Study on Thermal Comfort Control Based on Artificial Neural Network

In order to investigate control method for indoor thermal comfort, the experimental environment of indoor thermal comfort is designed in this paper. Thermal comfort data are collected based on questionnaire means, ANN (artifical neural network) is trained using the data, and the ANN model trained can accurately forecast thermal comfort index. Relevant input parameters is conversely computed based on fixed network output, optimum control parameters in indoor air-condition can be obtained in order to achieve desired control index. The experiment results indicate that the control method proposed synthetically considers the effect on thermal comfort owing to indoor air temperature, outdoor air temperature and clothing case. Therefore, the forecast precision for thermal comfort is improved and indoor personnel´s requirement to thermal comfort can be satisfied.