The Application of the GRNNFA to the Determination of Thermal Interface in Compartment Fire
Predicting the fire behavior is a new application area of Artificial Neural Network(ANN).This paper presents the application of a hybrid neural network model denoted as GRNNFA for predicting the height of the thermal interface between the hot and cold gases layers in a single compartment fire.GRNNFA is a hybridmodel to perform regression by the General Regression Neural Network(GRNN) based on the kernels obtained from compressing the training samples by Fuzzy ART (FA).This compression process facilitates the reduction of errors due to noise embedded in the data samples(e.g.experimental data).In this study,the trainingsamples were obtained from the fire experiments conducted by Steckler in 1982.The results confirmed that GRNNFA is capable to capture the behaviour of the thermal interaction during fire.
compartment fire Fuzzy ART general regression neural thermal interface
LEE Waiming YUEN Kityan CHENG Chorkwan HUANG Hechao
Fire Safety and Disaster Prevention Research Group,Department of Building and Construction,City University of Hong Kong,83 Tat Chee Avenue,Kowloon Tong,Hong Kong(SAR)
国内会议
北京
英文
95-101
2004-12-01（万方平台首次上网日期，不代表论文的发表时间）