基于多项式混沌理论的不确定度评定与分析

Measurement Uncertainty Evaluation and Analysis Based on Polynomial Chaos Approach

针对基于ISOGUM提供的测量不确定度评估方法所存在的局限，提出了采用多项式混沌方法进行测量结果不确定度评估的新方法。分析了GUM测量不确定度评估方法所隐含的假设条件，给出了多项式混沌方法的测量不确定度评估原理。通过GUM与多项式混沌方法扩展不确定度包含因子估计误差的数值计算表明，当随机变量的PDF不满足高斯分布时，多项式混沌估计方法所获得的扩展不确定度更能反映实际情况。

Deal with the limitations of measurement uncertainty evaluation method based on ISO GUM, A new approach to the evaluation of measurement uncertainty based on the polynomial chaos theory is presented. The main assumptions behind the measurement uncertainty propagation based on the GUM is analyzed. The measurement uncertainty evaluation used the PCT approach is provided. By the GUM and PCT method, the relative absolute difference between the confidence obtained in the two cases and the true value is computed. It is show that in case of a distribution very different from the Gaussian, the polynomial approach leads to results very close to the ideal case.