IoT For Building Energy Systems In Zero-Emission Buildings

By Dirk Mayer and Olaf Enge-Rosenblatt

How can buildings contribute to a significant reduction in global primary energy consumption?

Due to the global trend toward reducing CO₂ emissions and resource conservation, the demands are increasing with regard to the efficiency of heating, ventilation and air-conditioning (HVAC) systems operated in buildings.

In conflict with calls for faster measures to reduce CO₂ emissions and conserve resources, buildings are very long-lived objects that exist for 50 to 100 years. As a result, new concepts based on modern construction materials are only slowly gaining ground. On the other hand, it is much simpler to upgrade an existing building into a smart building. This allows significant energy savings of up to 20% to be realized within a short time. But four prerequisites are necessary for effective energy-management:

• Sensors for distributed acquisition of the relevant measurement values;
• A data network in the building and connection to the environment;
• Powerful computing systems for realizing advanced control and regulation functions, and
• Predictive algorithms that aggregate the information, determine optimal control programs for the energy system, and enable condition-based maintenance of components.

Because these measures require lower investments than large construction measures, they pay for themselves much quicker. IoT solutions also can play an important role here.

Wireless, energy-independent sensors can be installed in existing buildings to measure the temperature, humidity and air quality (for instance, in the form of CO₂ concentration). It is also important to determine the distribution of people within the building in order to take the impacts on the room climate into account, as well as for heating and cooling of only the necessary rooms. This can be achieved with a variety of sensors. However, the use of cameras, microphones or measures to determine the number of smartphones in the area always pose challenges with respect to data privacy. It is advantageous here to use sensor systems that preprocess and anonymize the data close to the hardware level or that are based on measurement principles that do not permit the identification of personal data.

In general, it can be expected that the use of IoT technologies in buildings will increase even further, driven by the rising importance of climate protection, creating a growing market for sensor systems, network technologies and embedded systems. The key challenges for the development of IoT components on the path to emissions-free smart buildings are:

• Economical realization: Current energy systems for buildings are designed according to the same paradigms as automation systems in industry, e.g. the hierarchical structure of an automation pyramid. Future developments in the area of industry are moving in the direction of networked, cognitive systems (cyberphysical systems). These allow a flatter network topology that can also be economically implemented in buildings.
• Reliability of the systems: In order to realize the efficiency improvements in operation and maintenance, the additionally installed components must be extremely reliable and nearly maintenance-free.
• Data sovereignty, data protection and IT security: The digitalization of buildings also creates new attack surfaces. In some cases, the systems collect personal data that must be protected from unauthorized access. The networking of a building also makes it a worthwhile target for cybercriminals. These concerns must be taken into account from the very start in the development of IoT solutions for buildings if the solutions are to meet with lasting success.
• Suitability for installation in existing buildings: Wireless sensor systems that harvest their own energy enable the retrofitting of instrumentation or temporary installations for determining optimal configurations. Naturally, it is also necessary to create options for integrating the energy producers into a networked automation system.
• Data fusion: Methods of machine learning and physical modeling merge heterogeneous sensor data to improve the precision of inputs, or estimate required values for which no sensors are available.

Olaf Enge-Rosenblatt is the group manager for computational analytics at Fraunhofer EAS. He holds a degree in electrical engineering from Chemnitz University of Technology.

Dirk Mayer is the Head of Department for Distributed Data Processing and Control at Fraunhofer EAS.