Predicting field performance based on FEA and Monte Carlo simulation techniques

Abstract

Today Active Implantable Medical Devices (AIMDs) are being used for a variety of therapies intended to enhance the quality for patients suffering from disabling diseases. The Implantable Neurostimulator (INS) is part of the Activa Deep Brain Stimulation (DBS) system developed for the treatment of movement disorders. A few years after initial release a small number of devices were returned for analysis of repeated power-on-reset (POR), intermittency, or loss of output. Analysis determined the cause of failure to be cyclic fatigue fracture of the B+ battery bond wires at the heel of the bond. Corrective action was to increase package stiffness and reduce the amount of micromotion between the hybrid and battery. Bench test results showed a significant improvement in performance. Analytical methods were then applied to predict how the improved lab performance would translate into field performance. The stresses at the heels of the wire bonds were modeled utilizing the COMSOL Multiphysics finite element analysis (FEA) package. The results were applied to a fatigue crack growth equation. Monte Carlo simulation of the variables shows a significant difference between the two different device configurations consistent with bench testing and field performance to date.