Research On Reliability Modeling Technology Of CNC Machine Tool And Its Application

CNC machine tool is the key fundamental equipment of modern manufacturing technology and equipment manufacturing industry, whose quality, modern level and ownership are the important symbols of industrial level in one country. Reliability, as a significant index to estimate product quality, has been the main bottleneck of CNC machine tool development in China. Reliability problems are widespread throughout every control link of product life cycle and its quality management process, and the correct reliability modeling is the prerequisite and basis for realization of reliability analysis and enhancement, which has important research significance.Based on the support of national major scientific technological special project “The high-end CNC grinding machine tool reliability enhancement project”, the following work is done from in-depth research on CNC machine tool reliability modeling and its application.①For considering the corresponding relationship of auxiliary systems and main systems of CNC machine tool, the Reliability Modeling Technology of Superelement is proposed, which embeds auxiliary systems into reliability model of main systems with the form of superelement, and a dynamic reliability analysis model of CNC machine subsystems based on Markov process is established, which reduces the system state space. Based on this, assuming that each main system is independent, the steady reliability index of CNC machine tool is calculated with Kronecker method. The modeling result is the basis of reliability design.②A reliability modeling and evaluation method based on mixture power law process model to solve reliability estimation of multiple CNC machine tool with minimal maintenance is proposed, and the model is classified into three categories on the basis of parameter similarities and differences. Trend test for different truncated failure data is performed to ensure that the application of mixture power law model is established correctly. According to the data homogeneity test result from two-step-test which is Common Beta Test and Common Lambda Test, the corresponding model is selected. Parameter estimation is solved by Expectation Maximization algorithm, and the goodness-fit-test is checked by Kolmogorov-Smirnov test. Finally, the reliability indexes of CNC machine tool are assessed based on the mixture power law process.③To establish an appropriate reliability model and make a reasonable preventive maintenance cycle of CNC machine tool under imperfect situation, the generalized proportional intensities models, including preventive maintenance factor and corrective maintenance factor, are proposed. The models for failure and maintenance process of machines based on three baseline intensity functions are built, the model parameters are evaluated and the goodness of fit is performed. Based on the selected optimal generalized proportional intensities model, an optimization model of the preventive maintenance cycle decision-making of CNC machine tool is presented, with the maintenance cost per unit time as the objective function and the reliability as the constraint.