Accuracy Quantification of the Reverse Engineering and High-Order Finite Element Analysis of Equine MC3 Forelimb.

The research article discussed the use of reverse engineering techniques to create an accurate model of equine third metacarpal bones (MC3) for biomechanical studies and aimed to minimize errors in the resulting finite element analysis (FEA). The study found minimal error in it’s model’s reconstructed geometry, suggesting this method could boost FEA accuracy without significantly increasing computational time.

Creating and Testing a Bone Model

• The study used reverse engineering techniques to reconstruct the diaphysis (shaft) of 15 equine MC3 bones, seeking to achieve a virtual model that accurately represents the bone’s shape for biomechanical studies.
• This involved extrapolating cross-sectional slices of the bone to recreate its geometry in a computer-aided design (CAD) environment.
• The model’s geometry was then compared against reference models derived from computed tomography (CT) scans of the same bones.
• The researchers then plotted the error map of the generated surfaces to quantify how faithfully the virtual model represented the original bone structure.

Findings on Model Accuracy

• The study found that the minimum errors in the reconstructed geometry of the bone models were +0.135 mm and -0.185 mm.
• A minimal amount of error was observed on the dorsal cortex for both the outside and inside surface.
• The deviations fell within acceptable ranges, suggesting that the reverse engineering method used in this study was capable of creating a highly accurate representation of the MC3 bones.

Minimizing FEA Errors

• The research used a displacement-based error estimation on 10 of the MC3 models to identify elements within the finite element analysis that might have been poorly shaped, which could lead to inaccuracies in the results.
• The relations of finite element convergence analysis were used to develop a framework meant to minimize stress and strain errors within the FEA.
• The literature often reports FEA errors of between 3% to 5%, but through this method, the researchers achieved much more accurate results, with FEA errors of only 0.12%.
• The proposed model will allow future investigators to maximize the accuracy of their FEA results without an increase in computational time, known as the runtime penalty.