As part of the GovTech4All initiative, this project addresses the environmental and financial impacts of energy usage in large recreational facilities and serves to broaden our expertise in sustainable and efficient infrastructure management. Recognizing substantial energy needs of the water part for maintaining optimal water and air temperatures year-round, we embarked on this research to develop solutions in the realm of smart energy management. Our approach focuses on optimizing the heating, ventilation, air conditioning (HVAC), and water filtration systems through predictive AI models. We propose real-time and historical data analysis to ensure operational efficiency and cost-effectiveness by continuously optimizing energy-intensive systems.

Technical solution

The system suggests three primary data sources — weather forecasts, visitor trends, and energy prices. By integrating historical and real-time data, our models will dynamically adjust target values for HVAC and water systems. Our research suggests that balancing energy efficiency with visitor comfort is achievable by dynamically adapting operations based on predictive models.

Data processing and prediction models

We envisage the usage of machine learning models, including Long Short-Term Memory  and Auto-Regressive Integrated Moving Average. These models predict optimal settings by analyzing temperature, visitor count, and historical data trends. The LSTM model, capable of capturing long-term dependencies, adapts to evolving patterns in visitor behavior and weather conditions, while ARIMA handles time series forecasting. Together, these models inform device regulation strategies that achieve a balance between energy conservation and visitor satisfaction, continuously improving as they incorporate new data.

Integration with EcoStruxure™ Building Operation (EBO)

The proposed system will interface with water park’s infrastructure using the EBO system, along with BACnet and Modbus protocols for seamless communication with HVAC and lighting controls. This setup will enable the system to send adjustments directly to each device in real-time.

Sustainability-focused Cloud optimization

We propose a cloud-based infrastructure designed to minimize computational resource use, focusing on sustainability through predictive modeling. The system activates only as needed, avoiding excessive computations and ensuring operational resource efficiency. Deployed on low-carbon cloud servers in Europe, the system aligns with environmental standards for sustainable technology. Monthly operational costs are projected to be lower than the obtained energy savings, thus supporting economic and ecological sustainability.