IoT-Based Sustainable Architecture: Energy Efficiency Optimization in Tropical Climate Buildings

Abstract

Buildings in tropical climates such as Southeast Asia exhibit cooling-dominated energy profiles, with HVAC systems typically accounting for 55 to 70 per cent of total building electricity consumption. The Internet of Things (IoT) combined with machine learning offers significant opportunities to optimize energy efficiency in such buildings while maintaining occupant comfort. This study proposes and evaluates an IoT-based building energy management framework integrating temperature, humidity, CO2, occupancy, illuminance, and energy meter sensors with an edge-cloud computing architecture and an LSTM-based cooling load predictive controller. The framework was deployed in three test-bed buildings: a 5,200 m² office building in Kuala Lumpur, a 3,800 m² commercial building in Surabaya, and a 4,500 m² mixed-use facility in Bangkok. Over a 12-month evaluation period (April 2023 to March 2024), the framework achieved energy intensity reductions of 21.4 per cent (Kuala Lumpur), 19.6 per cent (Surabaya), and 23.1 per cent (Bangkok) compared to pre-deployment baselines. Peak demand reduction averaged 28.7 per cent across the three sites. Thermal comfort indicators (PMV index, PPD percentage) remained within ASHRAE 55 acceptable ranges throughout the evaluation period. The LSTM cooling load forecasting model achieved R² of 0.92 with RMSE of 8.4 kW. Cost-benefit analysis indicates an average payback period of 3.4 years, supporting the economic feasibility of IoT-driven energy efficiency in tropical climate buildings.