Stream Of Variation Modeling And Application Of Multistage Machining Process

Machining processes generally consist of multiple machining steps and numerous interacting error sources. Error modeling and synthesis would be of great use for predicting product quality before machining and adjusting machining process according to the model output. This paper introduces several common quality control methods in machining process including their basic principles and applications. We focused on the variation synthesis within machining stage and the error propagation modeling between machining stages based on the homogeneous transformation theory.Geometric error modeling is mainly for the quasi-static errors. Quasi-static errors refer to the errors under the assumption that all system components are considered as rigid bodies. They are easy to analyze geometrically. In this paper, there are three main modeling objects, i.e. fixture error, datum error and machine tool error.The idea of geometric error modeling is using coordinate systems(CS) to represent part features, machine carriages, and clamps. Thus the whole machining system is simplified as a CS chain, and then the following transformation and derivation can be based on the CS chain using homogeneous transformation. Homogeneous transformation matrix(HTM) can express not only the coordinate system itself but also the location of one CS relative to another CS. With HTM, the model can be derived and expressed in a uniform format and the errors can be expressed by HTM or vector. Within machining step, there are three geometric errors, i.e. fixturing error caused by the imperfection of locators, datum error caused by the error of datum feature, machine tool error caused by structural loop of machine.Error propagation happens when the current datum is the feature generated in previous machining stage. Propagation routes are determined by the datum transformation routes. The error relative to previous datum is transformed to that relative to following datum. By substituting the measured error parameter into the model, we can predict the error after current machining step.At last, we validate the feasibility of the model through the instance of engine box by comparing error model output with the actual measurement.