Research On Evaluation Of Dynamic Measurement Uncertainty Based On Bayesian Theory

Uncertainty is the important parameter of the measurement-result and used to token the dispersal of measured substance, and is a evaluation of the measured substance-result in some range. Every measurement-result must have explaining of uncertainty. At present, the express and evaluation of uncertainty based on GUM which was publicized by ISO and other international organizers in 1993, but it obviates the evaluation of dynamic measurement uncertainty.With the development of technology and science, dynamic measurement is being regarded more and more, and has became the mainstream of the modern measurement. The research on evaluation of dynamic measurement uncertainty is showing its importance.This paper studys one method to evaluate dynamic measurement uncertainty based on Bayesian theory, and chooses different Bayesian models to evaluate for different random process. For exemple, choose Bayesian constant-mean model for ergodic stationary process, and choose Bayesian dynamic linearity model for non-placidity process. This method does not distinguish between A and B's evaluation, and, we throw away front swath and prior information after obtaining posterior probability, which does not effect posterior concluder.Uncertainty principle is applied in the analysis of dynamic measurement system, and this paper studys thoroughly the relationship between six indexes and dynamic characteristic which express magnitude properties and uncertainty of dynamic measurement system, the evaluation methods of uncertainty components caused by each index of magnitude properties are proposed.This paper adopts Monte Carlo method to synthesize uncertainty, and sets up model based on certainty menasurement and samples random for every input by computer through Matlab programme, then calculates sendout and sees as real measurement, which mades the measurement accord fact and the integration infection for synthesized uncertainty.