【摘要】 为实现高速飞行器承载结构在力热耦合载荷下全场响应的快速求解与精确预测，利用基于多维降阶与深度学习的代理模型技术，提出一种新型的舱体力热耦合响应瞬态重构方法。采用了顺序耦合方法降低力热耦合工况下的计算复杂度；采用了载荷-空间-时间三维降阶技术完成全局瞬态响应场的基模态特征分解，截断选取包含绝大部分样本特征的基模态以线性拟合所有样本；通过深度学习方法搭建了代理模型，从而建立起决定力热载荷输入参数与拟合系数之间的映射关系，实现对结构全局瞬态响应场的快速预测。针对三维舱体承载结构开展的验证表明：该方法能较精确地预示力热耦合载荷下高速飞行器承载结构全局瞬态的应力场与位移场，其中Mises应力场平均绝对误差为0.25 MPa，位移场平均绝对误差为0.0144 mm，且单工况响应重构耗时可控制在0.6 s以内，远低于传统数值计算方法，对后续舱段结构综合强度实时评估及其在轨数字孪生体的构建具有重要意义。更多还原

【Abstract】 To achieve the rapid solution and accurate prediction of full-field response of the bearing structure of high-speed aircraft under force-thermal coupling loads, a novel response reconstruction method based on multi-dimensional order reduction and deep learning is proposed. The sequential coupling method is adopted to reduce the computational complexity under the force-thermal coupling condition. The load-spacetime three-dimensional order reduction technique is adopted to complete the fundamental mode feature decomposition of the global transient response field, and the fundamental mode containing the characteristics of the vast majority of samples is truncated and selected to linearly fit all samples. A surrogate model is built through deep learning methods to establish the mapping relationship between the input parameters determining the force-thermal load and the fitting coefficients, thereby achieving rapid prediction of the global transient response field of the structure. The verification carried out for the three-dimensional cabin bearing structure shows that this method can accurately predict the global transient stress field and displacement field of the bearing structure of high-speed aircraft under the force-thermal coupling loads. The mean absolute error of the Mises stress field is 0.25 MPa, and the mean absolute error of the displacement field is 0.0144 mm. Moreover, the response reconstruction time for a single working condition can be controlled within 0.6 seconds, which is much lower than that of traditional numerical calculation methods. It is of great significance for the subsequent comprehensive evaluation and optimization design of the structural strength.更多还原