基于神经网络的三层织物结构消防服装的热防护性能预测  

Prediction of thermal protection performance of firefighter clothing with three-layer fabrics based on the neural network

作　　者：陈晓渊 鄢友娟 Chen Xiaoyuan;Yan Youjuan(Procurement Service Station of Support Brigade Directly under the Headquarters of Armed Police, Beijing 102613, China;Armed Police Research Institute, Beijing 100071, China)

机构地区：[1]武警总部直属保障大队采购服务站,北京102613 [2]武警研究院,北京100071

出　　处：《产业用纺织品》2021年第3期51-56,共6页Technical Textiles

摘　　要：制作了16种不同结构参数的三层织物结构消防服装试样,对试样的对流换热指数和辐射换热指数进行测量。构建了6种多层感知神经网络(MLPNN),每种神经网络均包含一个隐藏层,输入相同的数据,分别对三层织物的对流换热性能和辐射换热性能进行预测。结果表明,6种网络的预测值与试验值均具有很好的相关性;所构建的MLPNN能够有效预测织物的对流换热指数和辐射换热指数,且有两个输出项的神经网络具有更好的预测性能,能够有效评估消防服装的热防护性能。Firefighter clothing samples of three-layer fabrics with 16 different structural parameters were prepared.The convective heat transfer index and radiation heat transfer index of the samples were measured,and six multilayer perceptron neural networks(MLPNN)were constructed.Each neural network contained a hidden layer,and the same data were input.The convection and radiation heat transfer properties of the three-layer fabrics were predicted respectively.The results showed that there was a good correlation between the predicted values and the experimental values of the six networks.The constructed MLPNN could effectively predict the convective heat transfer index and radiation heat transfer index of the fabrics,and the neural network with two output items had better prediction performance,which could effectively evaluate the thermal protection performance of firefighter clothing.

关 键 词：消防服装 人工神经网络 三层织物 热防护性能 

分 类 号：TS107[轻工技术与工程—纺织工程]