Detection of friction variations in bolted joints during tightening
Abstract: Tightening describes the process of rotating a screw with the purpose of binding two surfaces together. It is widely applied in the assembly process of structures, where bolted joints are used to connect the different parts with each other, e.g. robot arms, vehicles, aircrafts. A certain torque is applied with a tool in order to develop the desired clamping force that keeps the surfaces together. A challenge during this process is the fact that friction variations occur unexpectedly, thus increasing the risk of not achieving the necessary clamping force to ensure successful tightening. In this thesis, a diagnosis method is implemented in order to detect friction variations during highly dynamic tightening. Different detection algorithms are investigated (e.g. CUSUM, Particle Filter, Linear regression), and an approach that makes use of the torque and angle signals while estimating the clamping force is implemented. Investigations of signal noise and filtering operations during data extraction are conducted, and the signal channels are evaluated with regards to accuracy and noise bias. An approach using a sliding window is used to estimate the torque rate, and the CUSUM detection algorithm is implemented to indicate variations and provide a diagnostic report. The analysis is performed using a highly dynamic tightening strategy programmed in an electrical tightening tool, allowing for the process to be conducted in milliseconds. Investigations of the tuning parameters of the detection algorithm are also conducted, and value thresholds are identified. Finally, a statistical analysis of the system’s behaviour, as well as the influence of the operator holding the tool, is performed for evaluation.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)