Given the built environment generates 42 percent of annual global CO2 emissions, it is increasingly relevant for owners and operators to drive forward the decarbonization of buildings. However, technological solutions do not only support in terms of sustainability, they also improve efficiency and value. It's no wonder therefore that Artificial Intelligence (AI) tools are indispensable in the buildings sector as well. Rahul Chillar, senior vice president and head of building X at Siemens Smart Infrastructure, explains how building owners and managers can make the best use of AI.

Green and cost-efficient buildings – this is what the future of buildings should look like. In order to reach this goal, the process to increase buildings’ efficiency needs to be an iterative one: Just as humans must learn to crawl and then walk, before they can run, building managers must first gather relevant energy consumption data associated with systems such as heating, ventilation, air conditioning (HVAC) and lighting to measure before being able to take action. As soon as the data is available, building operators can optimize areas, such as energy efficiency, operations, and comfort. “Running” takes place when data from different sources – even across several buildings – is integrated, analyzed, and acted upon in real-time, autonomously based on AI.

Phase #1: Sensors as the basis for data gathering

A lack of data on building systems and operations can be considered as one of the top hurdles towards sustainable buildings. The solution: Deploying sensors and connecting them to the existing on-premise building management systems.

Sensors can provide a wealth of information, for example with HVAC systems. A wide variety of factors such as air flow, humidity, pressure, and temperature affect the systems’ performance. AI algorithms can exploit data on these variables to increase efficiency tremendously. Meanwhile, occupancy and motion sensors can generate real-time data on the usage pattern of a building, which then can be leveraged to further enhance efficiency.

The collection of data is the groundwork needed to aggregate it from different sources, such as building automation and management systems. In the end, it multiplies the ability to increase efficiency and preserve value – either for a single building or across several buildings.

Phase #2: Making smart buildings smarter through AI-enabled analytics

Buildings generate a vast quantity of data. AI algorithms, which have become more common in recent years, can help to turn these into actionable information, that create value for companies and reduce energy consumption. Using AI-based management systems can improve the energy efficiency of buildings by 30 percent. Cloud-based platforms are one example: They can consolidate data from various sources into a digital twin of a building operations and can be open and interoperable to enable the integration of existing software and third-party applications.

In combination with AI, digital building platforms are able to optimize operations and sustainability: Machine-learning can optimize the operation of HVAC systems based on environmental and building data. AI algorithms can also use historical data to forecast and therefore support the reduction of expected energy consumption patterns, costs or CO2 emissions.

In addition, AI-enabled analytics can also help to reduce system downtime and enhance efficiency. By detecting irregular behavior in components such as valves, they are able to send maintenance alerts.

An example illustrates the above: A predictive cooling service learns that the occupancy of the co-working space on the seventh floor always peaks between 2 p.m. and 3 p.m. Hence, the HVAC system can predictively plan and account for the respective cooling load.

In the end, the ideal AI is the one that works in the background and thus allows operators to focus on their business. Therefore, maintenance and retraining of algorithms should occur automatically, while models continue to maintain their accuracy as conditions and their associated data change.

Phase #3: From smart to autonomous and beyond

Today, AI is able to proactively detect faults before they occur. Much more should be possible in future: For example, AI could be configured to take specific actions, such as triggering a task based on the collected data. It could also increase energy efficiency in a building through optimizing space utilization. Building occupants would be advised to work or stay in a central area, which is cooled or heated – depending on the season.

Although AI will become more autonomous, it will always be guided by the parameters that building managers define for it: It could adapt the air exchange rate in flu seasons to protect building occupants’ health. Or it could optimize HVAC components to extend the system’s lifespan.

Before AI can support building owners and managers to make their buildings smarter, they need to make use of the data available by technological solutions. In the end, the result will be autonomous behavior, where AI has learned, and all the actions are run continuously. There is certainly a long way to go until more capabilities are added and all single building assets are improved. However, the potential benefits are huge – be it in terms of value, productivity or sustainability.