Research Of Transient Characteristics Uncertainty Evalution Method In Digital Pulse Measurement System

Evaluation and expression of measurement uncertainty is a standard for thecalibration of sampling oscilloscope in metrology. That is to say a good evaluationmethod directly determined the quality of uncertainty evaluation. Traditionaluncertainty evaluation method written in GUM (Guide to the expression ofuncertainty in measurement) is linear delivery method, which based on uncertaintytransfer theory, and need to evaluate mathematical model. The mathematical modelis based on Taylor series to approximately determine. However, as more people usethis method, people find there are many deficiencies for linear delivery method.Firstly it is difficulty to process complex mathematical models, embodied in verylarge amount of computation, also because of mathematical model wasapproximately determined, which lead to an inaccurate result. Secondly thecalculation of freedom in class B standard uncertainty with a strong subjectivity,which also lead to the result inaccurate. Third most of the time people want to get thecoverage interval with a specified coverage probability, but not the standarduncertainty. Linear delivery method can directly get standard uncertainty, though wecan get the coverage interval through formula, but this is only established in themeasurement model for the symmetric case, that is to say, we can’t get the coverageinterval through standard uncertainty, when the input model is asymmetry, at thesame time it is difficult to calculate partial derivative.Based on those factors above, people begin to study other methods in additionalto linear delivery method to get measurement uncertainty. Monte Carlo Method(MCM) has proven to be a good method to substitute linear delivery method, whichmake up for the lacks of linear delivery method, that is to say, MCM don’t need toanalog input distribution, don’t need to calculate complex partial derivatives, candirectly get the coverage interval with a specified coverage probability and standarduncertainty. This paper based on experimental system platform, make a depththeoretical research and actual application.Firstly, background and issues related to this topic, as well as related concepts areintroduced briefly. This paper also studies the theory of linear delivery method, at thesame time identify the shortcomings.Secondly, in-depth theoretical study on MCM, from the development of MCM to the basic principles of MCM, as well as some key issues should be noted when useof MCM, such as: determining the distribution of input quantity, generation ofpseudo-random number, method of determining the value of sample size. AdaptiveMCM is an improved method for traditional MCM. This paper also studies adaptiveMCM. This paper studies how to use MCM to validate linear delivery method, andthrough example to prove that the feasibility of this method.Finally, by building experiment platform, this paper complete data acquisition,and get the data that using MCM to calculate the measurement uncertainty. Based onthe data firstly MCM is used to calculate measurement uncertainty with simulationdata. Then MCM is used to calculate measurement uncertainty with really data. Thisexperiment proved that MCM can finish the work of measurement uncertaintyassessment very well, this method expanded measurement uncertainty method.