Research On Adaptive NURBS Interpolation And Implementation

NURBS interpolation is an advanced CNC machining technology with high-precision and high-efficiency, which has been one of the most important functions of high performance CNC systems, as well as the research focus and challenge in the field of CNC interpolation. The dissertation is supported by the Science and Technology Project of Zhejiang Province-"Control system of high performance CNC machines" (no.2006C11067), researching on the NURBS interpolation and correlative technologies oriented to 3-axis machining. The innovation is developed both in theories and methods based on the existing achievement in order to achieve high precision, high efficiency and adaptive capability during NURBS interpolation. The study includes several parts about the theme:NURBS fitting from tiny NC codes, NURBS adaptive interpolation, error compensation based on the interpolated data, and the implementation and experiments based on DSP. The main contents and research results can be concluded as following.The predicting method with proportion corrector is adopted to avoid derivation calculus. The main calculation tasks in NURBS interpolation are analyed, then algorithms for parameter calculation with appointed feed speed and algorithms for coordinate calculation with appointed parameter are studied. In order to simplify the calculation process, the two mentioned algorithms are replaced by predictor-corrector technique with a proportion correction method, which is validated in the simulation.The optimized knots combination strategy (OKCS) is proposed to improve the precision in NURBS fitting. In order to solve the problems brought by tiny codes machining, NURBS curve is adopted to fit numerous tiny codes from CAM within the tolerance, and then it can be interpolated in the form of NURBS. As the knot vector of the NURBS is always pre-determined before fitting, the influence on the ultima precision by the knot vector can not been diminished during the fitting process. Therefore, the least squares NURBS fitting method based on the optimized knots combination strategy is proposed to solve this problem. Furthermore, to achieve higher precision, two different adjustment methods can be applied. The algorithm simulation is implemented compared with existing algorithms.Axis-based look-ahead NURBS interpolation (ALANI) is proposed to improve the adaptability, as well as take the kinematic parameters into consideration. The algorithm can calculate the curve ahead the interpolation, which guarantees the chord error, acceleration and jerk of each axis during the interpolation period. The checking of contour precision and the drive capability is integrated in the algorithm harmoniously. Comparison simulation is also realized to illustrate the advantage of ALANI.A robust strategy for solve the dangerous points is proposed based on the kinematics, which can deal with any kinds of dangerous points through the uniform method.7 paramenters are strictly confined with this strategy. On the other side, the tail phenomenon of the feed speed profile is analyzed, and a robust revising strategy is proposed at same time, which is also integrated into the ALANI.The on-line compensation strategy based on the rough interpolation data is proposed during the NURBS interpolation. The rationality of static and pseudo-static error compensation is analyzed, following by the presentation of space error calculation approaches. Subsequently, the on-line compensation strategy based on the rough interpolation data is proposed to solve the static and pseudo-static error. The subsection-combination method is proposed to establish the uniform compensation table, which contains both the static and pseudo-static errors. The compensation simulation is implemented to the rough interpolation data.Both the hardware and software of NURBS interpolation controller is designed and realized based on the TMS320F2812 Digital Signal Processor (DSP), following by the NURBS curve machining experiment. Furthermore, tiny NC codes machining and NURBS fitting machining are implemented, and then the quality in different ways is taken into comparison. Experiments indicate that NURBS fitting machining can obtain better quality with less code information, and the axis-based look-ahead NURBS interpolation algorithm is available and stable.