Practical thermal sensing measurement and neural-thermal comfort index

The primary purpose of heating, ventilating and air-conditioning (HVAC) system is to makeoccupants comfortable. Without real-time practical measurement and method to determinehuman thermal comfort, it may not be feasible that the HVAC system can provide humancomfortable all the time. This paper presents a practical measurement and model to determinehuman thermal comfort index for feedback control. The proposed model is developed basedon the original thermal comfort index called predicted mean vote (PMV) index by applyingfeed-forward neural network model. The model was proposed as an explicit function of theinteraction of the air temperature, wet-bulb temperature, global temperature, air velocity,clothing insulation and human activity. An experiment was done to demonstrate theeffectiveness of the proposed PMV index by comparing to the original PMV index. Theresults show good agreement between the PMV values calculated from the proposed PMVmodel and the original one.