This report proposes a model that precisely estimates the state variable of NOx storage reduction (NSR) catalysts (i.e., the NOx storage amount) for the purpose of improving fuel consumption by optimizing the control of an NSR catalyst. The proposed model combines neural networks (NNs) and a physical model, and it is based on the mass conservation law in order to estimate NOx storage amount. Compared with a model that does not consider mass conservation (i.e., a model composed only of an NN), the proposed model decreases both the maximum and mean absolute error of the NOx storage amount by at least 80%. This result confirms the effectiveness of the proposed model. To implement the proposed model in an in-vehicle ECU, it is necessary to use only variables that can be acquired on the vehicle as the input variables of the proposed model. The catalyst bed temperature, which is one of the input variables for the model, is difficult to measure on the vehicle. In this report, UKF (Unscented Kalman Filter) is designed for the purpose of estimating the catalysts bed temperature averaged in the flow direction. By inputting the catalyst bed temperature estimated using UKF into the proposed model, it was possible to estimate the NOx storage amount with good accuracy. The maximum absolute error was 30.46 mg (2.2% of the correct value) in all sections.
Ikedo et al. (Wed,) studied this question.