Identification of nonlinear hysteretic engine boost pressure characteristics using Preisach representations

Abstract

Increasing complexity of modern combustion engines lead to a significant augmentation of the number of physical setpoints, sensorics/actuators and control systems in electronic engine control units. Parameter calibration of these structures in order to adapt them to the current engine/vehicle configuration is a major task in automotive series development. Well established parameter-optimisation procedures based on knowledge and engineer's intuition no longer suffice in current and future calibration practise, therefore model-based approaches are used. The quality of the process model acquired by conventional linear or nonlinear system identification is often insufficient for systems showing rate-independent memory characteristics, also known as hysteresis. The paper describes model-based calibration approaches by process model identification using a hysteresis model. Modeling results by the example of a direct injection two-stage diesel engine boost pressure control are presented.