基于Matlab平台,开发了面向大范围降水空间插值的普通克里金模型—Matlab based ordinary Kriging(MatOK).与已有模型相比,MatOK的主要特点是:(1)在降水空间变异函数计算环节,采用SCE-UA算法拟合理论变异函数;(2)在OK方程组估值环节,引...基于Matlab平台,开发了面向大范围降水空间插值的普通克里金模型—Matlab based ordinary Kriging(MatOK).与已有模型相比,MatOK的主要特点是:(1)在降水空间变异函数计算环节,采用SCE-UA算法拟合理论变异函数;(2)在OK方程组估值环节,引入降水空间发生概率的计算,完善了日等短时间尺度降水空间估值方法;(3)通过对OK方程组的标准化处理,有效提高了模型数值计算的稳定性.同时,将MatOK初步应用于集水面积为83374km2的赣江流域年、月、日降水空间插值中,并重点讨论了MatOK的计算稳定性和计算效率,结果初步说明了采用MatOK进行大范围区域降水空间插值是可行的.展开更多
In the conventional absolute nodal coordinate formulation(ANCF), the model is pre-meshed, the number,distribution and type of elements are unchangeable during the simulation. In addition, the deformations of a flexibl...In the conventional absolute nodal coordinate formulation(ANCF), the model is pre-meshed, the number,distribution and type of elements are unchangeable during the simulation. In addition, the deformations of a flexible body are space-varying and time-varying, one cannot predict when, where, and how the deformations will occur. Therefore, in order to obtain a satisfactory accuracy during the whole simulation, the model is usually densely meshed, but it will result in a loss of computational efficiency. In this study,an adaptive absolute nodal coordinate formulation(AANCF)is proposed to optimize the accuracy and efficiency of flexible dynamics. The movement features of flexible bodies are analyzed, and the conventional and adaptive ANCF methods are compared. Then the adaptive computation strategy is presented. The discretization errors come from the inability of interpolation functions of individual elements to capture the complexity of the exact solution, so the mesh can be adaptively optimized by changing the element sizes or the orders of interpolation functions during dynamic computation. Important issues of AANCF, including error estimation,mesh updating, and performance of the AANCF model, are analyzed and discussed in detail. A theoretical model of a planar AANCF cable is presented, where the strategies of dividing and merging elements are discussed. Moreover, the continuity of dynamic variables is deduced, and the mean factors that affect the continuity are obtained, which is very important for the subsequent continuity optimization. Thesimulation results indicate that the distribution of elements varies with time and space, and the elements are denser in large-deformed domains. The AANCF model improved the computational accuracy and efficiency, but the system energy is discontinuous when the elements are merged. Therefore,a continuity-optimized AANCF model is given based on the previous continuity analysis, the results show that the accuracy and continuity of energy are further improved by the continuity-o展开更多
文摘基于Matlab平台,开发了面向大范围降水空间插值的普通克里金模型—Matlab based ordinary Kriging(MatOK).与已有模型相比,MatOK的主要特点是:(1)在降水空间变异函数计算环节,采用SCE-UA算法拟合理论变异函数;(2)在OK方程组估值环节,引入降水空间发生概率的计算,完善了日等短时间尺度降水空间估值方法;(3)通过对OK方程组的标准化处理,有效提高了模型数值计算的稳定性.同时,将MatOK初步应用于集水面积为83374km2的赣江流域年、月、日降水空间插值中,并重点讨论了MatOK的计算稳定性和计算效率,结果初步说明了采用MatOK进行大范围区域降水空间插值是可行的.
基金supported by the National Basic Research Program of China (Grant 2013CB733004)
文摘In the conventional absolute nodal coordinate formulation(ANCF), the model is pre-meshed, the number,distribution and type of elements are unchangeable during the simulation. In addition, the deformations of a flexible body are space-varying and time-varying, one cannot predict when, where, and how the deformations will occur. Therefore, in order to obtain a satisfactory accuracy during the whole simulation, the model is usually densely meshed, but it will result in a loss of computational efficiency. In this study,an adaptive absolute nodal coordinate formulation(AANCF)is proposed to optimize the accuracy and efficiency of flexible dynamics. The movement features of flexible bodies are analyzed, and the conventional and adaptive ANCF methods are compared. Then the adaptive computation strategy is presented. The discretization errors come from the inability of interpolation functions of individual elements to capture the complexity of the exact solution, so the mesh can be adaptively optimized by changing the element sizes or the orders of interpolation functions during dynamic computation. Important issues of AANCF, including error estimation,mesh updating, and performance of the AANCF model, are analyzed and discussed in detail. A theoretical model of a planar AANCF cable is presented, where the strategies of dividing and merging elements are discussed. Moreover, the continuity of dynamic variables is deduced, and the mean factors that affect the continuity are obtained, which is very important for the subsequent continuity optimization. Thesimulation results indicate that the distribution of elements varies with time and space, and the elements are denser in large-deformed domains. The AANCF model improved the computational accuracy and efficiency, but the system energy is discontinuous when the elements are merged. Therefore,a continuity-optimized AANCF model is given based on the previous continuity analysis, the results show that the accuracy and continuity of energy are further improved by the continuity-o
