应用反向传播神经网络算法的发动机动态总压畸变预测

为提高动态总压畸变预测的准确性,通过引入反向传播神经网络算法,研究其在紊流相关中的可行性和准确性.结果 表明:反向传播神经网络可应用于紊流相关以预测动态总压畸变;采用单个工况进行自我预测时,神经网络显示了良好的预测能力,预测结果与试验值吻合很好;采用反向传播神经网络法,5个工况作为样本预测各工况紊流度时,样本工况的平均...

Engine Dynamic Total Pressure Distortion Prediction Appling BackPropagation Artificial Neural Network

In order to improve the accuracy of dynamic total pressure distortion prediction, the feasibility and accuracy of BP artificial neural network in turbulence correlation are studied. The results show that: the BP artificial neural network can be used to predict the dynamic total pressure distortion. When a single condition is used for self-prediction, the BP neural network shoes good predictive ability, and the predicted results are in good agreement with the experimental values. When the BP neural network is used to predict the turbulence of 5 working conditions as samples, the average turbulence of the sample is consistent with the test value, while the average turbulence of individual conditions in the test conditions is somewhat different from the test value. The network setup and training need to be further studied to improve the prediction ability of the network. Compared with the least square method, BP neural network is a more promising method.