Cylindricity is a key index of precision of mechanical parts. For the convenience of computer programming and improving the robust of algorithm, the method of evaluation of cylindricity geometrical error base on computational geometry is presented in this paper. The transformation of coordinates in evaluation of geometrical error is discussed. Adopting the self-adapting variation, the mathematic model and criteria for evaluating of minimum zone of cylindricity are demonstrated in detail. The procedure of algorithm is given. Finally, a case study is given to validate the proposed method.

A tolerance synthesis model is established based on degree of freedom of geometric variations of features in this paper. The method allows a designer to analyze the relationship between geometric variations of features of a part and functional requirement of assembly (FRA). Firstly, tolerance is modeled with DOFs of geometric variations of features and the tolerance zone is expressed with six kinematic DOFs in three-dimensional (3D) space. Secondly, the stack-up of geometric variations of features is formulated as explicit tolerance analysis equations using kinematical coordinate systems associated with each feature. To express mathematically the relationship between given FRA values and the corresponding DOFs of geometric variations of features, the reverse synthesis equations are obtained using a matrix inversion scheme of the tolerance analysis equations. Finally, a case study is used to illustrate the proposed method.

Take the complex three-dimensional (3D) surfaces which consist of all kinds of curves, such as cosine, parabolic and elliptic as examples, the processing is analyzed to determine the reasonable feed routes. The method of parameter programming provided by SINUMERIK 802D system is applied to choose reasonable machining parameters. Using these parameters for NC programming, not only the roughing and finishing machining with different processing requirements can be met, but also the programming solutions are optimized and the NC program is briefer than that generated by CAM software. Furthermore, human-computer interaction can be achieved by means of assignment to the parameters.

To realize the management and reuse of the knowledge of straightness verification, the ontology theory is applied in the knowledge base system of straightness verification. The system is divided into three layers: domain layer, inference layer and application layer. The ontology of the straightness verification is established for formal description of the conceptions of straightness verification and the relationship among these conceptions. Related axioms, rules of straightness measurement are given out according to related standard documents. Finally, an application platform is provided for the digitized realization of the straightness verification.