神经网络的非设计点损失落后角模型在流场诊断中的应用

１前言流场诊断技术是利用部分实测结果并配以若干简化数值计算软件,判断并得出尽可能多的流场信息的一门综合技术［１］、对于多级轴流压气机,由少数易于测量的部位的流场参数,例如进出口全部参数及部分外端壁静压值,推算出其它部位流场参数,是流场诊断技术的典型示例。当压气机远离设计点工作时,其流场中存在着严重的粘性效应,即分离和旋涡,较为通用的损失与落后角模型迄今尚未见到。本文利用神经网络的映射特性来重新关联实验数据,获得了新的基于神经网络的非设计点损失与落后角模型,将流场诊断技术推广到了压气机非设计点。流…

USE OF LOSS AND DEVIATION MODEL BASED ON NEURAL NETWORK IN FLOW FIELD DIAGNOSIS OF MULTISTAGE AXIAL COMPRESSORS

Flow Field Diagnosis" is a unified technique which takes full advantages of the theo-retical calculations and experimental measurements, to rapidly and accurately get as muchas possible the flow information in the turbomachinery. System of flow field diagnosis contains S2 modal, off-design loss and deviation model, and optimization modal. This paperdeveloped a new off-design loss and deviation model based on the neural network. The diagnose system including the modal shows good anastomotic with the experiment result intotal characteristic and stage characteristic of multistage axial compressor.