正交约束下的D—最优试验设计  

Test design with orthogonal restrain based on D-optimal

作　　者：王超[1] 王跃钢[1] 

机构地区：[1]第二炮兵工程学院,西安710025

出　　处：《中国惯性技术学报》2011年第2期248-252,共5页Journal of Chinese Inertial Technology

基　　金：国防973项目(61355020301)

摘　　要：针对一类因素间具有正交约束的回归模型的试验设计问题,以D—最优为准则,以加速度计试验设计为背景,从试验点的均衡分布原则出发,以不同试验环境下合加速度大小为依据,给出了不同环境下球内接正多面体的试验方法,通过进一步计算和深入分析表明,单球面时为避免信息矩阵降秩应采用不同安装方式,降低模型维数以达到试验目的;两球面下试验设计的优劣与两个半径的大小有关,半径越大计划越优,而且信息矩阵的秩最大时两个球半径之间呈近似线性关系。试验结果的优劣还受到初始安装位置的影响,为了达到最优化,应使合加速度方向与加速度计的某个轴同向。Based on the D-optimal test design guide rule,a kind of test design problem that has orthogonal restrain between elements is studied.The error model of accelerometer is employed to embody the problem.The Spheroid Inscribed Regular Polyhedron method is presented,and the conclusion is given by rigorous calculation and deep analysis.To avoid singular information matrix,a suitable mounting scheme should be adopted from various alternatives.The bigger of the different radiuses under the double sphere situation,the more precise the result.The initial mounting status could also influence the quality of test results.To get the best quality,the direction of combined acceleration should be parallel to one of the accelerometer axes.

关 键 词：正交约束 D—最优 加速度计 正多面体 试验设计 

分 类 号：U666.1[交通运输工程—船舶及航道工程]