一种小样本虚拟仪器测量不确定度评定新方法.pdfVIP

29　4 　　　　　　　　　 　Vol.29, №4 2008 　9 ACTA 　METROLOGICA　SINICA 　 September, 2008 王中宇, 　葛乐矣, 　杨文平, 　李　萌 ( , 　100083) :, , (RBF) , ;, , GUM 。 , , , , 。 :;;;;RBF ; :TB9　　　　:A　　　　:1000-11 8(2008)04-0387-06 A Novel Measurement Uncertainty Evaluation Method for Small Sample Virtual Instrument WANG Zhong-yu, 　GE Le-yi, 　YANG Wen-ping, 　LI Meng (Key Laboratory of Precision Opto-mechatronics Techonology of the Ministry of Education, School of Instrumentation Science Opto-Electronics Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China) Abstract:There are three problems should be solved in the case of small sample virtual instrument measurement: establishing measurement model, calculating transfer coefficients and correlation coefficients.ISO GUM (Guide to the expression of uncertainty in measurement)method is used to evaluate the final uncertainty.To analyze the random errors of the virtual instrument measurement chain, composing mechanism of a general virtual instrument measurement system is discussed in the first.To establish the model of small sample virtual instrument measurement, a radial basis function (RBF)neural network is employed, and then the transfer coefficients are calculated by a difference equation for all error sources.To calculate the correlation coefficients for all error sources, a new method based on grey system theory is advanced.In the end, according to a special example analysis, the uncertainty evaluation results of the proposed method and Monte Carlo method are in good agreements with each other, and the validity of this method is also discribed. Key words:Metrology;Virtual instrument;Uncertainty evaluation;Small sample;RBF