Uncertainty Evaluation And Misjudgment Risk Estimation For Product Inspection With CMMs

The measurement uncertainty is an important parameter to characterize the quality of measurement results. Evaluating the measurement uncertainty reliably and estimating the misjudgment risks of product inspection based on it, have an important significance for enhancing the reliability of product inspection results and improving the quality and effectiveness of the product inspection work.In this dissertation, the uncertainty evaluation and misjudgment risk estimation for production are studied. The Coordinate Measuring Machines (CMMs), which are widely used in product inspection, are taken as the research object. This research is mainly focused on the issues of uncertainty evaluation for form measurement with CMMs, conformity assessment and misjudgment risk evaluation for product inspection.The optimization technologies for the evaluation of product inspection uncertainty are proposed. The main research contents are listed as follows:The product inspection uncertainty is evaluated through the measurement systems analysis method based on the black box model. After systematically analyzing the uncertainty evaluation method for dimension measurement with CMMs, an emphasis in this dissertation has been given to the uncertainty evaluation for form measurement. The sources of uncertainty in compliance with characteristics of form measurement are proposed. The uncertainty model is establishe, which contains indication error,repeatability, reproducibility, the uncertainty components caused by sampling strategy and probe configuration. Processes and methods for each uncertainty component evaluation are studied, which are different from those in dimension measurement.Based on the measurement uncertainty,the conformity assessment methods and misjudgment risk estimation models are proposed. According to the measurement uncertainty, the conformity assessment interval is divided into conformance zone,non-conformance zone and uncertainty zone. The misjudgment risks caused by the products whose measurement results fall into the uncertainty zone are studied. The misjudgment probability estimation formulas for total inspection are derived respectively based on absolute probability and conditional probability models. Based on the distribution of the measurement uncertainty, the conformity assessment and misjudgment probability calculation methods for the measurement results in total inspection are studied. The misjudgment probabilities in sampling inspection are estimated based on the calculation results in total inspection. The conformity assessment methods for the measurement results in sampling inspection are discussed. The misjudgment probabilities are calculated respectively based on the exhaustive method and the Monte Carlo method.The optimization technologies for the uncertainty evaluation in product inspection are studied. Based on the Bayesian statistical inference, by integrating the historical experiment information and the measured sample information, the value characteristic indices are updated in real time and continuously. Through integrating the statistical controlled production information into the product inspection results, the measurement results and their uncertainties are reestimated, which reduce the evaluation results of the measurement uncertainties significantly. The evaluation methods for the compliance uncertainty and other uncertainties proposed in GPS are discussed, which can provide the basis for the coordination between product design and inspection. In view of the uncertainty components caused by uncertain systematic errors whose influence directions are clear, the concept of unilateral uncertainty is proposed, and the method to calculate combined standard uncertainty on one side is used to reduce the distribution intervals of the measurement results reasonably and effectively.Through the product inspection case analysis, the proposed key technologies of uncertainty evaluation and misjudgment probability calculation for product inspection are applied, processes and methods of product inspection uncertainty evaluation and misjudgment risk estimation are standardized.