| 摘要 |
PROBLEM TO BE SOLVED: To accelerate discriminating a case where an element crack occurs to a downstream sensor from a case where air-fuel ratio imbalance among cylinders occurs.SOLUTION: If duration of a state in which an output value of an Osensor is lean as compared with a predetermined value is equal to or longer than predetermined time (S182), a speed for capturing a correction amount of a sub-feedback control into a learning value in a learning control is set higher than a speed if the duration is shorter than the predetermined time (S184). As learning progresses in the learning control, information on an output state of the Osensor can be acquired more promptly. It is, therefore, possible to accelerate discriminating a case where an element cracks occurs to the Osensor from a case where air-fuel ratio imbalance among cylinders occurs. Since execution of a correction amount guard control is suppressed until completion of the learning control, an air-fuel ratio correction amount sufficient to determine whether the air-fuel ratio imbalance among the cylinders occurs can be secured until the completion of the learning control. |
