Direct Monte Carlo simulation can be used to examine balance calibration system uncertainty. Work on this is largely theoretical in nature. Although this is useful, practical application of findings is of greater value. This paper describes the practical application of the description of calibration-model error as an angular value (C). This value is constant for a given load level. It is shown that C can be effectively determined by finding the coefficients of a trigonometric function that minimises the least square errors of balance dead-weight roll-polar data. The errors are determined using a trigonometric function as the reference. This reference is related to secondary standards in the most direct manner possible so that the relative errors reflect balance accuracy within determinable uncertainty levels. It is further described how multiple roll-polars can be used to generate a "predictor-corrector" balance calibration scheme that simultaneously verifies and improves balance accuracy. This is a significant step towards the quantification of balance "run-time" accuracy data.
展开▼
