Part 3

7 EXHAUST GAS SENSOR MONITORING Contd.

The offset adaptation is based on the oxygen sensor signal downstream of the catalyst which is monitored by the sensor's own monitoring functions, and which is a more simple oxygen sensor (LSF-type). It is positioned downstream of the catalyst to prevent it from aging and contamination shift down, which ensures that its signal matches the actual lambda (hence the LSF-type sensor is the master sensor for the offset adaptation).

A lean offset of the LSU-type sensor results in a rich lambda in the exhaust system and the oxygen sensor downstream of the catalyst indicates a high voltage. A rich offset of the LSU-type sensor will result in the LSF-type sensor showing the actual lambda by indicating a low sensor voltage.

The offset can only be adapted under largely stationary conditions, when the same exhaust gas lambda would be expected upstream and downstream of the catalyst according to the corresponding operating conditions. After the individual adaptation steps and dynamic lambda disturbances, a certain flow rate through the catalyst to the downstream oxygen sensor has to be achieved before another adaptation is feasible.

Adaptation mode
There are two adaptation modes, a slow (conventional) adaptation and a fast adaptation. The conventional adaptation is slow, as it only has to compensate for minor system tolerances or slow alterations due to aging or contamination. Short-term lambda disturbances are hence not considered for the slow adaptation. The fast adaptation mode is enabled when a permanent major control deviation of the sensor signal downstream of the catalyst and / or a major control action of the second control loop indicates a large offset.

Slow offset adaptation mode
In the slow adaptation mode, the offset is adapted by utilizing the integral part of the second control loop. Prerequisites are that the offset adaptation has to be generally activated and the I-part of the control downstream of the catalyst has to be activated.

With each individual calculation step, a calibratable portion of the I-part of the second control loop is taken into account in the value of the offset adaptation. This portion is subtracted from the I-part of the downstream controller so that the influence on the adjusted lambda remains the same.

Fast adaptation mode
The fast mode is activated when the feed-forward error of the second control loop, which consists of the control deviation (setpoint voltage value minus actual voltage value at the downstream sensor) and the control action of the controller, is greater than a threshold value. The characteristic curve of this threshold value for the predicted I-part depends on the already adapted I-part. Thus, the adaptation steps can be adjusted to the data learning progress. A further condition for the stepwise adaptation is that a defined exhaust gas mass throughput has to be achieved without a plus/minus sign change of the unfiltered feed-forward error. The characteristic curve of the exhaust gas mass put through the catalyst depends on the current feed-forward error. Thus, the adaptation steps can be accelerated, as in the case of a very big feed-forward error, no major exhaust gas mass has to be put through the catalyst before the next step is performed.
In case the filtered feed-forward error is greater than a threshold value, an applicable part is extracted and used step by step. Before the next adaptation step is effected, a certain exhaust gas mass has to be put through so that the mixture change has reached the downstream sensor.

When very large offsets of the front sensor's characteristic curve occur, the first control loop may reach its correcting limit with the second control loop being disabled. In this case, the offset adaptation may only be effected by the fast adaptation mode.

7.4.2 Oxygen sensor, signal too lean / rich, P2195 P2196
Too lean: Bank 1: P2195; Bank 2: P2197; Bank 3: P3144; Bank 4: P3146
Too rich: Bank 1: P2196; Bank 2: P2198; Bank 3: P3145; Bank 4: P3147

Monitoring strategy
The adaptation value extracted from the offset adaptation is evaluated. If the adaptation value is greater than a positive threshold value, a lean offset is present. If the adaptation value is less than a positive threshold value, a rich offset is present. If the absolute value of the offset is greater than a threshold value, a major offset fault is present (oxygen sensor fault).

