Improving reliability and robustness in ventricular assist device

Abstract

This paper proposes the development of a method for designing Ventricular Assist Devices (VAD) can measure "in vitro" the improvement of reliability and risks with inherent safety concept as well as criteria for design validation. The VAD models used and developed in this research project were: i) VAD.PETG-01 with a top design and a base that connects to the pump housing, a magnetic coupling with six magnets that allows the drive by an external motor and bearing shaft clearance adjustment; ii) VAD.PETG.02 with a brushless magnetic motor, rotor with bearings and rolling elements driven by direct conduction, which allows the operation of this internal motor. The method, with the use of HFA, provided the definition of failure frequencies for the models tested which were 5.0014E-02 for VAD-PETG-01 and 3.314E-03 for DAV.PETG.02. The failure frequency for the pumps from Reyes' research, used as a reference for applying the method, was 1.8228E-05. With the failure frequencies of the top events for each of the models, it was possible to calculate the reliability improvement index between them, which was 0.068 and with a % Reliability Improvement of 93.20%. The models DAV.PETG.01 and DAV.PETG.02 were plotted in the "unacceptable" region of the acceptability criterion, due to the high failure frequency. As a reference the Reyes pumps were also plotted, but in the "acceptable" region. For the research problem, the method presented measurable results of reliability improvement among the tested models. In addition to confirming the acceptability of the Reyes pumps, used as a case study. The acceptability criterion was shown to be consistent with the calculated reliability data.