Artificial intelligence (AI) opens up new possibilities for operating ventilation systems in a more energy-efficient and proactive manner. Its use in swimming pools is particularly interesting, as temperature, humidity, and air quality must be reliably maintained there, while at the same time high energy demands arise for ventilation, dehumidification, and heating.

HANSA Klimasysteme has been working on model-predictive control of HVAC systems for several years. With the etaSmart system, this technology is used specifically for the energy optimization of swimming pool ventilation systems.

Modern HVAC systems are already equipped with numerous sensors and measurement values. For example, they monitor temperature, humidity, airflow rates, damper positions, and other operating data. Integrating this information creates the conditions necessary for intelligent control.

In traditional control systems, operating modes and control parameters are determined based on proven experience. This allows the desired indoor air conditions to be reliably achieved. However, this does not automatically mean that the most energy-efficient operating mode is selected at all times.

AI and model-based control systems can consider multiple influencing factors simultaneously. In addition to the condition of the HVAC unit, factors such as the building’s behavior, weather conditions, and occupancy patterns can be taken into account. The system thereby attempts to optimize comfort, required indoor air conditions, and energy consumption in tandem.

Swimming pool ventilation is a relatively complex control task. The temperature and humidity of the pool area must be maintained within the specified ranges, even as outdoor conditions and pool usage constantly change.

In addition, a swimming pool HVAC unit has various options for achieving the desired indoor air conditions. These include, among other things, the proportion of outdoor air and recirculated air, air flow rate, heating capacity, and—for units equipped accordingly—the operation of a heat pump.

These options influence one another. For example, changing the position of a damper not only alters the air mix but also the flow resistance and, consequently, the power consumption of the fans. A heat pump can utilize additional heat from the exhaust air, but it requires electrical energy to do so. Therefore, many factors must be considered simultaneously to select the most efficient operating point.

This is precisely where intelligent control methods excel: they can compare different operating scenarios and determine how the required heating and dehumidification needs can be met as efficiently as possible.

One option is model predictive control, or MPC for short. Put simply, the controller not only reacts to the current state but also takes into account how the situation is expected to develop.

In HANSA’s etaSmart system, models of the HVAC unit and the building are combined for this purpose. The HVAC unit is described using a physical model. A neural network is used to model the building’s more complex behavior.

Based on available data, the building model can estimate how the pool temperature and humidity will change. Factors such as heating, dehumidification, weather conditions, and occupancy are taken into account. This information is used to determine future heating and dehumidification requirements.

etaSmart then calculates suitable operating points for the HVAC system. The goal is to maintain the required conditions in the swimming pool while using as little electrical and thermal energy as possible.

One advantage of the system is that the existing DDC control system can continue to be used. etaSmart interfaces with the control system via an industrial PC and transmits the optimized settings to the DDC.

The technology is not merely being studied in theory. HANSA has already implemented etaSmart in various swimming pools. Depending on the type of system and weather conditions, an average energy savings of about 20% was determined for the ventilation systems.

The Ramsloh indoor swimming pool was examined in particular detail. There, conventional and model-predictive control systems were compared over an extended period. HANSA reports average savings of 24% in electrical energy and 29% in thermal energy.

In another project involving three ventilation zones, etaSmart was expanded to include integrated zone control. This also made it possible to reduce the required air duct pressure. Initial results from this project showed savings of 25% in electricity and 42% in heat.

The results demonstrate the technology’s potential but should not be interpreted as a blanket promise of savings. The actual magnitude of the savings depends on the HVAC system, the building, its use, weather conditions, and the system’s previous operating history.

The models used in etaSmart open up additional possibilities beyond energy optimization. One future area of application, for example, is predictive maintenance.

When the mathematical model operates in parallel with the actual HANSA HVAC unit, calculated and measured system behavior can be compared. Significant deviations could, for example, indicate contamination, wear, or changes in individual components.

As a result, the intelligent control system is evolving toward becoming a digital twin of the HVAC system. In addition to energy optimization, this could also support system monitoring, fault detection, and maintenance in the future.

AI in ventilation technology does not simply mean replacing a conventional controller with “artificial intelligence.” The key advantage comes from considering the HVAC unit, the building, the weather, and building occupancy together. Particularly in the case of energy-intensive swimming pool ventilation, such predictive control can help maintain the required indoor air quality while using less energy. The practical projects described in the technical article show that this technology is already in use today and, at the same time, opens up further possibilities for digital twins and predictive maintenance.

This wiki article is based on the technical paper “Artificial Intelligence in Ventilation Technology” by Dr.-Ing. Matthias Lamping, published as a special reprint from cci Zeitung 02/2026.