Design and simulation of an energy efficient intelligent kitchen HVAC control system using RBFNN–PID and GA

Abstract

Abstract—Heating, Ventilation, and Air Conditioning systems are essential for maintaining thermal comfort, indoor air quality, and occupant health, particularly in energy-intensive environments such as indoor kitchens. However, conventional HVAC control systems often relyon fixed-gain PID controllers that struggle to handle non-linear dynamics, time-varying disturbances, and indoor air pollution, leading to reduced comfort and increased energy consumption. To address these challenges, this study designs and simulates an energy-efficient HVAC control system based on a hybrid Genetic Algorithm, Radial Basis Function Neural Network–PID approach for regulating indoor temperature, humidity, and carbonmonoxide concentration. The genetic algorithm is first employed to obtain optimal baseline PID gains, ensuring fast and stable initial system response, while the RBF neural network adapts these gains online to cope with environmental uncertainties and disturbances. The HVAC plant is modeled as a three-state dynamic system with realistic thermal, moisture, and pollution is turbances. The simulation results indicate accurate tracking of the reference set points across all controlled variables. The temperature regulation achieved a low RMS error of 0.26726 with a co-efficient-of determination of 0.75193, demonstrating good tracking performance. For relative humidity, the RMS errorwascalculatedas1.1548, while the corresponding R² value was 0 .11706, Carbon monoxide regulation exhibitedalowerRMSerrorof0.014568compared to temperature and humidity; andtheassociatedR²value of 0.99539 confirms satisfactory goodness off it and effective disturbance rejection despite intermittent pollutions pikes. The proposed GA–RBFNN–PID control system demonstrated rapid convergence characteristics, where the controlled parameters settled precisely at their respective points within approximately 1–2 minutes with no over shoot and observable steady-state error, thereby ensuring robust and stable indoor kitchen climate regulation.