Электронный каталог

Статья

Toth, T.
Application of Mathematical Methods in the Processing of Computer Tomography Free Form Surfaces Measurement Data / T.Toth, V.Rajtukova, M.Dovica, J.Busa. – Text : electronic // Results in Engineering. – 2026. – Vol. 32. – P. 112463. – URL: https://doi.org/10.1016/j.rineng.2026.112463. – Bibliogr.: 56.

The main objective of the work is to measure free surfaces obtained by CT scanning using RMSE minimization for fine-tuning the coordinate system. The secondary objective is to evaluate the effect of build orientation on the geometric accuracy of parts manufactured by selective laser sintering. Twelve parts of Polyamid 12 were produced in vertical, inclined, and horizontal orientations using an EOS P 396 system. The parts were digitized by Carl Zeiss Metrotom 1500 micro-CT, and surface datasets were exported at two quality configurations. A coordinate system fine-tuning framework –progressing from an initial 3-2-1 registration to 3D translation and full 6-DOF rigid transformation using an inexact Newton method – was proposed and implemented in Octave. Statistical validation was conducted using 95% confidence intervals, Fisher’s F-test, independent Student’s t-tests, and Hedges’ g effect sizes. The application of the proposed optimization algorithm reduced the total RMSE by 7.0% to 30.7% compared to the initial 3-2-1 alignment. Vertical orientation achieved the highest absolute geometric accuracy (minimum mean RMSE of 98.0 µm), compared to horizontal orientation (up to 207.4 µm). Student’s t-test and Hedges’ g effect size (g > 1.21) statistically confirmed the results achieved. The optimization framework is highly robust because the mesh downsampling caused a negligible mean RMSE difference under 1.2 µm. The proposed Newton-based fine-tuning approach effectively isolates manufacturing errors from positioning offsets. For the analyzed complex workpiece geometry, horizontal build orientation is unsuitable, while the vertical orientation provides the smallest mean RMSE, and the inclined orientation provides the smallest standard deviation.



Спец.(статьи,препринты) = С 17 и1 - Математическое моделирование
Спец.(статьи,препринты) = С 344.1ш - Методы обработки результатов измерений
ОИЯИ = ОИЯИ (JINR)2026