摘要
针对变风量空调系统风口结构研究中,要求室内环境参数多点测温的特点,采用数字化测温技术和单总线传输方式,构成了矩形测温网络,并在3种风口工况下,进行了大量实验。测量过程中,由于测温网络出现的个别传感器故障而导致不能正确读数时,可通过BP人工神经网络预测该点的温度值予以补充。另外对半导体传感器最高采样频率的限制也进行了一定研究。
In the study of air opening structure of indoor variable air volume in the air conditioning system, it is required that environmental parameters are measured from multiple points. The aim of this study was to establish a measuring method using digital temperature measurement technology and 1-wire bus data communication to form rectangle temperature measuring network. The extensive experiments were carried Out under 3 working conditions of air opening conditions. During the measurements, when the system gave an incorrect value due to a single sensor malfunction, the value in that sensor was obtained through BP neural network prediction. The study also investigated the limit of the sample frequency through a semiconductor sensor.
出处
《西安科技大学学报》
CAS
北大核心
2008年第3期564-568,共5页
Journal of Xi’an University of Science and Technology
基金
陕西省教育厅专项科研计划项目(04JK214)
关键词
矩形测温网络
BP人工神经网络
采样频率限制
rectangle temperature measuring network
back-propagation artificial neural network
sample frequency limit
