| Summary: | 博士 === 國立臺灣科技大學 === 機械工程系 === 95 === The main purpose of this paper is to optimize the design parameters of linear motion guide. The experiment is planned via the Taguchi experiment design method. Through the S/N ratio, grey-based Taguchi methods and analysis of variance, the optimal parameter combination is identified. The design of experiment employs the Taguchi method. Control factors include the preload grade, lubricant viscosity, ball precision grade, and with/without ball cage and with/without flange types. Its quality evaluation indicators comprise base smoothness, horizontal combination error, movement error, thrust value, noise value and motion guide surface roughness. Based on the quality parameters expectation value outputted by the linear motion guide user, current test data is utilized in conjunction with linear motion guide design parameters input and quality parameters output for identification of the optimal parameter combination through the grey-based Taguchi methods. In practice optimization of multiple quality characteristics is generally expected. Due to the fact that the Taguchi method is unable to process multiple quality characteristics directly, the grey relation analysis is employed to deal with multiple quality issues in order to understand what the optimal linear motion guide process condition will be when multiple quality characteristics are taken into consideration simultaneously. The analysis of variance is then conducted to identify the contribution of control factor. Moreover we established the neural network prediction modeling of the linear motion guide. This method has the accurate forecast ability. The analytical method incorporates Genetic Algorithm and neural network in order to find the optimal design parameters. We anticipated the optimal quality characteristics that can better meet the requirement of the linear motion guide user that what have been identified through the grey-based Taguchi methods. This analysis follows to equip it with the ability to predict more accurately. A better linear motion guide design parameters combination can be acquired as a result to enhance the entire linear motion guide process in terms of precision, noise and motion guide surface roughness.
|
|---|
