Systematic modelling of a three-axes, non-contact coordinate measurement system
A coordinate measurement machine (CMM) is an advanced, multi-purpose quality control system used to help inspection keep pace with modern production requirements. These machines can provide repeatable dimensional and geometric accuracy (micro- or nano-metre accuracy) of everything from small engine blocks, to sheet metal parts, to circuit boards. However, their performance may degrade when subjected to different forms of external disturbances, including geometric, kinematic, thermal and dynamic forces. By studying the influences of these disturbances, the impact on the machine measuring accuracy can be understood. The research work described in this thesis investigates two important aspects of machine accuracy of a newly developed coordinate measurement machine. The Small Coordinate Measuring Machine (SCMM) consists of a non-contact laser displacement probe supported on three mutually perpendicular (X, Y & Z) axes. The geometric and kinematic errors, which have been shown to account for more than two-thirds of the machine errors, of the SCMM were studied. A mathematical model has been proposed and incorporates geometric, kinematic and orientation errors of the probe that may have been introduced during the assembly process. Using the Renishaw performance measurement system, the error components of each individual carriage were diagnosed. The resulting study of these errors provided a better understanding of the performance of these positioning carriages. The mathematical model was also able to show the extent of the positioning errors of the carriages and the combined volumetric error of the SCMM. The second part of this research investigated the influence of drift on the probe output. This study aimed to demonstrate the magnitude of these environmental changes and the impact it has on the stability of the output of the non-contact, laser displacement probe. System Identification has been used to identify the correlation of the drift in the output of the probe and the changes reflected in the environment. A linear, 3rd order AutoRegressive model with eXogeneous input (ARX) has been chosen based on its merits. The results obtained from both the mathematical and linear models were eventually applied to a practical measurement problem, leading to a significant improvement in the measurement accuracy.