Development of an Additive Manufactured Fitting Sensorized with Optical Fibres for Load Recognition

In this work an additive manufactured fitting sensorized with optical fibres was developed with the objective to recognize the applied loads in terms of load magnitude and orientation. The component was designed with grooves to host the optical fibre sensors. A technological study to assess the strain transfer between the component and the fibre was carried out by comparing two strategies. Such evaluation was based on 3 point bending tests on specimens and numerical studies. Results showed that embedding the optical fibre into a composite patch bonded to the component guarantee good strain transfer. The component was then modeled with Finite Elements and influence functions, which relate the load orientation to the strain read by the sensors, were created. Load identification is obtained minimizing the difference between the strain read by the sensors and the influence functions. A correction procedure for the influence functions was set up using a Genetic Algorithm, which allows taking into account for the discrepancies between numerical model and real component. Virtual experiments with Finite Elements models were performed to assess the correction procedure introducing discrepancies like variation in material properties, contacts and noise. Despite such variability loads were correctly identified, moreover, a high sensitivity to the sensor position and contact clearance was observed. Correction procedure was further assessed experimentally by loading the component with different orientation. Overall results are satisfactory, the algorithm is able to identify the applied loads with sufficient accuracy and only some problems were encountered during a test on a specific configuration.