Development and validation of an algorithm for identifying critical sections in orthopedic plates to design the experimental setup according to ASTM F382-17 standard

Orthopedic plates represent a fundamental form of internal fixation utilized in orthopedic surgery. Their primary goal is to stabilize fractures, promote bone healing and reduce the risk of non-union by generating compressive forces between adjacent unconnected bone fragments. To assess specific mechanical characteristics crucial for in vivo performance and regulatory compliance, the ASTM F382-17 standard delineates appropriate test methods . The execution of these methods necessitates a clear understanding of the setup, as only a specified region of the plate undergoes testing. Therefore, it becomes crucial to identify the most critical area across the entire plate, ensuring a worst-case consideration. Considering these factors, the thesis work, conducted in partnership with Intrauma company, suggests the development of an algorithm to identify the critical section along a plate, corresponding to the area with the lowest flexural strength modulus. The objective is to designate the surrounding area of this identified section as the test region, facilitating the design of the experimental setup. The Intrauma company previously attained this result manually using SolidWorks software, but it faced drawbacks such as prolonged working time and subjectivity in results. Consequently, the algorithm offers a faster and automated approach to achieve the same objective. In addition, two algorithmic variations (automatic and guided) have been developed, differing in the level of interaction required between the user and the program. Despite this distinction, both variations share a common operational rationale. They focus on a designated region of the plate containing potentially critical elements, estimating the plate's longitudinal curvature using a defined number of macrozones corresponding to the vectors generated to achieve this approximation. Subsequently, the plate is sliced with planes perpendicular to the approximate curvature and the identified sections are evaluated in Matlab using the Polygon Moments function to select the critical section. The automatic and guided codes have been used synergistically, with the guided method addressing limitations arising from complete automation, such as potential overestimation of the test region and curvature approximation for specific plates which lack longitudinal symmetry. When both codes are applicable to the same plate, the automatic method has demonstrated performance similar to the guided one. Results obtained for 57 plates through the developed algorithms showed no significant differences from the reference manual method. Indeed, errors in section moment of inertia and the furthest point of the plate section from the principal axes of inertia were generally below 1%. However, in specific cases, higher errors were observed, primarily due to the resolution of the STL file during the export process from SolidWorks. Based on the conducted analyses, it can be concluded that the two algorithms are validated and can serve as a substitute method for the previous manual approach.