Simulation of Wear in Turbocharger Wastegates

Abstract

This thesis report presents a Finite-Element based methodology for simulating wear damage in the wastegate system of automotive turbochargers. The derived simulation routine is able to reproduce the changes in geometry at contact interfaces of dynamic systems that are subject to relative motion and material ablation. The wastegate is a regulatory device that determines the boost pressure output of a turbocharger. It consists of several distinct components that are in constant relative motion due to external dynamic loads, making their contact surfaces very susceptible to material depletion. The most wear-critical location in the system is the interface between the wastegate's lever and bushing, whose cylindrical outer surfaces slide against each other due to the intended action of the wastegate's regulation actuator as well as unintended effects of exhaust gas pulsations. Severe degradation of the contact surfaces leads to significant changes in the parts' geometry, which inhibits the proper functioning of the device: It becomes increasingly difficult to achieve the optimum part configuration for the requested amount of boost pressure, which in extreme cases might result in device failure. The long-term goal at BMW's department for exhaust system development is to achieve the capability to predict the extent of wear damage at different engine operating conditions. Subject of the current thesis project is the development of a fundamental simulation routine that is able to qualitatively generate the expected geometric changes with a Finite-Element model. The most significant product of this study is a Python script that uses the commercial ABAQUS FE-solver to process the effects of an arbitrary dynamic load within a short period of time, calculate the expected geometric changes, extrapolate them to a large number of load cycles, and transfer the modified geometry back into the original analysis. The script includes three fundamental elements that are essential to any wear simulation concerned with realistic dynamic systems: 1. an accurate representation of the geometric transformation of a three-dimensional model with arbitrary spatial orientation, using a universal method for finding node-displacement directions at part boundaries, 2. a global routine to import and transfer the modified geometry between dynamic and static analyses, and 3. a sensible technique to extrapolate the impact of a load case within a short period of time to represent the effect of a large number of load cycles, such that substantial wear damage - which is typically incurred over several days or weeks - can be simulated with acceptable computation times. The implementation of the individual features have been verified numerically, and it has been shown that the overall approach itself is able to produce legitimate results if certain precautions are taken. The derived "wear simulation tool" can be applied to any generic FE-model with a defined load case and therefore provides a practical basis for experimental validation, feature extensions and fine-tuning.