A New Feedback Control of a Lean Nox Trap Catalyst 2004-01-0527

A new feedback control for an LNT has been developed. This control adapts the rich spike (regeneration) operation in accordance with conditions of the engine and the LNT to realize high precision and robustness. The control consists of three components. First is a reference model composed of an inlet Nox concentration model and an LNT model. Second is a controller for rich spike timing and its duration. The start timing of the rich spike and its duration are determined based on the reference model. Third is a self tuning function. It adapts the parameters of the reference model using an Nox sensor downstream from the LNT. In particular, it can distinguish between deterioration of the LNT and unexpected change (increase or decrease) of Nox concentration in inlet gas in lean operation. To evaluate the performance of the proposed control, two other types of controls (type 1 and type 2) are discussed. Type 1 is a control based on the reference model without a sensor. Type 2 is a control based on the reference model and a self tuning function for the LNT model using an O2 sensor downstream from the LNT. Emissions (Nox, HC, CO) and fuel consumption performance are evaluated between type 1, type 2 and the proposed control. Only the proposed control is able to adapt the rich spike timing and duration when deterioration of the LNT and change of Nox concentration occurred. Therefore emissions and fuel consumption are optimized when the proposed control is used.