In-process forecasting compensation control of machining center volumetric error including workpiece inspection

The quality of a manufactured part can be substantially improved without costly machine tool redesign or remanufacture by compensation. Current off-line methods of machine error compensation cannot improve the accuracy of a machine beyond a certain limit. And, this limit can only be reached if the reproducibility of a machine's movement is improved by both real-time machine error compensation and in-process workpiece inspection.; To improve workpiece accuracy without costly machine tool structural changes, the machine control system must be made to more accurately position the cutting tool relative to the workpiece within the work volume of the machining center. In this thesis a strategy is proposed for improving the accuracy of a multi-axis machining center by on-line measurement and compensation for the machine's deterministic and stochastic errors in positioning. An on-line error forecasting compensation control system is proposed to enhance the real-time positioning accuracy of the machining center.; A volumetric error model is developed which uses the axis encoder zero readings as a reference rather than using the structural elements of the machine. Based on this volumetric error model, the total volumetric error map for the multiaxis machining center is determined. From the volumetric error map the error vector can be obtained for any position of the machine axes within the entire work space of the machining center.; Recursive autoregressive (AR) models, along each of the three encoder referenced axes, are used to fit the 21 volumetric error components in real time. One-step-ahead forecasting is used for on-line compensation. From the on-line forecasting compensation control simulation results from three typical cutting paths, the error was reduced between 86-93%.; Even with both the in-process measurement of machine error and compensation for this error, volumetric error from sources that are not covered by the on-line measuring system will still remain. In order to reduce the error further, on-machine workpiece inspection was added to the overall compensation strategy. Simulation shows an additional 74% reduction in workpiece error is possible with the use of time series techniques applied to the inspection data.