The Study of Neural Network in the Application of PMV Index

With the occupant´s pursuit of indoor thermal comfort, the conventional air-conditioning control system that only control air temperature and air humidity can´t satisfy people´s require. On the basis of the study on the thermal comfort theory and thermal comfort index, a new type of air conditioning control mode, thermal comfort index control was proposed which takes thermal comfort index as the control parameter. This model takes the six factors that affect human thermal comfort as input and PMV value as output. The thermal comfort index control not only can meet the thermal comfort requirement of human bodies, but also has great energy saving potential. As air-conditioning control system has the characteristics of time delay, large inertia, and highly nonlinear, it is difficult to get accurate mathematical model if we use traditional control theory. Neural network has highly adaptive and on-line self-learning ability, and it can approximate any nonlinear mapping. If we use neural network to train, we can get input-output model of air-conditioning control system. So we can realize the intelligent control of air-conditioning.