BMS Scheduling and Optimum Start: How a Building Wakes Itself Up

The building that switches itself on

Imagine arriving at the office at 8 a.m. and finding it already cool, comfortable and quiet — even though nobody was there at 6 a.m. to flip anything on. Something switched the cooling on for you, chose the right moment to do it, and will switch it back off after the last person leaves tonight. That "something" is the scheduling logic inside a Building Management System (BMS) — the same brain that, in earlier parts, we watched read sensors and command valves and dampers.

So far in this course the BMS has been controlling things: temperature, flow, pressure. This part is about the BMS controlling time. It turns out that deciding when the plant runs is one of the cheapest and biggest savings a building can get — no new hardware, just better instructions. In a tropical climate, where there is no heating season to worry about, cooling plant left to its own devices would happily run 24 hours a day, seven days a week. Scheduling is how you stop paying for cooling that nobody is around to enjoy.

Let's build it up piece by piece.

Time schedules: telling the building its working hours

The most basic layer is a time schedule — a calendar the BMS follows that says, for each zone, when it should be occupied (cool it to the comfort setpoint) and unoccupied (ease off). A typical office schedule might read: occupied 07:00 to 19:00 Monday to Friday, unoccupied all other times.

The important word is per zone. A retail podium, a Level 20 office floor and a 24-hour data room have completely different hours, so each gets its own schedule. The ground-floor lobby might stay cool longer than the upper floors; the gym might come alive at 6 a.m. while the offices are still dark.

A good schedule also carries two extra calendars layered on top:

- A weekend calendar, so Saturdays and Sundays follow reduced or zero occupied hours.
- A public-holiday calendar, so the plant doesn't fire up on Hari Raya, Deepavali, Chinese New Year or a National Day when the building is empty. In Malaysia, with our generous spread of public holidays, forgetting this single calendar can mean cooling an empty tower a dozen-plus full days a year.

The elegance is that a schedule is just a set of instructions the BMS obeys automatically. Set it once, maintain the holiday list once a year, and the building stops working weekends without anyone touching a switch.

Why this matters so much in the tropics

In a temperate country a building has an off-season — a stretch of the year when neither cooling nor much heating is needed. Malaysia has no such luxury. Every single day is a cooling day. That means an unscheduled building doesn't just waste a little energy at the margins; it runs its chillers, pumps and fans against the heat around the clock, all year. The plant never gets a natural break.

That is exactly why scheduling is described as free savings. You are not making the plant more efficient — you are simply not running it when there is no reason to. Every hour of avoided run-time is energy you never buy and wear you never put on the machines.

Unoccupied setback: letting the setpoint drift up

Turning cooling fully off overnight sounds tempting, but it has a downside in our climate: a building that sits with no cooling for twelve hot, humid hours can get uncomfortably warm and damp inside, and pulling it back to a comfortable state the next morning takes a big, expensive burst of effort. Damp is the real enemy — humidity creeps into a building that isn't being conditioned.

So instead of switching off completely, many buildings use unoccupied setback — and here we have to be careful with words. In heating climates the equivalent trick is called "night setback," meaning you let the building get colder overnight to save on heating. In Malaysia we never heat anything, so the term doesn't apply. Our version is the exact opposite direction: we let the cooling setpoint drift up when the building is empty.

Concretely, during occupied hours the setpoint might be 24 degC. During unoccupied hours the BMS sets it up to, say, 27–28 degC. The cooling doesn't stop — it just relaxes. The plant only kicks in if the building gets warmer than 28 degC, which keeps humidity and heat in check without paying to hold a crisp, comfortable temperature for empty rooms. Because it is a set-up (raising the target), not a heating-style setback, calling it "night setback" here would plant exactly the wrong mental picture. It is unoccupied setback / setup, and the direction is up.

This gentle relaxing of the target is the unoccupied-hours partner to the smarter idea we'll look at next.

Optimum start: starting as late as you can get away with

Here is the clever part. A fixed schedule that says "start cooling at 07:00 every day" is better than nothing, but it is blunt. Why? Because 07:00 is a guess — someone picked an hour early enough to be safe on the worst day. On most days, the building doesn't need that much lead time, and every extra minute of pre-cooling an empty building is wasted money.

Optimum start fixes this. Instead of a fixed clock time, the BMS learns how long this particular building actually takes to cool from its warm, unoccupied state down to the comfort setpoint. Then it starts the plant only as late as it can while still hitting setpoint by the time people arrive.

The key idea to hold onto: optimum start uses the building's measured pre-cool response, not a fixed clock. Over days and weeks, the controller watches how the space responds — "yesterday it took 55 minutes to go from 28 degC down to 24 degC" — and builds up a picture of the building's thermal behaviour. Then each morning it looks at how warm the building currently is and how humid the day is shaping up, and works backwards from the occupancy time:

- On a hot, muggy morning, the building is warmer and slower to cool, so optimum start fires the plant earlier — maybe 06:10 to be ready by 07:00.
- On a cooler, rainy morning, the building barely warmed overnight, so it starts later — maybe 06:40 for the same 07:00 target.

