Physics-based lumped-parameter modeling of automotive canister fuel purge

A physics-based canister fuel purge flowrate model based on fluid mechanics and thermodynamics principles is presented in this paper to improve the tailpipe emission control in automotive applications. The aim of the paper is to derive a lumped-parameter model for the canister that estimates the flowrate out of the canister purged into the intake manifold. The lumped parameters of the model, including canister capacitance and flow resistance are employed to obtain a first-order multi-input and single-output (MISO) dynamic model. The vacuum pressure in the intake manifold and the fuel tank pressure serve as inputs and the mass flowrate is considered as the model output. Gradient descent system identification method is used to estimate the model parameters based on experimental data collected at Ford Motor Company. The advantage of the model is that it does not require cumbersome integral or differential computational methods and is easy to implement for fueling control purposes. Hence, it allows direct implementation in the fueling control to compensate for the extra fuel benefiting the stoichiometric air/fuel regulation in the engine catalyst.