炉膛声学层析测温探头分布方式研究  

Study of Probe Distribution for Furnace Acoustic Tomographic Temperature Measurement

作　　者：张立峰[1] 苗雨 ZHANG Lifeng;MIAO Yu(Department of Automation,North China Electric Power University,Baoding 071003,Hebei Province,China)

机构地区：[1]华北电力大学自动化系,河北省保定市071003

出　　处：《发电技术》2021年第3期329-335,共7页Power Generation Technology

基　　金：国家自然科学基金项目(61973115)。

摘　　要：炉膛声学层析测温系统可获得炉膛温度二维/三维分布图像,超声波探头分布方式将直接影响其成像精度。研究了8、12及16个超声波探头的4面与4角分布方式,先采用Tikhonov正则化重建算法获取粗网格下的温度分布,再采用局部加权回归法预测网格细化后的温度分布。选取典型的温度场分布,对所研究的6种探头分布下温度场重建后的均方根误差及相关系数进行了比较,结果表明,采用16探头进行4面分布时,其重建精度最高。The two/three-dimensional distribution image of furnace temperature can be obtained by furnace acoustic tomographic temperature measurement system.The distribution of ultrasonic probes will directly affect its imaging accuracy.Four-side and four-corner distribution of 8,12 and 16 ultrasonic probes were studied,respectively.Tikhonov regularization algorithm was adopted to obtain the temperature distribution under coarse mesh,and then the local weighted regression method was used to predict the temperature distribution after mesh refinement.Typical temperature field distributions reconstruction was carried out for the 6 probe distributions,and the root mean square error and correlation coefficient were analyzed.The results show that the highest accuracy can be obtained using 16 four-side distribution probes.

关 键 词：锅炉 声学层析成像 探头分布 温度场重建 TIKHONOV正则化 

分 类 号：TK311[动力工程及工程热物理—热能工程]