The Comprehensive Research Of Numerical Control System Reliability Evaluation Scheme Based On Likelihood Ratio Method

NC machining industry is always extremely important in the modern manufacturing industry.For it’s relatively high cost, to numerical control machine and numerical control systemexperiment, it is particularly important to ensure and test its reliability. In terms of the currentsituation, the numerical control system of the assessment is diverse. The method of differentcriteria of home and abroad. The authority of this to evaluate the reliability of the numericalcontrol system had a serious impact. Reality requires us to devote more energy to research andpromote the development of reliability evaluation scheme. This paper adopted itself has excellentfeatures of the reliability evaluation of the likelihood ratio theory as the final interval statisticsmethod. First of all, theoretically it prove the superior performance of the likelihood ratio testand there may be insufficient, to make a supplement to make up.According to the different types of data to select different estimation methods, looking forthe meet must be reliably point estimate and interval estimates.Function for failure to obey and exponential distribution data, according to the index distributionfunction in the form of seeking failure number and the cut-off time is derived, the relationshipbetween the sample. The nature of the intermediate process using poisson distribution,Function for weibull distribution types of failure data, the data from a simple screening,selection of a suitable point estimation method, the first point estimation of the parameters.Forsmall amount of data samples, using the least squares estimate, other cases, adopts maximumlikelihood estimation. On the basis of the point estimates obtained, using the theory of likelihoodratio for interval estimation of the parameters. In this process, first of all, the basic principle ofusing the likelihood ratio, the relation between different parameters in the structure of the todetermine one of them, by using iterative way to look for other parameter estimates. Secondly,comparing the different parameter estimation, select one of the maximum and minimum valuesas an interval estimation of two endpoints. Then, from the likelihood ratio in the originalequation for different iterative equation, a similar iteration and value respectively, comparevarious iterative convergence speed and stability, and to select one of the best performers as thefinal iteration equation. Then to the combination of the likelihood ratio and the maximumentropy theory made some prospects.Finally, the limitation of the theory of likelihood ratio are discussed and possiblemodification, points out that the likelihood ratio test of different types of data, there is a certainerror.