Real-Time Desorption Flow Prediction of Automobile Carbon Canister for Virtual Calibration

Fuel evaporation emission control system (EVAP) is very important in automobile. The air desorption flow of EVAP during driving under the regulatory driving cycle is a key parameter to provide useful guidance in optimizing canister design and reach the limit requirements of the Type IV Test and Type VII Test of China 6 emission regulation. Based on this, a real-time desorption flow prediction model is proposed via the basic structure and control strategy of the EVAP. Considering implementation of real-time desorption flow calculation in ECU, the model is set up and optimized properly based on the model identification and the parameter estimation function in MATLAB/Simulink. After then, three experimental tests which represent three different control strategies of desorption process including normal desorption mode, moderate desorption mode and aggressive desorption mode, are designed for the model validation. A light-duty gasoline vehicle which was equipped with a 6-speed automatic transmission and a naturally aspirated engine, was selected to conduct these three mentioned tests on a chassis dynamometer testbed. And WLTC was chosen as the driving cycle. The experimental validation results of WLTC tests show that the RMSE of predicted desorption flow could be controlled within 2.736 L/min and the precision of total predicted desorption volume could be controlled within ±3.867 %. And the predicted desorption flow can rapidly track the actual desorption flow with a percentage tracking error less than 4 %. Therefore, the desorption flow prediction model has high prediction precision and fast dynamic response, which could be used for virtual calibration of EVAP control strategy in ECU. Finally, the proposed real-time desorption flow prediction model turns to be simple enough to be effectively implemented in ECU, thus reducing the development time and costs.