基于BP神经网络的三角机翼飞行载荷模型研究

Study on Flight Load Model for Delta Wing based on BP Neural Network

针对三角机翼结构设计复杂、传力路径多和大载荷引起结构非线性等因素,采用BP神经网络方法建立飞行载荷模型。通过地面静力试验,利用某型三角机翼载荷标定试验数据,建立机翼根部剪力和弯矩的载荷模型,并对多个校验工况进行验证。经与多元线性回归载荷预测结果对比分析,结果显示,BP神经网络模型预测误差都在3%以内,特别对于剪力,BP神经网络预测误差明显小于多元线性回归,表明BP神经网络可作为测量三角机翼飞行载荷的一种更加有效的工程方法。

Aiming at the complexity of structural design of delta wing, the multiple paths of strength transfer, and the structural nonlinearity under big loads, and so on, the flight load mod- el is built by using the BP neural network method. Through the static test, the root load model of shear and bending moment is built and verified by using loading calibration test data of a delta wing. Comparing and analyzing the results between load prediction and multiple linear regression, it is found that the prediction errors of BP network model are less than 3%. Especially for shear load, the prediction errors by BP model are less than the result by multiple linear regression obviously. So the BP neural network can be applied as an effective engineering method for obtaining the flight load of delta wing.