Insights · Automation

Why Do Building Controls Drift, and How Do You Spot It?

Every building drifts from how it was set up. The question is whether anyone notices.

Focus: Controls drift and fault detectionAudience: Owners · Property Managers · Facility TeamsPublished: November 16, 2026Published by: Airiva Inc.

Introduction

Tuned Once, Then Left Alone

If a building’s automation system was ever set up properly, it usually happened once. A commissioning team tests the sequences, checks the sensors, sets the schedules and hands over the keys. Some buildings never even get that. Then the building lives with those settings for twenty years or more.

A lot happens in twenty years. Tenants move in and out. Office hours change, and then hybrid work changes them again. Someone raises a setpoint to stop a complaint. A contractor swaps a sensor with whatever was on the truck. A valve starts to stick. None of these is a big event on its own, and that’s the problem.

It’s a lot like a car’s wheel alignment. Nothing breaks. The car still drives. But it pulls a little more to one side every month, and you only notice when the tires are worn. In our view, controls drift is one of the most common reasons a building that performed well on day one is wasting energy by year five.

Section 01

How Common Is It? More Than Most Owners Think

The best large-scale picture comes from a 2023 study by Lawrence Berkeley National Laboratory and partners, published in Science and Technology for the Built Environment. The researchers pulled at least a year of data, and often two, from fault detection software in 317 buildings, covering nearly 57,000 air handlers and zone units, and sorted over 90 types of fault.

Their headline finding: on any given day, 40 percent of air handling units and 30 percent of the smaller air terminal units that serve individual zones had a reported fault of some kind.

Some faults showed up almost everywhere. Problems with the supply air temperature setpoint appeared on 55 percent of air handlers at some point. Cooling coil valve faults appeared on 35 percent, and heating coil valve faults on 22 percent. In the zones, faulty temperature sensors and abnormal zone temperatures each appeared on 46 percent of terminal units. And many of these faults weren’t brief. Eighteen types of air handler fault were present for more than a fifth of the time covered by the data.

The authors are careful about what this does and doesn’t prove. The buildings aren’t a random sample: owners who pay for fault detection software may run better buildings than average. And the software can raise false alarms or miss real faults. Even so, early field checks suggested the software data gave an accurate picture of how common faults are.

40% Air handlers with a reported fault on any given day, across 317 buildings
46% Zone units that showed a temperature sensor fault at some point in the study
55% Air handlers that showed a supply air temperature setpoint fault at some point
Section 02

Four Ways a Building Drifts

Drift doesn’t have one cause. It usually comes from four places, and each needs a different fix.

Type of drift What it looks like How to spot it
People Overrides left on after a complaint, setpoints nudged up or down, schedules extended for one event and never reset An override report from the automation system, reviewed every month
Sensors Temperature, humidity, CO₂ or airflow sensors slowly reading wrong, so the system makes good decisions on bad numbers Spot checks against a calibrated handheld meter, on a set schedule
Equipment Valves that leak through when closed, dampers that stick, actuators that no longer reach full travel Trend data showing heating and cooling running at the same time, or a valve “closed” while the coil still changes temperature
Use The building is used differently than when it was set up: new tenants, new hours, fewer people on Fridays Comparing schedules and ventilation rates with how the space is actually used today

The first type is the most human, and the hardest to fix with technology alone. Operators are judged on comfort calls, not on energy bills. A 2019 ASHRAE Journal study went back to 28 buildings that had been recommissioned and measured how much of the savings were still there. About 61 percent of the savings had held overall. Three years after the work, 76 percent remained. Six years after, only 49 percent did. The causes the authors pointed to were comfort complaints and overrides, staff turnover, renovations that changed how spaces were used, and a lack of training. Nobody is doing anything wrong. They’re solving today’s problem, and the cost lands somewhere else.

Section 03

Why Drift Stays Hidden

Most automation systems are good at shouting when something fails. A fan stops, a pump trips, a space overheats, and an alarm goes off. Drift rarely trips an alarm, because the building usually stays comfortable while it drifts.

That’s the clever and annoying part of a modern control system. If a heating valve leaks through, the cooling valve opens a little more to compensate. If a zone sensor reads two degrees high, the system pushes more cold air into that zone. The room feels fine. The meter spins faster. It’s like driving with the handbrake half on: the car still gets you there, it just burns more fuel and wears out sooner.

