A sensorless physiologic control strategy for continuous flow cavopulmonary circulatory support devices

Abstract

ObjectiveMechanical circulatory support devices using continuous flow blood pump technologies are currently being developed to provide cavopulmonary support in patients with Fontan circulation. While cavopulmonary support can functionally replace the missing native right ventricle and biventricularize the Fontan circulation, there is a need for the pump to adapt to variable physiologic demands. A constant pump speed limits the ability to provide sufficient physiologic perfusion over a range of physiologic conditions. To address this challenge, a physiologic control algorithm for a Fontan pump was developed. MethodsThe proposed algorithm uses a gain-scheduled, proportional-integral controller that generates a user defined cavopulmonary pressure head (CPPH = pulmonary artery pressure - vena caval pressure), to provide physiologic pressure rise while avoiding suction. The approach uses only intrinsic pump parameters and does not require the use of pressure or flow sensors. Performance and robustness of the sensorless control algorithm was quantified in-silico by simulating the following conditions: (1) Directly measured CPPH with pressure sensors; (2) CPPH estimation using the intrinsic pump measurement of pump speed and extended Kalman filter (EKF); (3) constant speed control algorithm; and (4) rapid four-fold increase in vena caval resistance (VCR) for (1) to (3). ResultsThe results demonstrated that the sensorless control algorithm maintained physiologic perfusion while simultaneously preventing vena caval suction without the need for pressure sensors. Conclusion and significanceThe proposed algorithm was superior to the constant speed control strategy for physiologic perfusion and suction prevention, and warranted further investigation in vitro and in-vivo.