An Investigation of Various Methods for Evaluating the Measurement Uncertainty

This study presents an assessment of various techniques for evaluating the measurement uncertainty that are internationally accepted to fulfil the requirements of metrological and various industrial applications. A standard document, the Guide to the Expression of Uncertainty in Measurement (GUM) describes the elementary phases that are widely adapted for measurement uncertainty. Moreover, the prescribed technique is based on law of propagation of uncertainty (LPU). In this, repeated measurements are taken into account and the measurement procedure is performed using statistical model and sampling distribution. It gives basic model for estimating uncertainty of measurement requiring a fair amount of knowledge about measurand. Also, it provides best estimate of a quantity under measurement to calculate standard uncertainty related with it. An alternative method, named as Monte Carlo Method (MCM) has also been described that assign suitable probability density function (PDF) to input quantities and output quantity. The approach provides reduction in mathematical calculations and efforts. In Bayesian approach, prior distribution is calculated from the past information and posterior distribution is the integration of prior distribution and updated sample data. It uses different interpretations of probability as compared to other methods. Using Bayes' theorem, it is convenient to develop model for uncertainty of measurement as compared to other methods. Among different evaluation methods of Bayesian approach, maximum entropy prior gives more reliable and practical results.