Abstract: A two-dimensional physical and mathematical transient heat transfer model for the pipeline with irregular inner geometry boundary is built.The identification problem of pipeline inner boundary is solved by using finite element method and conjugate gradient method(CGM)-based inverse algorithm.The differences of sensitivities of the transient and steady pipeline outer surface temperature to the inner boundary change are analyzed and certified by a series of numerical simulation experiments,in which the influences of temperature measurement errors and initial inner boundary guesses are tested.It is found that the temperature measurement errors and initial inner boundary guesses have no obvious effect on the pipeline inner boundary identification results when the infrared inspection is carried out in the transient heat transfer state,while the accuracy of the boundary identification results is poor with temperature measurement errors when the infrared inspection is carried out in the steady heat transfer state.Therefore,in the engineering application,in order to identify the inner geometry boundary accurately,the infrared inspection in the transient heat transfer state is more appropriate than that in steady heat transfer state.

Key words: boundary identification, inverse heat conduction problem, finite element method, conjugate gradient method, infrared inspection

摘要： 建立了带有不规则腐蚀内边界的管道二维瞬态传热模型,基于有限元法和共轭梯度法的导热反问题求解方法对管道内边界识别问题进行了研究,比较分析了管道外表面稳瞬态温度对内边界变化敏感程度的差异性,指出内边界瞬态检测识别比稳态检测识别更具优越性并利用数值试验进行了验证。同时,考察了测温误差、初始假设等因素对管道内边界瞬态检测识别结果的影响。在较大测温误差的情况下,采用瞬态检测条件从不同初始假设值出发都能准确地反演识别出管道内边界腐蚀后的几何形状,从而证明了算法的有效性和稳定性,表明了红外瞬态检测定量识别管道内边界的可行性。

关键词: 边界识别, 导热反问题, 有限元法, 共轭梯度法, 红外检测