为优化燃气导流器表面热环境,对某型车载单面导流器的热环境影响因素进行数值计算分析,确定了导流器高度、导流器折转半径和导流器背板与发射架距离为重要结构参数,它们与导流器表面温度之间存在复杂非线性关系。基于BP神经网络和RBF径向基神经网络建立相应代理模型进行对比,选择了决定系数更高、稳定性更好的RBF径向基神经网络代理模型进行导流器型面优化,并对优化结果进行了流场计算分析,结果表明,在满足导流器机动性要求的前提下,优化后的导流器表面最大温度降低了约400 K,热环境得到了明显改善,且导流效果更佳。优化误差在1% 之内,验证了优化的有效性和准确性,为今后燃气导流器型面的设计及优化提供了参考。

To optimize the thermal environment of gas deflector,the thermal environment influencing factors of avehicle-mounted single-sided deflector are analyzed by numerical calculation,which proves that the height of the deflector, the turning radiusof the deflector and the distance from the deflector backplane to the launch frame are important parameters, having complex nonlinear relationship with the surface temperature of deflector.Based on the comparison of BP surrogate mathematic model and RBF surrogate mathematic model,the RBF surrogate mathematic model with higher determination coefficient and better stability is selected to optimize the deflector. The flow field numerical calculation of optimization results is carried out,which shows that, on the premise of meeting the maneuverability requirements, the optimized max surface temperature of the deflector is reduced by about 400 K, the thermal environment is significantly improved and the diversion effect is better.The optimization error is within 1%, verified the effectiveness and accuracy of the optimization, which provides a reference for the design and optimization of the gas deflector in the future.