摘要
针对发动机特性计算中数据插值精度不高和部件特性的小转速数据难以获得的问题,建立了对部件特性数据进行识别学习的BP神经网络,从而实现了精确插值和对未知特性数据的推测。通过对网络输出结果的分析,表明该网络具有较强的实用性和准确性。
In order to improve precision of data interpolation and overcome the difficulty of the acquisi-tion of low rotate speed data in engine characteristic computing,a Back-propagation network model is es-tablished to study and identify the data of engine components characteristic.Then,an accurate interpola-tion and speculation on the unknown characteristic data is achieved.The analysis of network ' s output re-sults indicates a better practicability and veracity of the network.
出处
《燃气涡轮试验与研究》
2003年第4期15-17,共3页
Gas Turbine Experiment and Research
