A direct adaptive control architecture for buildings thermal comfort and energy efficiency optimization using multilayer perceptron neural networks and model reference learning

This paper presents an original direct adaptive control strategy applied to a building heating system, using an intelligent control approach based on Multi-Layer Perceptron (MLP) neural networks. The main contribution of this work is the integration of a newly developed climate database, representing the coldest regions of Algeria, into the Simbad simulation environment. This enhancement significantly increases the realism and robustness of the simulation, allowing for a more accurate assessment of control performance under extreme and challenging climate conditions. The proposed control architecture follows a model reference structure in which an MLP-based adaptive controller ensures real-time adjustment to varying thermal demands and external disturbances. Unlike conventional fixed-parameter controllers, the system learns and adapts online to maintain thermal comfort while optimizing energy usage. Simulation results demonstrate the effectiveness of the proposed method in achieving precise temperature regulation and substantial energy savings. This study highlights the importance of coupling intelligent control techniques with realistic environmental data to develop energy-efficient solutions tailored to diverse climatic contexts.