> For the complete documentation index, see [llms.txt](https://docs.buildings.ability.abb/collection/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.buildings.ability.abb/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings.md). # Introduction to Anomaly Detection in Smart Buildings Most building problems do not begin with a dramatic failure. They start quietly. A ventilation system runs longer than it should. Cooling stays active during a mild holiday. A floor develops a slightly higher night load than usual. Nothing looks critical in the moment, but over days and weeks the result becomes expensive: wasted energy, longer equipment runtime, and time lost trying to understand what happened. This is why anomaly detection matters in smart buildings. It helps reveal behavior that does not fit expected operation before that behavior turns into a larger operational or financial problem. ## Why this matters Modern buildings generate a constant stream of data. Energy meters, sensors, schedules, weather signals, and technical systems all describe how the building behaves over time. Hidden in that data are patterns that can point to waste, faults, inefficiencies, or unusual operation. The challenge is that many of these patterns are easy to miss. Some do not cross a hard alarm threshold.\ Some only become visible when viewed over several days.\ Some look normal at first glance and only become suspicious when seen in the right context. A building may consume the same amount of energy on two different days, and one case may be normal while the other is not. That is the core reason anomaly detection is useful in buildings. It helps answer not only whether something changed, but whether the behavior still makes sense. ## Why buildings need a different approach Buildings are not simple systems with one stable baseline. Their behavior changes with: * weather * occupancy * schedules * holidays * system states * control actions Because of this, normal behavior is not fixed. A load that is expected on a cold weekday morning may be unusual on a mild Sunday night. A value that looks acceptable in absolute terms may still reflect persistent waste. A problem that becomes visible at the main energy meter may actually begin deeper in the building, for example in sensing, control, or equipment behavior. This makes anomaly detection in buildings fundamentally contextual. It is not only about finding extreme values. It is about finding behavior that does not fit the situation. ## Why simple alarms are not enough Traditional threshold alarms still have their place. They are useful when there is a clear limit that should never be crossed. But many important building issues do not look like that. Simple thresholds often miss: * recurring night load that stays below a fixed alarm line * weekend operation that looks normal globally but is wrong in context * slowly increasing baseline drift * persistent inefficiencies that never create one dramatic spike This is why anomaly detection matters. It helps move from simple limit checking to a more meaningful question: **Is this building behaving the way it should right now?** ## Why AI changes what is possible This is also where Artificial Intelligence (AI) becomes important. Because building behavior depends on many interacting factors, anomaly detection benefits from methods that can work with context, changing operating patterns, and large volumes of time series data. AI makes it much more practical to detect subtle deviations, compare actual behavior with expected behavior, and turn complex patterns into something an operator can understand and act on. ABB positions ABB Ability™ BuildingPro Suites as a unified cloud intelligence layer with proactive AI capabilities that help streamline data collection, support analytics, and uncover anomalies before they become issues. ## What good anomaly detection should do A useful anomaly detection approach should do more than raise alerts. It should help answer practical questions such as: * Is this behavior really unusual? * Is it unusual for this time and condition? * Is it short-lived or persistent? * Does it likely matter financially? * Where should the investigation start? In other words, the goal is not just to find statistical oddities. The goal is to find behavior that is operationally meaningful. ## Why this matters in ABB Ability™ BuildingPro Suites ABB Ability™ BuildingPro Suites connects building automation, Heating, Ventilation and Air Conditioning (HVAC), energy, Information Technology (IT), and Operational Technology (OT) systems into one connected intelligence layer. That makes anomaly detection a natural and important topic in this context. When building data is connected and interpreted together, it becomes possible to move beyond isolated alarms and toward a clearer understanding of how a building is actually operating. ## What this academy will explain This academy section explores anomaly detection from first principles to practical implementation. The goal is not only to explain the concept. It is to explain why anomaly detection has become an important capability for smart building operations. Use these pages to move step by step through the topic: {% content-ref url="/pages/yrA3nz5PK3w5i3OSSWEF" %} [Foundations of Anomaly Detection](/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings/foundations-of-anomaly-detection.md) {% endcontent-ref %} {% content-ref url="/pages/GQ3mmddxaIMC2CbI7Ie6" %} [Understanding Building Data](/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings/understanding-building-data.md) {% endcontent-ref %} {% content-ref url="/pages/1ZOENvRd7SwzxjeLqmXz" %} [From Building Data to Detection Challenges](/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings/from-building-data-to-detection-challenges.md) {% endcontent-ref %} {% content-ref url="/pages/4MTnKYAMfs1GSUXr5ugx" %} [How BuildingPro Suites Addresses the Challenges of Anomaly Detection](/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings/how-buildingpro-suites-addresses-the-challenges-of-anomaly-detection.md) {% endcontent-ref %} For the product view, continue here: {% content-ref url="/pages/Wt9AF6p1ggbPrdI12gPq" %} [Energy Anomaly Detection](/collection/suites/energy-and-sustainability-intelligence/energy-anomaly-detection.md) {% endcontent-ref %} {% content-ref url="/pages/9vMvp1iU0OmKO5DY5Xep" %} [Configure Energy Anomaly Detection](/collection/suites/energy-and-sustainability-intelligence/energy-anomaly-detection/configure-energy-anomaly-detection.md) {% endcontent-ref %} ## Conclusion Anomaly detection matters in smart buildings because many important problems are not obvious, not immediate, and not well captured by simple alarms. Buildings are dynamic, context-dependent systems, and meaningful deviations often only become visible when data is interpreted in the right way. AI makes this increasingly practical by helping turn large amounts of building data into useful signals, explanations, and actions. This academy explains what that takes, why it is difficult, and how a practical building-focused solution approaches it. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.buildings.ability.abb/collection/academy/building-intelligence-ai-and-machine-learning-in-buildingpro-suites/machine-learning/introduction-to-anomaly-detection-in-smart-buildings.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.