Programa del congreso
Resúmenes y datos de las sesiones para este congreso. Seleccione una fecha o ubicación para mostrar solo las sesiones en ese día o ubicación. Seleccione una sola sesión para obtener una vista detallada (con resúmenes y descargas, si están disponibles).
|
Resumen diario |
| Sesión | |
|
Vi-S6.2-ModGD: Modelado cardiovascular y gemelos digitales Lugar: Aula 0.01 Presidente de la sesión: Esther Pueyo Presidente de la sesión: José María Ferrero de Loma-Osorio | |
    | |
9:45 - 10:00
Calibrating arterial networks via inverse-adjoint methods 1: Instituto de Investigación en Ingeniería de Aragón, Universidad de Zaragoza, C. de Mariano Esquillor Gómez, s/n, Zaragoza, 50018, España; 2: Instituto Universitario de Matemáticas y Aplicaciones, Universidad de Zaragoza, C. de Pedro Cerbuna, 12, Zaragoza, 50009, España Central aortic pressure and cardiac output are key indicators of cardiovascular health, but they are usually obtained through invasive techniques. To enable real-time diagnostics, it is essential to develop fast calibration methods based on reduced-order cardiovascular models, in order to lower the computational cost of simulations. We present an efficient and innovative calibration method for arterial networks modeled with lumped parameters. Inspired by the variational adjoint method, it evaluates the sensitivity of the difference between simulations and measurements with respect to the parameters, without the cost increasing with their number. Based on the physical laws describing blood flow (mass and momentum conservation), a mathematical tool is constructed that propagates information from the peripheral arteries, where pressure measurements are available, to the network inlet. The procedure successfully calibrates the inlet boundary condition of the network and the peripheral parameters of the vascular beds, represented through Windkessel models, assuming the geometry and mechanical properties of the network are known. Comparing with data derived from a high-fidelity three-dimensional model, errors in the inlet flow waveform are below 5%. This demonstrates the potential to calibrate reduced cardiovascular models more quickly and cost-effectively than with neural networks, with the possibility of extending the approach to more complex networks.
| |