摘要 |
An ECU (30) executes PM regeneration control for adjusting the bed temperature of a DPF (22) to a target bed temperature by adding unburnt fuel to exhaust gas. During execution of PM regeneration control, the ECU (30) learns the learning value of the quantity of fuel addition for correcting the gap between the bed temperature of the DPF (22) and the target bed temperature. In case of YES at step S510, i.e. a judgment is made that the exhaust gas flow to the DPF (22) is in steady state, the ECU (30) stores a correction value K as a steady learning value Ka at step S520. In case of NO at step S510, i.e. a judgment is made that the exhaust gas flow is in transient state, the ECU (30) stores the correction value K as a transient learning value Kb at step S530. During execution of PM regeneration control, the ECU (30) reflects the transient learning value Kb on the quantity of fuel addition. NOx reduction control is executed based only on the steady learning value Ka. |