IoT Sensors in a Building — How Continuous Monitoring Replaces Breakdown Maintenance

Most buildings in Israel are "silent." A pump runs until it burns out, an HVAC motor vibrates for months until the bearing seizes, a leak creeps behind a wall until a stain appears on the ceiling of the floor below. In every one of these cases the building knew about the problem long before we did — it simply had no mouth to speak. IoT sensors are that mouth: they turn a system discovered only when it breaks into a system that reports on itself in real time, and allow the shift from "breakdown maintenance" to proactive maintenance that treats a fault while it is still a small sign.

What IoT Actually Means in the Context of a Building

IoT (Internet of Things) sounds like a buzzword, but the idea is simple: a small sensor that measures some physical quantity — temperature, moisture, vibration, current draw, pressure, occupancy count — and sends its reading over a wireless network to a central place that collects, stores, and alerts. The sensor is not "smart" on its own; the intelligence is that instead of a person physically going once a month to measure, the measurement happens every few seconds, automatically, around the clock.

It is important to distinguish IoT from a classic building management system (BMS). A BMS controls: it turns chillers on and off, opens and closes dampers, manages HVAC and lighting logic. IoT, by contrast, mostly monitors — a layer of sensors that can be installed even in an old building without touching the existing infrastructure, and provides a continuous status picture from it. The two complement each other: in a modern building the BMS and IoT talk to each other; in an old building, IoT is often the only cheap and fast way to start measuring at all. I expanded on this difference in the guide to BMS building management systems.

The Real Problem: The "Wait Until It Breaks" Culture

Many buildings in Israel act on a system only when something forces them — a regulator demands, an inspector arrives, or — the worst version — an event happens. Until then the pump "works," the elevator "runs," the electricity "flows," and there is no reason to touch anything.

From my experience as a building manager, breakdown maintenance is always more expensive than preventive maintenance in three dimensions: the secondary damage (a seized bearing burns out a whole motor, a small leak ruins a ceiling and floor), the inconvenient timing (it always happens on a Friday afternoon or a holiday), and the legal exposure — liability that could have been prevented entirely.

Continuous monitoring breaks this pattern because it provides the missing "forcing factor" — not a regulator and not an event, but data. When you see that a pump's current draw has gradually risen over two weeks, you do not need anyone to force you; the data itself is the reason to act. This is exactly the principle of planned preventive maintenance, on which Israeli Standard (SI) 1525 is built — only here the data arrives in real time instead of in a periodic inspection.

What IoT Sensors Really Measure in a Building

There is no single "IoT sensor" — there is a family of sensors, each catching a different type of failure. These are the most common and useful in an office building:

• Vibration: installed on motors, pumps, fans, and chillers. A gradual rise in vibration amplitude is the earliest sign of bearing wear, imbalance, or looseness — sometimes months before the actual failure. In my experience, the first alert arrives when the bearing starts to drag at night, when there is no background noise to mask it.
• Thermal temperature on panels and terminals: a loose connection in an electrical panel heats up before it burns — sometimes weeks in advance. A temperature sensor (or a periodic thermal scan) catches the "hot spot" in time, before it becomes a fire.
• Moisture and leaks: leak-detection strips on the floors of machine rooms, under boilers, near water reservoirs, and at pipe crossings. They alert on the first drop, not on a puddle. In a building where only protective bunds are laid — it takes weeks until someone sees it.
• Electricity consumption (current monitoring): a current meter on a single circuit. A change in a motor's consumption profile betrays an abnormal load, a dropped phase, or wear — before the motor gives out. A 10-15% rise in consumption sustained over a few days is a clear red flag.
• Air quality (CO2, particles, humidity): a direct measure of HVAC function and occupant comfort, and also a tool for detecting ventilation problems or moisture leading to mold. CO2 climbing during working hours indicates a failure in air supply — before anyone feels "less well" and complains.
• Water pressure and filter differential pressure: a pressure drop betrays a leak or a blockage. A pressure differential across an HVAC filter tells you exactly when to replace it — not too early and not too late. This way you replace by condition, not by schedule.
• Open/close and presence: machine-room doors, access to sensitive areas, occupancy counting to calibrate HVAC and lighting by real occupancy. An empty building on a Friday afternoon does not need to run like a full building.

Note the pattern: almost every sensor here catches a trend, not an event. The value is not in the single reading but in its change over time — and that is exactly what a person with a clipboard once a month will never see.

Three Maintenance Approaches — and Why Only the Last Closes the Gap

To understand why IoT is a step-change, it helps to understand the sequence:

• Breakdown maintenance: you fix it when it breaks. The cheapest on the day of purchase, the most expensive over time.
• Schedule-based maintenance: you replace a filter every three months, service a pump every six months — regardless of the actual condition. A huge step up, and it is the basis of every annual preventive maintenance checklist. But it has waste: you replace a perfectly good component, and sometimes a component wears out faster than expected and breaks between two service dates.
• Condition-based maintenance: you neither fix it when it breaks nor replace it by an arbitrary schedule — you act when the data indicates real deterioration. The sensor says "this bearing has started to deteriorate" — and then, and only then, you schedule a replacement, at a convenient time, with a spare part ready, before the failure. This is the most mature form of maintenance, and it requires no guessing — it requires measurement.