Typical Enable Conditions (Details see Summary Table)
- Upstream, LSU-type sensor is ready for operation
- Enable conditions of the lambda control:
- first control loop is active and the I-part (in systems with natural frequency control the P-part is sufficient) of the second control loop is active for longer than a threshold value

OR
- first control loop is active and is at the controller's maximum limit with the lambda value of the upstream sensor indicating a lean mixture and the sensor voltage of the sensor downstream of the catalyst indicating a rich mixture for longer than a threshold value

OR
- first control loop is active and is at the controller's minimum limit with the lambda value of the upstream sensor indicating a rich mixture and the sensor voltage of the sensor downstream of the catalyst indicating a lean mixture for longer than a threshold value

- the exhaust gas mass flow is within a defined range
- the modeled catalyst temperature is within a defined temperature range for longer than a time threshold value
- misfire-based fuel disturbance compensation for each individual cylinder is not active
- after disturbances or unsteady operating conditions a certain exhaust gas mass has to be put through the catalyst
This applies after the following conditions:
- operating points with a setpoint lambda value not equal to 1
- catalyst purge after deceleration fuel cut-off
- Deceleration fuel cut-off
- gradients of the relative engine load > threshold value
- Engine temperature < threshold value
- catalyst heating
- secondary air injection
- Long enrichments caused by acceleration or long enleanments caused by deceleration (surface film effects that lead to mixture disturbances)
This enable condition is monitored by incrementing a counter by a certain amount every 100 ms when an acceleration enrichment or a deceleration enleanment occurs.
In case these conditions are currently not present, the counter is decremented every 100 ms. As long as the counter value is greater than a threshold value, the adaptation is disabled.

Malfunction Criteria
If the enable conditions are fulfilled, the monitor is started.

If the adaptation value extracted from the offset adaptation is greater than a high, positive threshold value, and if no "sensor signal too rich (stuck rich)" fault is present at the downstream oxygen sensor, the fault "sensor signal too lean (stuck lean)" is detected.

If the adaptation value extracted from the offset adaptation is less than a high, negative threshold value, and if no "sensor signal too lean (stuck lean)" fault is present at the downstream oxygen sensor, the fault "sensor signal too rich (stuck rich)" is detected.

To be able to detect the lean LSU sensor offset, it has to be ensured that the sensor signal can still fall below the setpoint threshold value for the second control loop. Therefore, the downstream sensor must be tested by using the oscillation check with active enleanment or by using the deceleration fuel cut-off check of the diagnostic for the downstream sensor, before the fault for the upstream sensor is registered.2 The deceleration fuel cut-off check is carried out if the vehicle is in deceleration fuel cut-off mode at the time of the planned oscillation check.

To be able to detect the rich LSU sensor offset, it has to be ensured that the sensor signal can still fall below the setpoint threshold value for the second control loop. Therefore, the downstream sensor must be tested by using the oscillation check with active enrichment of the diagnostic for the downstream sensor, before the fault for the upstream sensor is registered.3

For the confirmation of a major offset fault, it must be ensured that the offset adaptation has been enabled long enough to be able to adapt a major offset and that it has executed a sufficient number of adaptation steps in the fast adaptation mode. For this purpose a counter that counts the adaptation steps is incremented, as soon as the conditions for overriding the adaptation value with the offset value are fulfilled. (In this context), it must be ensured that a certain exhaust gas mass has to be put through the catalyst between two adaptation steps.






7.4.3 Lambda correction downstream of catalyst, control threshold "rich" exceeded, P2096 P2097
Threshold "rich" exceeded: Bank 1: P2096; Bank 2: P2098; Bank 3: P117C; Bank 4: P117E
Threshold "lean" exceeded: Bank1: P2097; Bank 2: P2099; Bank 3: P117D; Bank 4: P117F

Monitoring strategy
The adaptation value extracted from the offset adaptation is evaluated. If the adaptation value is greater than a positive threshold value, a lean offset is present. If the adaptation value is less than a negative threshold value, a rich offset is present. If the absolute value of the offset is less than a threshold value, a small offset value is present.

