论文部分内容阅读
变异函数套合的合理性是Kriging进行有效预测的根本保证.针对现有方法的线性处理和人为假设所导致的套合结果的不确定,提出了一种统一的套合方法,该方法通过径向变异率和轴向变异率的定义来定量刻画变异函数空间变异特征.在此基础上,将套合过程分解为宏观变换和微观计算两部分.前者用来解决尺度效应,后者则采用积分的思路来实现变异函数的套合计算,最终构建出只与欧氏坐标变量x,y,z有关的套合模型.结果表明:该方法普适性强并满足Krig-ing方法中的非负定条件,其合理性和有效性通过交叉验证实验得到了证实.
The rationality of the variogram fitting is the fundamental guarantee of Kriging’s effective prediction.Aiming at the uncertainty of the fitting result caused by the linear processing and the artificial hypothesis of the existing method, a unified method of nesting is proposed, Quantitatively characterize the spatial variability of the variogram based on the definition of the rate of variability and the rate of axial variability.On this basis, the process of nesting is decomposed into two parts: the macroscopic transformation and the microcosmic calculation.The former is used to solve the scale effect while the latter is to use the integral , And the fitting model of variation function is constructed, and finally the construction model is established, which is only related to the Euclidean coordinates variables x, y, z.The results show that this method is universal and satisfies the non-negative Set conditions, its rationality and effectiveness through cross-validation experiments have been confirmed.