人工神经网络在热舒适实验研究中的应用

使用人工神经网络(ANN)方法处理影响人体热舒适的多种环境因素和人体的热反应之间的关系,以探讨用于各类环境热舒适性预测的可行性.针对稳定热环境和非稳定热环境下的热舒适实验,建立了BP神经网络模型,利用实验数据对网络进行训练和测试,检验其预测的准确性.评测结果表明,对稳态环境,该模型较传统的线性回归方法有更高的预测精度;对非稳态环境,避免采用回归方法遇到的非线性关系处理问题,能合理地预测非稳态条件下热反应的变化.因此,合理建立各类热环境的ANN模型,用已有的热反应数据训练该模型,实现对环境热舒适性的预测是可行的.

Feasibility of Artificial Neural Network on Thermal Comfort Research

The possibility of using artificial neural network (ANN), which models the relationship between the thermal environmental factors and the residents' thermal reaction to predict the environmental thermal comfort, is discussed.The ANN models based on BP algorithm were respectively built according to the thermal comfort experiments done under steady and unsteady conditions.The experimental data are used to train and test the model.The test result showed that the ANN model had higher precision in prediction than the traditional method-linear regression under steady condition.And under unsteady condition,the model can also be correctly used to predict the change of the residents' thermal reaction with avoiding dealing with the complicated nonlinear relation while using the regression method.So it is feasible to predict the environmental thermal comfort with building and testing ANN model to different thermal environment.