Modeling of cardiovascular system, pulmonary mechanics and gas exchange

The present dissertation is a modeling effort to mimic cardiovascular and pulmonary systems to understand the physics of systemic circulation and to predict acute cardiopulmonary complications accurately and noninvasively.; To determine a minimal lumped model/s of systemic circulation that can be characterized by parameters well correlated to the vascular properties, all possible three- and four-parameters lumped models were constructed and compared for fits of aortic input impedance to 100 subjects. The model formed by adding a serial inertance element to the aortic resistance of the traditional three-element Windkessel model (RLRC), inertance viscoelastic Windkessel model (IVW), modified Windkessel models (R-LC-R and LRC-R) represent physiologic features of the vasculature and matched the experimental impedance curve more accurately than the other models. On correlating the model parameters to the vascular properties, it was observed that (1) peripheral resistance is significant and included in all models, (2) RLRC and R-LC-R models include aortic resistance, (3) RLRC and IVW models include fluid inertia and (4) nonlinear arterial compliance is produced by IVW and LRC-R models.; A five-element lumped pulmonary model based on airway pressure and flow was used to estimate respiratory mechanics automatically and noninvasively and hence to detect an obstructed endotracheal tube and endobronchial intubation. Bronchospasm, a stiff chest wall and positive end expiratory pressure were also tested to determine the specificity of the diagnosis. Endotracheal tube obstruction is indicated if airway resistance increased more than 30% and alveolar gas compression (C1) decreased more than 10% and lung and chest wall compliance increased more than 10% from baseline. The method has a sensitivity of 90% and specificity of 97%. Endobronchial intubation is indicated if C1 decreased ≥50% and in change in lung and chest wall resistance is ≤10-fold from baseline. The above rule has 90% accuracy.; A mathematical gas exchange model based on conservation of mass, ideal gas law, Fick's principle, alveolar air equation and empirical relations was developed to predict ventilation-perfusion (V/Q) mismatch. The V/Q decreased during embolism and increased during edema. Based on V/Q, the model could identify embolism and edema in six out of seven dogs.