针对超声速流场智能重构方法存在的复杂波系结构特征丢失、无法有效捕捉非定常流场的时间演化特性,以及共同导致的无法准确辨识激波串前缘位置(shock train leading edge,STLE)等问题,提出基于组合式细节特征增强的神经网络模型。基于...针对超声速流场智能重构方法存在的复杂波系结构特征丢失、无法有效捕捉非定常流场的时间演化特性,以及共同导致的无法准确辨识激波串前缘位置(shock train leading edge,STLE)等问题,提出基于组合式细节特征增强的神经网络模型。基于稀疏压力数据实现密度梯度场的高精度预测,模型通过多层卷积网络串联建立流场的主要波系结构特征,利用残差网络通过跳跃连接将不同尺度感受野的特征进行融合,增强重构流场的细节特征表达能力。基于冲压发动机数值模拟计算构建的数据集进行验证,结果显示,与多层卷积神经网络相比,该方法在整个测试集上的平均峰值信噪比提升了9.5%。重构流场的STLE与数值计算结果高度吻合,进一步证明了所提方法的有效性。展开更多
This study investigates the performance boundaries of ramjet and scramjet engines fueled by boronbased propellant through full-scale engine modeling and three-dimensional computational fluid dynamics simulations.Resul...This study investigates the performance boundaries of ramjet and scramjet engines fueled by boronbased propellant through full-scale engine modeling and three-dimensional computational fluid dynamics simulations.Results show that the performance boundary between ramjets and scramjets occurs near Mach 7.Specifically,at Mach 6,the ramjet exhibits a 1290 m/s higher specific impulse than the scramjet;however,at Mach 7,their performance becomes comparable.The ramjet's higher static temperature promotes boron particle vaporization and B_(2)O_(2) dissociation,limiting the total temperature increase,unlike in scramjets.The boron vapor mass fraction significantly impacts this temperature difference,with ramjets exhibiting values 8.5 and 3.9 times higher than scramjets at Mach 6 and Mach 7,respectively.Despite lower total temperatures,ramjets achieve more efficient boron combustion due to the combined effects of higher pressures and longer particle residence times.These findings offer valuable insights for engine designers in selecting ramjet or scramjet configurations for boron-fueled propulsion systems.展开更多
文摘针对超声速流场智能重构方法存在的复杂波系结构特征丢失、无法有效捕捉非定常流场的时间演化特性,以及共同导致的无法准确辨识激波串前缘位置(shock train leading edge,STLE)等问题,提出基于组合式细节特征增强的神经网络模型。基于稀疏压力数据实现密度梯度场的高精度预测,模型通过多层卷积网络串联建立流场的主要波系结构特征,利用残差网络通过跳跃连接将不同尺度感受野的特征进行融合,增强重构流场的细节特征表达能力。基于冲压发动机数值模拟计算构建的数据集进行验证,结果显示,与多层卷积神经网络相比,该方法在整个测试集上的平均峰值信噪比提升了9.5%。重构流场的STLE与数值计算结果高度吻合,进一步证明了所提方法的有效性。
基金funded by the National Natural Science Foundation of China(Grant No.U21B2086)。
文摘This study investigates the performance boundaries of ramjet and scramjet engines fueled by boronbased propellant through full-scale engine modeling and three-dimensional computational fluid dynamics simulations.Results show that the performance boundary between ramjets and scramjets occurs near Mach 7.Specifically,at Mach 6,the ramjet exhibits a 1290 m/s higher specific impulse than the scramjet;however,at Mach 7,their performance becomes comparable.The ramjet's higher static temperature promotes boron particle vaporization and B_(2)O_(2) dissociation,limiting the total temperature increase,unlike in scramjets.The boron vapor mass fraction significantly impacts this temperature difference,with ramjets exhibiting values 8.5 and 3.9 times higher than scramjets at Mach 6 and Mach 7,respectively.Despite lower total temperatures,ramjets achieve more efficient boron combustion due to the combined effects of higher pressures and longer particle residence times.These findings offer valuable insights for engine designers in selecting ramjet or scramjet configurations for boron-fueled propulsion systems.
