On-Machine Measurement Method And Experimental Research For Manufacturing Of Complex Surface

The components with complex surfaces are widly used in the areas of aerospace, energy and transport, precision mould and etc. and the more excellent manufacturing technology is demand on their fabrication accuracy, efficiency, and reliability. With the developments and applications of NC machine with high accuracy, high precise probe, superpower data processing and transmission technology, and powerful algorithm software, it not only is posible to realize on-machine measurement of the components with complex surfaces through integrating the measurement into NC machining process of the components, but also is an effective means to improve manufacturing accuracy and efficiency of the components. In this thesis, the methods of surface measurement and accuracy assessment, based on on-machine measurement system, are studied theoretically and experimentally. Moreover, it is investigated to realize the integration of machining-measuring surfaces, combining on-machine measurement technology with machining technics of surface.Parametric error model of machine tool is studied in this thesis. After the topology structure of a five-axis NC machine is equivalent to a multi-body system, a model of spatial geometric error of the machine is created based on product-of-exponential formula (POE formula). By means of this model, the law transmiting error from each axis to end effector of the machine can be normally described and also the error of each axis can be inversely solved from measured pose of the end effector. For validating the model, spatial geometric errors of the machine are measured with laser interferometer.Methods of accuracy analysis and error compensation for touch trigger probe are studied. On the basis of analysis on the errors from touch trigger probe of on-machine measurement system, a method to correct the offset of probe is given, and the actual functional radius of probe is calibrated by means of a calibration ball. Oriented to practical measurement task, a probe radius compensation method is given. The method realized pointwise compensation for the errors by mens of establishing the mapping between measured points and calibration surface after the normal vector information of measued surface are discreted based on bilinear interpolation technique. The verification experiments of probe radius compensation show that the accuracy by this method, compared to the method provided by the commercial software, can be much improved. In order to evaluate the extent proximal between the results measured on-machine and real measured surface, considering the influences of instrument and method on measured results, the task specific uncertainty evaluation for on-machine measurement is studied. In the thesis a virtual machine model is created by Monte Carlo Method. According to the error distributions of machine tool and touch probe, the uncertainty oriented to measurement task on-machine is evaluated by means of Multi-measurement Method and Monte Carlo Virtual Method respectively. Then evaluation experiments are done by virtual machine technology respectively according to least square and minimum zone criterions for form error evaluation of surface.In view of the characteristics that contact measurement has higher accuracy and laser measurement has higher efficiency, multiple measurement on-machine is carried out respectively by touch and laser probes, then data fusion and surface localization technologies are studied. Through three procedures i.e. firstly matching, secondly partitioning, and thirdly reducing and optimizing for measured point clouds, thus a mehtod fusing data with high and low accuracies is given. Simulation results show that fused points cloud can describe the measured surface more accurately. Then a surface localization process is done for the fused point cloud based on distance function method, and characteristics of localization algorithm are analysed based on Monte Carlo Virtual Method. Finally, uncertainty of the localization algorithm using Bootstrap methods is assessed for a given set of real measued data.Using the methods given in this thesis, the integration of on-machine measurement technology and machining process is realized. In allusion to the problems in machining process such as raw material position, machining allowance uniform, workpiece reposition and machined error compensation, the on-machine measurement technology is added into surface machining process so the integration of machining-measuring-compensation machining comes into being. Meantime, the analysis and calculation methods in the thesis are also integrated into the software package of on-machine measurement system. The real case given in the thesis shows that on-machine measurement technology is an effective method for machining-measuring integration.