Creating patient specific cardiovascular lumped parameter models in Fourier space using harmonic balance CFD results

The lumped parameter modeling technique has demonstrated great potential in supporting diagnosis and treatment planning of cardiovascular disease. These models offer low computational cost, fast computing time, and relative simplicity, in contrast to complex 3D hemodynamics simulations. Traditional construction of such models relies on solving an error minimization problem to determine the model parameters from clinical or simulation-derived measurements. However, this approach is impractical in patient-specific clinical settings, where obtaining these measurements is both cost prohibitive and time consuming. In this study, we propose a new method of constructing accurate and patient-specific lumped-parameter models. The proposed method takes advantage of the periodic nature of cardiovascular flows and models the hemodynamic responses of patient specific geometries in the Fourier/frequency space. Using a limited number of frequency-space harmonic balances CFD results, we directly generate an impedance-based lumped parameter network that captures the transient flow response of the blood vessels. The resulting model is inexpensive and can adjust to changes in heart rate and flow signals. In this talk, we will present realistic case studies to demonstrate the cost-efficiency, accuracy, and flexibility of the lumped parameter models generated from our proposed method.