五孔探针的神经网络处理方法  

Five-hole probe and its neural network algorithm

作　　者：王洪伟[1] 陈小龙[1] 

机构地区：[1]北京航空航天大学能源与动力工程学院航空发动机气动热力科技重点实验室,北京100191

出　　处：《航空动力学报》2010年第2期417-423,共7页Journal of Aerospace Power

摘　　要：对于五孔探针的数据处理方法就行了详细的研究,发展了一种人工神经网络的数据处理方法,并将其与三维线性插值法进行了全面的对比,证明了所发展的神经网络法的精度更高,可靠性更好,而且经过简单改动就可以应用在其他多孔探针上.Data reduction methods for five-hole probe are investigated and a new method based on artificial neural network is developed. The developed method is then compared with a three-dimensional linear interpolation one. Conclusion is drawn that the neural network method is superior to the interpolation one in accuracy and robustness. The new developed method is also applicable to multi-hole probe through minor modifications.

关 键 词：五孔探针 数据处理 人工神经网络 精确度 

分 类 号：V211.71[航空宇航科学与技术—航空宇航推进理论与工程]