Adaptive prediction of intake manifold absolute pressure for a coal-bed gas engine

In order to correctly calculate the engine cylinder charges for determining fueling, a precise real-time estimation of the engine intake manifold absolute pressure is required for speed-density methods. With the nonlinear time-varying characteristics of the engine intake manifold system, an intake manifold adaptive model is constructed and an adaptive identifier based on the recursive identification algorithm is used to estimate intake manifold pressure during different engine operating conditions. For comparison, two recursive identification methods, Kalman Filter algorithm and Forgetting Factor algorithm are adopted. Subsequently, the predictor based on the parameters estimations is used for real-time prediction of the manifold pressure before sensor measurements. According to the manifold pressure state equation, the manifold pressure is treated as a bivariant function of throttle valve angle and engine speed, so a linear structure of equation error model is selected. Based on experiment data on engine test bench, the adaptive prediction results by recursive algorithms are compared with the estimates from conventional IIR Filter. The result shows that both of the recursive adaptation algorithms can be used for real-time prediction of manifold pressure of the coal-bed gas engine during different engine operating conditions.