In the building I manage, the shift to condition-based maintenance on the HVAC motors changed the frequency of urgent calls: when you catch deterioration three weeks in advance, the technician arrives by appointment, not on an hour's notice.

Where IoT Pays Back Fastest — By System

Not every system needs monitoring, and not every monitoring is worth the investment. Here is where IoT pays back fastest in a typical office building:

• HVAC: the most expensive and sensitive system in the building. Vibration monitoring on chillers and fans, differential pressure on filters, and supply temperatures catch expensive failures early. A chiller failure struck in summer is a major expense — constant monitoring significantly lowers the probability of it. More on the system in HVAC maintenance in office buildings.
• Electricity and panels: monitoring panel temperature and current on critical circuits catches loose connections and overloads — the two main sources of electrical fires in buildings. I expanded in electrical system maintenance.
• Plumbing and water: leak sensors in machine rooms and under critical piping prevent the most expensive damage — water seeping for weeks without anyone seeing. More in water system maintenance.
• Elevators: vibration monitoring, motor temperature, and trip counting betray wear before a fault and complement — do not replace — the legally required inspection by a licensed inspector.

The Big Mistake: Buying Sensors Without Building a Process

This is where most buildings that do try fall down. They buy a crate of sensors, install them, open a nice dashboard with graphs — and then nothing happens. The reason is simple: a sensor without a process is noise. It produces alerts that no one reads, or worse — produces so many false alerts that people get used to ignoring all of them, and then the real alert is lost too.

A sensor produces value only when it is connected to a full chain:

Measurement → threshold → alert → service call → documentation → loop closure

That is: when the reading crosses a predefined threshold, a service call is automatically opened with the appropriate party, someone responsible handles it, the treatment is documented, and the loop is closed with confirmation that the problem was resolved. Without the last links, the sensor is a toy. With the full chain, it becomes a worker who never sleeps.

A point I learned in practice: calibrating the threshold takes time. In the first weeks, almost any threshold set too early is either too high (misses) or too low (floods with alerts). It takes a few months of data to understand what is "baseline" and what is "abnormal" for each specific system. I expanded on building the process in preventive maintenance by property type.

Practical Considerations Before Installing

Before committing to a monitoring project, there are questions you must answer — otherwise it becomes an expense without a result:

• Connectivity: how do the sensors send data? Long-range, energy-efficient wireless protocols (such as LoRaWAN or Zigbee) suit a whole building; a WiFi-based network is limited in range and battery. The wrong choice leaves sensors "dead" at the edges of the building or in shielded machine rooms.
• Battery and maintenance of the sensors themselves: a wireless IoT sensor lives on a battery. If replacing the battery requires a ladder, access coordination, and a locked room — plan for it in advance. A sensor whose battery dies goes into "silent mode" without anyone noticing, exactly like the problem it is supposed to detect.
• Information security: every sensor connected to the network is a potential entry point. The sensor network must be isolated from the office network, encrypted, and managed with firmware updates. A cheap sensor with poor security is a risk, not an asset — the matter is also grounded in the Privacy Protection Law and the guidelines of the National Cyber Directorate.
• Privacy: presence and counting sensors are subject to the rules of the Privacy Protection Law, 1981. Anonymous occupancy measurement is fine; tracking identified individuals requires a clear legal basis and a transparency policy.
• Data ownership: if the vendor "holds" the data in their cloud and you leave them — you lose the history. Trend history is the value; make sure the data is yours and that there is an export capability.
• Start small: do not wire a whole building on the first day. Choose 2-3 critical systems, monitor them for six months, learn which alerts are real and which are noise — and only then expand. A project too big on the first day almost always ends in graphs no one reads.

What IoT Does Not Replace — and Why It Is Important to Know the Limit

Here there is a dangerous temptation to watch out for: sensors do not replace the statutory legal inspections, nor the licensed professionals. A licensed elevator inspector on behalf of the Administration of Engineering, Planning and Environment (the body that oversees elevator safety), inspection of the fire detection and suppression system per Fire and Rescue Authority requirements, an electrical inspection by a licensed electrical engineer — all these remain a legal obligation, and a vibration sensor on an elevator motor exempts you from none of them.

What IoT does do is fill the gap between the periodic inspections. The elevator inspector arrives, say, once every six months per the requirements; the sensor keeps an eye on the motor every day in between. Real fire safety requires both periodic inspections and daily vigilance, as I described in fire safety in a building. This way you do not choose between continuous monitoring and legal inspection — you get both, and both are essential.

The Bottom Line for the Building Manager

IoT sensors are not a technological gimmick — they are the tool that finally allows the shift from maintenance that reacts to faults to maintenance that prevents them, without waiting for a regulator or an event. But they are only a tool. The value is created when the measurement is connected to an orderly process — a calibrated threshold, an alert, a service call, documentation, and loop closure — and to a single party responsible for the whole fabric.

Without that, the sensors are just more nice graphs no one reads. With it, the building truly speaks — and you truly hear, before it breaks.

At Domera we build a monitoring layer connected directly to the maintenance log and service calls — every alert becomes a documented task, not a forgotten graph. To arrange a review — see comprehensive property management or get a quote.