Monitoring Process
1. Computation of the Amplitude RatioThe first step is the computation of the amplitude of the signal oscillations of the upstream and downstream lambda sensors. This is accomplished by extracting the oscillating signal component, computing the absolute value and averaging the signal over time. The quotient of downstream amplitude value divided by upstream amplitude value is called amplitude ratio (AR).
This AR is the raw information necessary for the TWC monitoring method. It is computed continuously over a certain engine load and speed range. The signal paths for both sensor signals are identical. Thus, variations like an increase of the control frequency, affect both signal paths in the same way and are compensated by the division.
2. Post processing
The actual amplitude ratio is compared with a limit value according to the load and speed range the engine is operating in. The result of this comparison, the difference of both values, is accumulated separately for each range. Thus, even short time periods of driving in a certain range yield additional information.
By using separate load and speed ranges in combination with the accumulation of information, results can be obtained during a single drive cycle.
3. Fault evaluation
The accumulated information about the amplitude ratio becomes more and more reliable as different load and speed ranges are used during a drive cycle. If there are high values of the amplitude ratio in a certain number of load and speed ranges, a fault is detected and an internal "pending DTC" will be set. If the fault is detected again in the next drive cycle, the MIL will be illuminated.
4. Check of monitoring conditions
The monitoring principle is based on the detection of relevant oscillations of the downstream sensor signal during regular lambda control. It is necessary to check the driving condition for exceptions where no regular lambda control is possible, e.g. fuel cut-oft. During these exceptions and for a certain time afterwards, the computation of the amplitude values and the post processing is halted. Thus, a distortion of the monitoring information is avoided.