting to every brief wobble.

Getting these numbers right is most of what commissioning and later optimisation actually involve. Tighten a deadband too far and equipment short-cycles; loosen it too much and comfort drifts. Stage up too eagerly and you run machines you did not need; stage up too slowly and the building overheats before help arrives. The sequence defines the structure; the tuning makes it liveable.

There is an energy angle here too. On buildings billed under a maximum-demand tariff — as most medium and large commercial buildings in Malaysia are, at the RP4 demand charge of RM89.27–97.06 per kW effective 1 July 2025 — the sequence can be written to avoid starting several large machines at the same instant, smoothing the demand peak. You can see how that peak translates into ringgit with our maximum-demand calculator. Studies cite plant-wide savings from well-tuned sequences, but the figures vary and are genuinely site-specific, so treat any single percentage with caution.

The sequence is a contract

Here is the part that matters most in practice, and the reason experienced building owners obsess over this document. A sequence of operation is not just a technical artefact — it is effectively the contract between the owner and the automation contractor.

The contractor is paid to make the building behave the way the sequence describes. If the sequence is clear — every mode defined, every interlock spelled out, every setpoint and delay listed — then "done" is unambiguous, and both sides know when the job is finished. If the sequence is vague — "the chillers shall stage as required," with no rule for what "as required" means — then every dispute about how the plant behaves becomes an argument with no reference to settle it. Vague sequences cause most controls disputes, and they are entirely avoidable.

Writing and implementing a genuinely clear sequence of operation is a defining part of building-automation work — it is one of the core deliverables of Cobler's Automation Services, where the plain-English intent and the coded reality are made to match. A good sequence is worth more than any single clever control trick, because it is the thing everyone can point to and agree on.

The takeaway

A sequence of operation is the written recipe that turns a collection of independent control loops into a building that behaves as one system. It defines the "when X, do Y" logic, enforces the interlocks and safeties that protect equipment, stages capacity up and down to keep machines efficient, switches the plant between occupied, unoccupied and startup modes, and exposes the setpoints, deadbands and delays that engineers tune to make it all liveable. Above all, written clearly, it is the contract that keeps owners and contractors honest.

Next, in Part 7, we look at how a BMS actively manages a building's electrical demand — and why timing the plant's biggest loads is where controls meet the electricity bill.