Three dimensional integrated software development for air-particle flow simulation through image-based upper human airways

摘要

The effort to reconstruct and simulate flow-particle behavior in realistic patientspecific airway system requires multi-software skills. Conventionally, pre-processing, simulation and optimization and post-processing stages are carried out explicitly via a combination of commercial, open source and/or in-house engineering software. The tedious procedure had left more significant medical analysis such as flow pattern classification, patient group-based flow analysis and statistical flow studies at bay. In this work, the focus is on the development of a dedicated software that is capable of performing all the three stages for any patient-specific data set. A novel approach of combining the efficient Immersed Boundary method and Finite Difference Splitting solver within a matrix-based open source programming platform has radically simplified the procedure especially in the pre-processing stage. The air and particle interactions are based on Eulerian-Lagrangian technique with comprehensive validations for each stage of the solvers integration. A non-dimensional convergence error of less than 1 x 10-6 was consistently set for all the validations. An air flow rate of 30 litre / minute was used throughout the analyses representing the normal inhalation condition while a number of 10,000 and 5,000 micro particles were modeled for simplified and image-based airways respectively. The assessment analysis showed that 42.35% of the particles inhaled by female subject managed to reach the end of trachea while male subject with epiglottis blockage recorded only 0.43%. None of the inhaled particles managed to pass through the trachea of the oversized male subject. This work suggests that such pattern analyses are crucial to facilitate medical practitioners in their patient-specific diagnosis and decision making process of airway flow related diseases.