UKF-based estimation of indicated torque for IC engines utilizing nonlinear two-inertia model

The estimation of the indicated torque for internal combustion (IC) engines is important for control, diagnosis and performance evaluation. If adequate sensors were available for pressure measurement in each cylinder, the indicated torque estimation would be easier. Those sensors, however, cannot be utilized generally from cost and installation space viewpoints. Hence, this paper considers an indicated torque estimation method utilizing only the crank-shaft angle information, which can be obtained even for commercial cars, instead of pressure sensors. The use of the crankshaft angle causes a problem that an unexpected vibration contaminating the angle information deteriorates the torque estimation accuracy because vibration frequency is relatively close to the combustion cycle. Thus, the engine torque estimation can be formulated as an observer design under model uncertainty and nonlinearity. In addition, this observer design problem is classified into an unknown input observer design because the engine torque or the combustion pressure as input to the engine system cannot be measured directly. To solve the design problem, this paper proposes the introduction of a nonlinear two-inertia model to catch the unexpected vibration, and an UKF-based I-type disturbance observer to estimate the indicated torque and the system state. The effectiveness of the proposed method is verified through experiments.