Reconfigurable machining systems (RMS) provide exact amount of flexibility and capacity exactly when needed by reconfiguring the machining system in response to product changes within a pre-defined part family. In order to meet short reconfiguration and ramp-up time requirements. Reconfigurable Machine Tools (RMTs) are designed to quickly and accurately reposition, re-orient, and replace various modules such as a spindles, slides, tool-support and work-support sub-systems. One of the key challenges in the design of RMTs is to ensure ease of reconfiguration without any loss of accuracy. Resulting accuracy depends on geometry, assembly and random motion-related errors. Error also depends on the manner in which various motions and modules are nested in a machine tool. It is therefore critical to evaluate alternate machine tool designs early in the design process in terms of resulting accuracy (tool-tip error) in each of their intended configurations.; This research presents a systematic methodology to estimate the accuracy of candidate designs early in the machine tool design process thereby avoiding extensive trial-and-error based error compensation after the machine is fabricated. A generalized mathematical framework for prediction of tool tip errors in multi-axis machine tools is developed using screw theory. Using the same mathematical framework, a method of determining the nature and extent of control compensation required to improve the machine tool accuracy is also developed. In contrast to conventional homogeneous transformation matrix (HTM), employed by other researchers for error prediction, Screw Kinematics is used in this research since it offers several design advantages including: (i) accurate modeling and solution of complex configurations with rotational axes and (ii) modular and functional representation of motions as screws in a global reference frame enabling quick reconstruction of the model as the machine is reconfigured (iii) Automatic generation and simple solution of inverse kinematics, (iv) Kinestatic Filtering method for estimation of compensatable portion of predicted errors. Deterministic and probabilistic methods are used to compute the total tool tip error. Design examples highlight the benefits of the error prediction methodology and the error compensation strategy.
展开▼
