Research And Realization Of Human Thermal Comfort Model Based On Multiple Physiological Parameters Fusion

The energy issue is the key problem to every country in today’s world, China as adeveloping country which is advancing rapidly, the energy problem directly restrictsthe economic development speed and quality. Along with quickening step of ruralurbanizatation in china, the building volume is also growing rapidly, causes energyconsumption growing fast. Building energy conservation is beneficial to the people’slivelihood problems must be solved for in china. But, building energy saving is notonly energy-saving, but also should assure comfort, health and green, and should havemore connotation of building energy conservation.People has a increasing demanding to the comfort and health of buildingenvironment, indoor thermal comfort evaluation has also attracted more and moreattention. Although the study of thermal comfort is relatively mature, predecessorshave established a number of thermal comfort model, the thermal comfort model basedon physiological parameters is more accurate, can reflect the thermal comfort ofindividual in the indoor more accurate and other advantages, this paper emphasizes thethermal comfort model of multiple physiological parameters based on research.Dissertation aims to study the identification methods of physiological parameters,through the recognition of physiological signals to give the accurate comfort feeling.Model can not only as a scientific explanation of human thermal comfort and provide atheoretical basis, but also through the human body physiological parameters todetermine the parameters for the environment, air temperature, humidity, wind speed,to provide theoretical guidance, and eventually to build a comfortable indoorenvironment and provide the theoretical support for the formulation of relevantstandards, and provide guidance for more scientific and building energy saving.Dissertation designed the thermal comfort experiment first, designed two differentenvironmental parameters experiment environment but other conditions are the same,collected indoor environment parameters, two in the environment of humanphysiological signals and fill out the subjective questionnaire. Analyzed the reliabilityand validity of questionnaire results, two indicators are qualified; we can regard thedata scientific and effective, and can be used to study. We got the ECG, BP and GSRsignal data through experiment. We choose the more suitable feature extraction andfeature combination for establishing model through theoretical analysis andexperimental verification, that is, extracted energy feature using the wavelet packet extraction method. Then we uesd the BP neural network based on MIV value methodto do the fusion and reduction of three kinds of physiological signals. BP neuralnetwork as a classifier, select appropriate parameters of the neural network. The newfeature vector as input, the subjective thermal comfort questionnaire survey results asoutput. We selected65of the80experiment samples as the training signal, theremaining samples as test signal. Summer and winter thermal comfort model ofthermal comfort model obtained by experiment, compared with the classical PMVformula calculation, to verify the effectiveness of our thermal comfort model.