Byounggul Oh

Real-Time IMEP Estimation and Control Using an In-Cylinder Pressure Sensor for a Common-Rail Direct Injection Diesel Engine

An in-cylinder pressure-based control method is capable of improving engine performance, as well as reducing harmful emissions. However, this method is difficult to be implemented in a conventional engine management system due to the excessive data acquisition and long computation time. In this study, we propose a real-time indicated mean effective pressure (IMEP) estimation method using cylinder pressure in a common-rail direct injection diesel engine. In this method, difference pressure integral (DPI) was applied to the estimation. The DPI requires only 180 pressure data points during one engine cycle from top dead center to bottom dead center when pressure data are captured at every crank angle. Therefore, the IMEP can be estimated in real time. To further reduce the computational load, the IMEP was also estimated using DPI at 2 deg, 3 deg, and 4 deg crank angle resolutions. Furthermore, based on the estimated IMEP, we controlled IMEP using a radial basis function network and linear feedback controller. As a result of the study, successful estimation and control were demonstrated through engine experiments.

Development of Turbine Mass Flow Rate Model for Variable Geometry Turbocharger Using Artificial Neural Network

지구온난화 문제와 에너지고갈 문제에 대응하기 위하여 세계 각국에서는 차량용 내연기관에 대한 연비규제를 점차 강화하고 있다. 디젤엔진은 가솔린엔진 대비 높은 열효율에 의한 우수한 연비, 낮은 이산화탄소 배출 등의 이유로 차세대 자동차의 Key Words: Turbine Model(터빈 모델), Artificial Neural Network(인공신경망), Diesel Engine(디젤엔진), Variable Geometry Turbocharger(가변 기구 터보차저), Mass Flow Rate(질량 유량) 초록: 이 논문에서는 인공신경망을 이용하여 가변 기구 터보차저(VGT)의 터빈 질량유량을 추정하는 모델을 제안하고자 한다. 터빈 질량유량을 추정하기 위한 모델의 입력변수는 VGT 베인 개도량, 엔진 회전속도, 배기매니폴드 압력, 배기매니폴드 온도, 터빈 출구 압력이 사용되었으며, 터빈 입구 유효 단면적을 추정하는 부분에 인공신경망을 적용하였다. 실험을 통하여 이 논문에서 제안한 모델의 터빈 질량유량 추정 성능을 검증하였으며, 터빈 맵을 이용하여 추정한 결과와 비교를 통하여 제안한 모델의 우수성을 확인하였다. Abstract: In this paper, we propose a turbine mass flow rate model for a variable geometry turbocharger (VGT) using an artificial neural network (ANN). The model predicts the turbine mass flow rate using the VGT vane position, engine rotational speed, exhaust manifold pressure, exhaust manifold temperature, and turbine outlet pressure. The ANN is used for the estimation of the effective flow area. In order to validate the results estimated by the proposed model, we have compared estimation results with engine experimental results. The results, in addition, represent improved estimation accuracy when compared with the performance using the turbine map.