针对工程中的反问题,以POD(最佳正交分解,Proper Orthogonal Decomposition)技术为基础设计了一种反问题的低阶计算模型,通过采集到的有限输出结果,仅采用简单的非线性规划与插值法即可求解该模型,大幅度地降低了工程中的反问题的求解难度与计算时间.由于模型求解过程不涉及原物理问题的控制方程,因此,该模型在实际中具有一定通用性.以圆管内强制对流换热问题为应用实例,结果表明,基于一些热电偶测量的温度数据,低阶反问题计算模型能够较精确地反算出未知的流体进口温度与圆管外壁热流密度,与真实值相比仅相差1.0％左右.此外,与经典的共轭梯度法相比,低阶反问题计算模型的计算速度可提高1200倍以上.“,”This paper proposes a reduced-order computational model for reverse problem mainly based on POD (Proper Orthogonal Decomposition),through finite output data collected in advance,the model is solved only using non-linear programming and interpolation methods,which considerably reduces the difficulty and computational time of reverse problems in engineering.The governing equations of the original problem are not considered in the model,which suggests the universality of the reduced-order reverse problem computational model.The reduced-order reverse problem computational model is approved on a forced convection and heat transfer problem in a circular pipe,the results show that the reduced-order reserve problem computational model enables us to accurately calculate the unknown inlet temperature and wall heat flux based on some temperature data measured by thermocouples,where the error between the computational results and the actual results is only about 1.0％.Besides,it is quite fast,where the achieved increase in calculation speed is more than 1200 times compared with the classical conjugate gradient methodology.