Research And Development Of 3D CAD-Based Inspection Planning System For Intelligent Coordinate Measuring Machines

The manufacture of high-quality products depends not only on high-precision processing equipment, also requires high-precision inspection equipment. In order to improve the competitiveness of enterprises and guarantee the quality of products, it is essential to inspect the products or parts, while the development of rational inspection plan for the parts before the right implementation of inspection would be particularly important. The traditional manual way to develop parts' inspection plan is time-consuming and nonstandard, and the accuracy and consistency is very difficult to guarantee. As the rapid development of China's machinery manufacturing industry, advanced manufacturing technology and computer technology, the quality control of products has become increasingly demanding, and the high-speed inspection with high-precision, automation and intelligence has become an irresistible trend. The speed and accuracy of the traditional approaches in dimensional inspection have been serious bottlenecks of the modern production line efficiency. To overcome these bottlenecks, CAD-based computer-aided inspection planning system was proposed with the coordinate measuring machines (CMM) being the key device, and its basic task is to solve the problem of that how to generate the inspection procedures and inspection commands for CMM automatically, with the part's information provided by its CAD model and the measurement-based knowledge.In the dissertation, the general structure of the inspection planning system for intelligent coordinate measuring machines based on Open CASCADE is designed in detail, and an system architecture which consists of four layers, i.e. interface layer, function layer, service layer and data layer, is proposed, the support relationship between the CAD system Open CASCADE and the CAIP system is also established. In addition, the realization interface and the work flow of this system are designed, and the framework of inspection planning algorithm's principle with planning service in system's service layer is proposed.Based on the DMIS formats for nominal features definition, the dissertation analyzes and determines the contents of the geometric information that should be extracted from the eight basic measurement features, i.e. point, line, circle, ellipse, plane, cylinder, cone and sphere, and constructs the corresponding data structures. Two ways for information extraction are proposed based on the shape type, i.e. inquiring geometric information directly from the shape and from the least squares fitting shape which is fitted by the points generated by the parametric discretization of the original shape, expanding the scope of the shape type suited for geometric information extraction. A concept of "visible half-space of point on surface" is proposed, and based on that, the dissertation solves the problems to determine automatically whether the surface normal vector points to inside or outside, and whether the feature property is inside or outside for circle, ellipse, cylinder, cone and sphere features.In the dissertation, four requirements for sampling points' distribution are summarized, and aiming at the shortages and limitations of some conventional sampling strategies on general surfaces, a sampling method with step's adaptive subdivision is proposed, providing the solution of sampling planning on general surfaces. This sampling method is then discussed in two ways, i.e. edge-based sampling and face-based sampling for geometric elements measurement, and the detailed sampling algorithms are presented. The experimental results show that the method has strong adaptability for general surfaces with different complexity, and can generate sampling points that scattered on the whole surface as evenly as possible and located apart from the discontinuous areas of the surface.For the inspection path planning, the associated issues such as inspection sequence planning, collision detection and collision avoidance are studied in the dissertation. A genetic algorithm optimization method based on row-and-column planning is proposed to realize the inspection sequence planning, in which the initial sequence is got by planning the sequence of measurement points in row-and-column way according to the values of each point's weight sum of parameters, and then optimized by genetic algorithm. A collision detection method based on intersection test with traversed faces is proposed, in which the collision is detected by B-rep intersection test between the probe swept-volume surfaces and each traversed face of the part, with axis-aligned bounding box (AABB) filtering to reduce the test complexity. A collision avoidance method taking the outer-circle obstacles and general obstacles into account together is proposed, in which obstacle-avoidance planning is implemented firstly aiming at the possible outer-circle features, and then at other collisions, for which the obstacle-avoiding points are set with the heuristic rules and a designed expanded AABB. The computer implementations results show that the methods proposed in the dissertation can generate efficient optimum path in a short time and avoid collision effectively.The dissertation designs the realization process of automatic programming based on Dimensional Measuring Interface Standard (DMIS). An automatic programming method based on variable's setting and filling is proposed, in which the string formats for DMIS statements and the keywords and variables in the strings are set, according to the corresponding functions in each system's function module and the DMIS formats, the system will fill up the variables with the obtained planning information so as to realize the automatic programming. Then, the string formats and their keywords and variables that set for DMIS statements generation are discussed in detail, which provide detail bases for DMIS-based programming with the similar CAIP system development, and have certain value of reference and guidance.At the end of the dissertation, experiments are carried out with the related system operation and the inspection planning for a work-piece with regular features, the reliability of this system for inspection planning and the feasibility of the related planning algorithms are verified, and the correctness and reliability of the DMIS-based automatic programming method are also verified.The 3D CAD-based inspection planning system studied and developed in this dissertation for intelligent coordinate measuring machines can extract inspection information of the part's model from the CAD design result by the corresponding CAD interface, with the realization of the integration between this system and CAD system; and then generate measurement points and inspection path automatically in accordance with the corresponding inspection task, with collision detection and collision avoidance for the generated path. After no collision is confirmed by the simulation, the DMIS measurement program is generated and sent to CMM. Finally, the inspection task will be implemented by CMM and the inspection result will be got and analyzed. The whole process is done with high automation and high intelligence.