The dual ball screw feed mechanism (DBSFM) is commonly employed in high-end machining applications. However, its accuracy is challenging to enhance due to variations in the axial stiffness of the dual actuators under load, making geometric error compensation (GEC) difficult. To address this, our research establishes a comprehensive error model that integrates geometric error and load-induced error (LIE) to improve feeding accuracy. The axial stiffness matrix of the DBSFM is identified using motor current information, and an active prediction error matrix is defined based on the geometric error model, milling force model, and axial stiffness matrix. We propose and analyze the active prediction compensation (APC) strategy as a new method to enhance accuracy under load. Experimental validation using a gantry-type machine tool demonstrates that the APC strategy can reduce the positioning error of the DBSFM to ±2 μm under load, representing significant improvements over the established GEC method. Specifically, the APC strategy achieves a 53.07% and 57.04% enhancement in accuracy along the two axes. This work illustrates the effectiveness of the APC strategy in reducing positioning errors and improving feeding accuracy in DBSFM machining processes.
Read full abstract