Modeling And Analysis Of Dimensional Variation In Multi-variable Multi-station Manufacturing System

Multi-variable multi-station manufacturing system is a system in which product is usually done through several interconnected stations. Product dimensional variation is not only influenced by variations such as tool, fixture and machine in a single station but also by coupled ones between different stations, and all of the variations mentioned above will be produced, increased, accumulated, and propagated as the station goes on, which will eventually cause the dimensional variation of the final product.A lot of research works on the dimensional variation modeling in multi-variable multi-station manufacturing system using different methods, but the state space model based on the stream of variation theory is the most classical and efficiently used among those models derived, however, there commonly existing two problems in this model, one is that it is expressed in a nonlinear form which prevents the classical modern control theory being applied in the manufacturing system; the other one is that the system matrices of this model are implicitly expressed, which also hinders its further application in the real manufacturing system. Therefore, some researches have been done in following aspects:Firstly, key product characteristics and key control characteristics are sufficiently analyzed on the basis of key characteristics analysis flow chart. Part model and its variation are defined on the definition of several coordinates such as part coordinate, fixture coordinate, machine coordinate and etc. Then, homogeneous transform method is used to establish the mutual transformation between part coordinate, fixture coordinate and machine coordinate.Secondly, linear state space model and even linear input-output model with the newly achieved model are derived fully using homogeneous transform method and matrix theory.Thirdly, for the problem that implicit system matrices exist in the linearizing model, explicit system matrices are eventually achieved with the analysis of errors of fixture parameters, locating datum features, part coordinate and measurement datum features.Finally, through analysis of 3-station manufacturing process of cylinder head of engine, the effectiveness and accuracy of the derived explicit and linear model on describing the multi-variable multi-station manufacturing system is ultimately validated.