Straightness Error Assessment Model of the Linear Axis of Machine Tool Based on Data-Driven Method

In batch assembly, fast and accurate assessment of MT-LA straightness error is significant important for controlling of MT-LA assembly quality. In this study, in order to construct MT-LA straightness error assessment model, a data-driven method based on the bootstrap resampling approach improved fast correlation based filter (BR-FCBF) algorithm and genetic algorithm optimized multi-class support vector machine (GA-MSVM) algorithm is proposed. Firstly, the BR-FCBF algorithm is used to select the key assembly parameters that affect the straightness error. Secondly, the GA-MSVM algorithm is applied to construct the straightness error assessment model. Finally, the assembly-related data collected on a MT-LA assembly workshop is used to verify the proposed method. The experimental results show that the constructed straightness error assessment model has shown good performance in straightness error assessment.