A building can be comfortable and wasteful at the same time. Comfort complaints only catch half the problem.

Sensor faults deserve special attention, because every other decision depends on them. The Berkeley Lab researchers concluded that their results emphasize the importance of scheduled sensor calibration, and noted that sensor problems can also stop fault detection software from catching more complex faults. If the thermometer is wrong, everything downstream is wrong too.

Section 04

How to Spot It Before the Bill Does

Spotting drift doesn’t require expensive software to start. It requires looking at the right things on a schedule.

Three regular checks that catch most drift A starting routine, not a full program
Every month
Review overrides

Print the list of points in manual or override. Every item should have a reason and an end date. Anything older than a month gets reset or justified.

Every quarter
Read the trends

Look for heating and cooling at the same time, equipment running when the building is empty, and valves or dampers stuck in one position.

Every year
Check the sensors

Compare key temperature, humidity, CO₂ and airflow sensors against a calibrated meter. Fix the sensors before trusting anything else.

For larger buildings or portfolios, software can do much of the watching. The U.S. Department of Energy’s Smart Energy Analytics Campaign, a four-year program with 104 organizations and more than 6,500 buildings, reported in 2020 that participants using fault detection and diagnostics software saw median energy savings of 9 percent a year, and that their investment in analytics paid for itself in about two years.

Ongoing monitoring also seems to make commissioning itself work better. A 2020 Berkeley Lab review of about 1,500 North American buildings found median energy savings of 9 percent for existing building projects in monitoring-based utility programs, against 5 percent for standard utility program projects. The median payback across existing building commissioning was 1.7 years.

What This Changes

Treat Controls Like Equipment, Not Furniture

Most owners maintain boilers and chillers on a schedule. Far fewer maintain the controls that tell those boilers and chillers what to do. The controls get treated like furniture: installed once, expected to last.

The shift is small but real. Put sensor calibration on the maintenance plan. Make override reports part of the monthly routine. Keep at least a year of trend history so this winter can be compared with last winter. And plan to recommission every few years, or after any big change in how the building is used, rather than waiting for the energy bill to ask the question.

At Airiva we work by a rule we call performance before plaque: find out how a building actually behaves before giving it credit. Controls are where that rule matters most, because the system’s own screens will usually tell you everything is fine.

Conclusion

Drift Is Normal. Not Checking Is a Choice.

Every building drifts. The sensors age, the people change, the use changes. That isn’t a failure of the building or the operators. It’s just what happens to a complex system left on its own for years.

The difference between a building that holds its performance and one that slowly loses it usually isn’t better equipment. It’s someone checking, on a schedule, whether the building is still doing what everyone thinks it’s doing.

About Airiva Airiva is an Alberta-based building intelligence advisory firm. We read every building through five lenses (environment, energy, automation, compliance, and insight) and show owners where to act first. From signal to action.
Building Controls Building Automation Fault Detection Sensor Calibration Commissioning Airiva

Published by Airiva Inc., an Alberta-based building intelligence advisory firm. No equipment brands, software products or service providers are recommended, and no endorsement is intended.

Sources read: Crowe et al., “Empirical Analysis of the Prevalence of HVAC Faults in Commercial Buildings,” Science and Technology for the Built Environment 29, 2023, Lawrence Berkeley National Laboratory with Pacific Northwest National Laboratory and the University of Nebraska-Lincoln (317 buildings; fault prevalence and persistence figures; dataset limitations; sensor calibration finding); Gunasingh, Hackel and Zhou, “Persistence in Energy Savings from Retro-Commissioning Measures,” ASHRAE Journal, December 2019 (28 buildings; 61 percent overall persistence, 76 percent at three years and 49 percent at six; causes of lost savings); Lawrence Berkeley National Laboratory, “Berkeley Lab Building Efficiency Campaign Drives $95M in Annual Energy Savings,” October 2020 (final Smart Energy Analytics Campaign results: 104 organizations, 6,500+ buildings, 9 percent median savings with fault detection, about two-year payback); Crowe et al., “Building Commissioning Costs and Savings Across Three Decades and 1,500 North American Buildings,” Energy and Buildings 227, 2020 (median savings by program type and 1.7-year median payback).

← Back to Insights