Vývoj simulačního SW modelu umělého krevního řečiště s integrovaným metabolismem glukózy

Abstract

This thesis deals with the development of a simulation software model of an artificial bloodstream with an integrated glucose metabolism. Based on the mathematical description of the cardiovascular system and glucose kinetics, a modular model was assembled consisting of five interconnected blocks, namely the heart model, reservoir, dialyzer, catheter and flow cell with GOD sensor. The model was implemented in the REXYGEN environment with numerical integration using the fourth-order Runge-Kutta method. The thesis includes a description of the physical model and experimental measurements performed during two bolus glucose applications with concentrations of 555~mmol/l and 1110~mmol/l. Parameter identification of the software model was formulated as an optimization problem minimizing the deviation between the simulated and measured steady-state output, using the global evolutionary optimizer jDE. The software model successfully reproduced all three reference steady states of the physical model PRE, POST and POST2 with a deviation of less than 3~\%. The dynamics of the transient response after bolus application were not reproduced at this stage due to limited identifiability of the optimization problem. The thesis proposes a further development plan including identification from dynamic data and the use of a dynamic catheter model as a prerequisite for the transition to a digital twin of the system.