Failure of Bonded Composite Joints at Cryogenic Temperatures Loaded in a Modified Arcan Fixture
Topic(s) : Experimental techniques
Co-authors :
David DAVID BREARLEY (UNITED KINGDOM), Tobias LAUX (UNITED KINGDOM), M'Hamed M'HAMED LAKRIMI , Janice M. DULIEU-BARTON (UNITED KINGDOM), Ole T. THOMSEN (UNITED KINGDOM)Abstract :
A bore magnet within a magnetic resonance imaging (MRI) machine consist of coils of superconducting wire, impregnated with an epoxy resin to hold its shape and simultaneously acts as an adhesive to bond the solenoids to electrically insulating and supporting rings of glass fibre reinforced polymer (GFRP). To reach a superconducting state, the assembly operates at a cryogenic temperature of 4K (liquid Helium). During operation, the coils induce significant electromagnetic forces, on the inert GFRP ring. Because of these forces, stress concentrations occur in the interfacial region between the GFRP rings and the coil, which could initiate fracture, when the electromagnetic force is applied.
A modified Arcan fixture (MAF), as shown in Figure 1, is employed to assess the load response and failure behaviour of the solenoid/GFRP spacer joint at representative bi-axial stress states and cryogenic temperatures. This is done in a specially designed cryostat enabling in-situ testing with the MAF, isolating the specimen at 170K and up to ambient temperature. To observe the change in load response and failure behaviour extrapolated down to the operational conditions of the magnet. The experimental methodology incorporates in-situ use of digital image correlation (DIC), enabling detailed full-field observations of the strain state and failure behaviour of the solenoid/GFRP spacer. To evaluate the response of the epoxy adhesive alone, an adhesively bonded steel joint, subjected to analogous load and thermal conditions was used to examine the effect of varying stress ratios and temperatures. The resulting deformation and failure of these specimens provide a simplified approach for defining the procedure, removing the mismatch in materials and the complexity of the copper coil. The configuration provides a well-defined plane of adhesion where fracture can occur. This is contrary to the composite joint configuration, which has a complex geometry, stiffness distribution and failure behaviour unique to the MRI magnet.
Primary experimental observations indicate that the failure load and overall stiffness of the composite joints increases with decreasing temperature, for both joint configurations. The failure load observed for composite joints cut from an MRI magnet is lower than that of the bonded steel specimens, displaying different failure mechanisms. Instead of the crack being forced along the adhesive plane, like with the bonded steel specimens, failure in the contemporary specimens can occur propagate into the adherends where the crack follows the path of least resistance. The paper will describe the failure envelopes for both material cases at varying temperatures, and how the results can be used to inform predictive models.
This work was supported by UK Engineering and Physical Sciences Research Council (EPSRC), through an iCASE studentship with the Bristol Composites Institute, University of Bristol and Siemens Healthineers Magnet Technology.
A modified Arcan fixture (MAF), as shown in Figure 1, is employed to assess the load response and failure behaviour of the solenoid/GFRP spacer joint at representative bi-axial stress states and cryogenic temperatures. This is done in a specially designed cryostat enabling in-situ testing with the MAF, isolating the specimen at 170K and up to ambient temperature. To observe the change in load response and failure behaviour extrapolated down to the operational conditions of the magnet. The experimental methodology incorporates in-situ use of digital image correlation (DIC), enabling detailed full-field observations of the strain state and failure behaviour of the solenoid/GFRP spacer. To evaluate the response of the epoxy adhesive alone, an adhesively bonded steel joint, subjected to analogous load and thermal conditions was used to examine the effect of varying stress ratios and temperatures. The resulting deformation and failure of these specimens provide a simplified approach for defining the procedure, removing the mismatch in materials and the complexity of the copper coil. The configuration provides a well-defined plane of adhesion where fracture can occur. This is contrary to the composite joint configuration, which has a complex geometry, stiffness distribution and failure behaviour unique to the MRI magnet.
Primary experimental observations indicate that the failure load and overall stiffness of the composite joints increases with decreasing temperature, for both joint configurations. The failure load observed for composite joints cut from an MRI magnet is lower than that of the bonded steel specimens, displaying different failure mechanisms. Instead of the crack being forced along the adhesive plane, like with the bonded steel specimens, failure in the contemporary specimens can occur propagate into the adherends where the crack follows the path of least resistance. The paper will describe the failure envelopes for both material cases at varying temperatures, and how the results can be used to inform predictive models.
This work was supported by UK Engineering and Physical Sciences Research Council (EPSRC), through an iCASE studentship with the Bristol Composites Institute, University of Bristol and Siemens Healthineers Magnet Technology.