The building, in effect, gives itself exactly the head start it needs and not a minute more. It is the difference between a blunt alarm clock and someone who checks the traffic before deciding when to leave for work.

Under the hood this is the same feedback control thinking we met earlier in the course — the controller compares where the building is against where it needs to be and adjusts. If you want the deeper foundation on how a controller closes that loop, the Electricity Fundamentals series covers it in feedback control and PID explained.

Optimum stop: coasting on the building's own mass

The mirror image of optimum start is optimum stop. Near the end of the occupied day, the BMS eases off the cooling before closing time, because the building will coast on its own stored coolness for a while.

Think of a large concrete-and-glass building as a cool battery. All day, the structure, the floors and the furniture have been chilled. When you stop actively cooling at, say, 18:30 for a 19:00 close, the space doesn't instantly turn warm — it drifts up slowly, riding on all that stored coolness in its thermal mass (the building's ability to soak up and hold coolness, like a cold brick that stays cold for a while after you stop chilling it). By stopping early, you get the last half-hour of comfort essentially for free and shave run-time off every single day.

Optimum stop then hands over to the unoccupied setback we described above: cooling relaxes, the setpoint sets up to 27–28 degC, and the building idles gently through the night until tomorrow's optimum start wakes it again. Notice that none of this is "night setback" in the heating sense — the whole cycle only ever raises the cooling target when the building empties out.

Occupancy-based control: don't cool an empty room

Schedules assume a floor is either "in use" or "not in use" by the clock. Real buildings are messier — a meeting room booked till 19:00 but vacated at 16:00, a whole wing empty because half the staff are travelling. Occupancy-based control lets the BMS respond to who is actually there, not just the timetable.

It does this by borrowing signals the building already produces:

- Motion sensors (the same kind that switch lights on and off) tell the BMS whether a zone has anyone moving in it.
- Card-access data from the door system tells it how many people badged onto a floor this morning — a rough headcount.

If a zone reads empty for a sustained period, the BMS can set its cooling up early, close down its air supply, or skip cooling it entirely — without waiting for the scheduled end of day. On the flip side, if someone badges in on a Saturday, the system can bring just their zone to comfort rather than the whole tower. This is scheduling that reacts to reality, and it plugs the biggest leak in any fixed timetable: rooms that are technically "occupied hours" but actually deserted.

The demand connection: don't let everything start at once

There is one more reason the timing of start-up matters, and it points straight at your electricity bill.

Recall the distinction from the Electricity Fundamentals series between energy (kWh) — how much electricity you use over time — and demand (kW) — how hard you are drawing at any single instant. (If that split is fuzzy, power vs energy: kW and kWh explained is the primer.) Scheduling saves energy by cutting run-hours. But careless scheduling can quietly inflate demand.

Here's how. If the BMS starts every chiller, every pump and every air handler at the exact same instant — 06:00 sharp — the building draws a huge, simultaneous gulp of power as all those big three-phase motors spin up together. That coincident burst sets a peak demand far higher than the building's steady running load. And on a demand-metered tariff — the medium-voltage commercial and industrial tariffs most larger buildings are on — TNB bills that peak. Under the RP4 structure effective 1 July 2025, the maximum-demand charge runs at roughly RM89.27–97.06 per kW. A needless spike of a few hundred kW at start-up, caused by nothing more than bad timing, can cost real money every month.

The fix is staggered start-up: the BMS brings plant online in sequence — one chiller, let its pumps stabilise, then the next, then the air handlers — so the loads don't all pile onto the same instant. Same building, same comfort, lower billed peak. You can see how a coincident start-up moves the billed number using Cobler's maximum-demand calculator.

We're only touching demand here; a later part of this course is devoted to demand limiting — how a BMS actively caps peak kW during the day — and this staggered-start idea is its first taste.

Where Cobler fits

Getting schedules, optimum start and staggered sequencing right is squarely the job of the control system itself — designing and implementing that logic is what Cobler's Automation Services do. It is exactly the kind of "free savings" that comes from better instructions rather than new machinery, and it lives or dies on how well the sequences are written and commissioned.

The takeaway

A BMS doesn't just control temperature — it controls time. Time schedules stop the plant running when nobody's in. Unoccupied setback relaxes the cooling by letting the setpoint drift up (to 27–28 degC) when the building is empty — a set-up, never a heating "night setback." Optimum start learns the building's own measured cooling response and gives it exactly the head start it needs, earlier on hot days and later on mild ones. Optimum stop lets the building coast on its stored coolness. And occupancy sensing cools only the zones with people in them. In a tropical climate with no off-season, all of this is close to free money — you're simply not paying to cool an empty building around the clock. The one catch is start-up timing: fire everything at once and you spike the demand TNB bills you for, which is why the plant is brought online in sequence.

Next, we look at how all these controllers actually talk to each other — the communication protocols like BACnet and Modbus that let a mixed bag of equipment act as one coordinated system.