Analysis of three-dimensional grids: Five-point cubes and curvature coefficients

A method for estimating curvature coefficients on five measurements in prismatic array is illustrated with non-monotonic data. The method turns on repositioning the data from the prismatic array to the Latin square array. A polynomial interpolating equation is applied to the latter design to estimate surrogate numbers for data not originally present in the prismatic array. The five data and four surrogate numbers are then repositioned in the original nine-point cube. The cube is represented by another polynomial in order to estimate the three curvature coefficients. The method is capable of yielding estimates that are within an order of magnitude of the true values. (c) 2005 Elsevier Inc. All rights reserved.