基于改进布谷鸟算法的热物性参数反演

发布时间：2019-01-05 03:40

【摘要】：热物性参数反演具有重要的理论意义和工程应用价值,可以通过迭代方法或优化算法来数值求解。优化算法主要包括梯度类算法和启发式算法。共轭梯度法因其储存空间和计算量小等优点被广泛使用。共轭梯度法属于局部搜索算法,其对于迭代初始值敏感。作为一个新的启发式算法,布谷鸟算法带有全局搜索的天性,这为反问题的求解提供了一个强有力的工具。针对布谷鸟算法收敛速度慢的问题,将共轭梯度法法引入到布谷鸟算法中形成改进布谷鸟算法,可以高效准确地求解反问题。本文主要基于边界元法和改进布谷鸟算法反演二维热传导问题热物性参数。针对非线性稳态热传导问题,采用Kirchhoff变换将非线性热传导问题转化为线性热传导问题,反演导热系数。针对瞬态热传导问题,考虑不含热源和含有热源的情况,引入一个变换将带热源热传导问题转换为无热源热传导问题,反演热扩散系数。针对非线性瞬态热传导问题,也采用Kirchhoff变换将非线性热传导问题转化为线性热传导问题,反演导热系数。算例讨论了单元数量、测点数量、鸟巢数量、测量误差等对结果的影响。增加单元数量,迭代次数减少,计算结果更精确;增加测点数量,迭代次数增加;增加鸟巢数量,迭代次数减少;随着测量误差的增大,结果精度降低并且迭代次数增多。数值算例验证了改进布谷鸟算法反演热物性参数的准确性和有效性。
[Abstract]:The inversion of thermal properties parameters has important theoretical significance and engineering application value. It can be solved numerically by iterative method or optimization algorithm. The optimization algorithm mainly includes gradient class algorithm and heuristic algorithm. Conjugate gradient method is widely used because of its advantages of less storage space and less computation. The conjugate gradient method belongs to the local search algorithm and is sensitive to the iterative initial value. As a new heuristic algorithm, Cuckoo algorithm has the nature of global search, which provides a powerful tool for solving inverse problems. In order to solve the problem of slow convergence of cuckoo algorithm, the conjugate gradient method is introduced to form an improved cuckoo algorithm, which can solve the inverse problem efficiently and accurately. In this paper, the boundary element method and the improved Cuckoo algorithm are used to retrieve the thermal properties of two-dimensional heat conduction problem. For the problem of nonlinear steady state heat conduction, the nonlinear heat conduction problem is transformed into a linear heat conduction problem by Kirchhoff transform, and the thermal conductivity is retrieved. For the transient heat conduction problem, a transformation is introduced to transform the heat conduction problem with a band heat source into a non-heat source heat conduction problem, and the thermal diffusion coefficient is retrieved. For the nonlinear transient heat conduction problem, the nonlinear heat conduction problem is transformed into a linear heat conduction problem by Kirchhoff transform, and the thermal conductivity is retrieved. The effects of the number of units, the number of measuring points, the number of bird's nests and the measurement errors on the results are discussed. With the increase of the number of units, the number of iterations is reduced, the calculation results are more accurate; the number of measured points is increased, the number of iterations is increased; the number of iterations is reduced by increasing the number of bird nests; with the increase of measurement error, the precision of the results decreases and the number of iterations increases. Numerical examples verify the accuracy and validity of the improved Cuckoo algorithm in inversion of thermal properties.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;O551.3

【参考文献】