Research On On-site Measurement Complex Surface Distribution Planning Based On Bayesian Network

With the rapid development of modern manufacturing industry, On-site Measurement technology has been widely applied in the measurement of complex surface parts. On-site Measurement technology can reduce the enterprise capital investment, and avoid parts generated errors because of secondary clamping and real-time feedback of measurement results and many other advantages; Establishing CAD, CAM and CAI integrated system, to achieve the designprocessing- measuring integration, which conforms to highly integrated requirements of modern manufacturing. On the basic theory research of machine measurement technology has been carried out from all aspects. This paper proceed from two aspects of measured surface extracting and measuring point sampling plan, in-depth discuss and research on some technical problems of On-site Measurement system.This paper introduces the basic theoretical knowledge, and analyzes the On-site Measurement system, which does the groundwork to solve the following key problems:Systematically introduced NURBS theory knowledge, and researched the method to construct,learning model and reasoning method of Bayesian network model. Analyzed and selected the construction steps, parameter learning mode and network inference methods of network model.Complex surfaces On-site measuring system principle, the realization way and the advantages were described in detail. Through of analysis of On-site Measurement system hardware and software, puts forward the key problems that need to be addressed for complex surface On-site Measurement machine, and determine the overall program planning.According to the results of the above analysis demonstrates, I study thoroughly in a number of key technologies issues of On-site Measurement system, which mainly including: the extraction of to be measured surface of CAD model and measuring point sampling planning.(1) The extraction of to be measured surface of CAD model: Through study thoroughly of the structure of IGES files and analysis of IGES file entity record of topological structure, to achieve the visible geometry entity information extraction; This paper puts forward a excellent NURBS surface information and crop information storage method. The main function of the IGES interpreter is to stabilize the NURBS surface information in the IGESV5.3 file.(2) Measuring point sampling planning:Through the analysis of present sampling strategies. It is pointed out that single sampling scheme tends to be less robust problems, and put forward a method of sampling strategy selection based on Bayesian network, which according to the analysis of the differential geometry parts of curvature and surface patches and other information and on the machining accuracy of the workpiece surface requirements analysis, and according to the relationship between differentfactors of influence using Bayesian network to select the optimal sampling strategy. Only in this way can give full play to the advantages at various sampling strategies.