Cylinder Pressure-based Virtual Sensor for Gas State Estimation During Compression Stroke

摘要

The gas state during the compression stroke can vary depending on the operating conditions and cycle-to-cycle variations. In this research, determination of polytropic exponent, trapped mass, and gas temperature are addressed. A golden section search method is applied to cyclic polytropic exponent estimation, and then a statistical filter is employed for cyclic estimation to filter out the estimation noise. A novel iterative-Δp-method is finally presented for the determination of the trapped mass and gas temperature simultaneously during the compression stroke. A sequence of trapped mass estimates and a sequence of gas temperature estimates can be obtained along crank angle position. Experimental validations carried out on a gasoline engine demonstrate the effectiveness of the presented methods.