Typical Enable Conditions (Details see Summary Table)
- Upstream, LSU-type sensor is ready for operation
- Enable conditions of the lambda control:
- first control loop is active and the I-part (in systems with natural frequency control the P-part is sufficient) of the second control loop is active for longer than a threshold value

OR
- first control loop is active and is at the controller's maximum limit with the lambda value of the upstream sensor indicating a lean mixture and the sensor voltage of the sensor downstream of the catalyst indicating a rich mixture for longer than a threshold value
OR

- first control loop is active and is at the controller's minimum limit with the lambda value of the upstream sensor indicating a rich mixture and the sensor voltage of the sensor downstream of the catalyst indicating a lean mixture for longer than a threshold value

- the exhaust gas mass flow is within a defined range
- the modeled catalyst temperature is within a defined temperature range for longer than a time threshold value
- misfire-based fuel disturbance compensation for each individual cylinder is not active
- after disturbances or unsteady operating conditions a certain exhaust gas mass has to be put through the catalyst
This applies after the following conditions:
- operating points with a setpoint lambda value not equal to 1
- catalyst purge after deceleration fuel cut-off
- deceleration fuel cut-off
- gradients of the relative engine load > threshold value
- Engine temperature < threshold value
- catalyst heating
- secondary air injection
- Long enrichments caused by acceleration or long enleanments caused by deceleration (surface film effects that lead to mixture disturbances)
This enable condition is monitored by incrementing a counter by a certain amount every 100 ms when an acceleration enrichment or a deceleration enleanment occurs.
In case the conditions are currently not present, the counter is decremented every 100 ms. As long as the counter value is greater than a threshold value, the adaptation is disabled.

Malfunction Criteria
If the enable conditions are fulfilled, the monitor is started. If the absolute value of the offset is greater than or less than a threshold value, a fuel system fault (fuel trim fault) is present.

If the adaptation value extracted from the offset adaptation is greater than a low, positive threshold value, and if no "signal too rich (stuck rich)" fault is present at the downstream oxygen sensor, the fault "mixture too lean (control threshold "rich" exceeded) is detected.

If the adaptation value extracted from the offset adaptation is less than a lower, negative threshold value, and if no "signal too rich (stuck rich)" fault is present at the downstream oxygen sensor, the fault "mixture too rich (control threshold "lean" exceeded) is detected.

To be able to detect the lean offset, it has to be ensured that the sensor signal can still fall below the setpoint threshold value for the second control loop. Therefore, the downstream sensor must be tested by using the oscillation check with active enleanment or by using the deceleration fuel cut-off check of the diagnostic for the downstream sensor, before the fault for the upstream sensor is registered.4 The deceleration fuel cut-off check is carried out if the vehicle is in deceleration fuel cut-off mode at the time of the planned oscillation check.

To be able to detect the rich LSU sensor offset, it has to be ensured that the downstream sensor signal can still exceed the setpoint threshold value for the second control loop. Therefore, the downstream sensor must be tested by using the oscillation check with active enrichment of the diagnostic for the downstream sensor, before the fault for the upstream sensor is registered.

For the confirmation of a fuel system fault (fuel trim), it must be ensured that the offset adaptation has been enabled long enough so that the adaptation value for the currently present offset can be learned. For that purpose, a time counter must have exceeded a calibratable threshold. This counter is incremented when the enable conditions of the offset adaptation are fulfilled and when the second control loop is active OR when the setpoint lambda value equals 1 and the first control loop is not at its minimum or maximum limit. This fault confirmation period is only required if a fault had already been present in the preceding driving cycle.






7.4.4 In-Use Monitor Performance Ratio (IUMPR)

Incrementing the numerator
The numerator is only generated for the large offset diagnostic as the faults of the small offset are treated as fuel trim faults and thus, this diagnostic is classified as continuous monitor of the fuel delivery system.

The numerator is incremented if the following conditions are fulfilled:
- a "stuck rich" sensor offset fault has been detected OR
- a "stuck lean" sensor offset fault has been detected OR
- a PASS result is present

AND
- the simulation of the check with a FAIL result is completed
- the defined master fault paths have been checked, too (valid for a FAIL result)

Simulation of the check with a FAIL result
The numerator is incremented if the following three partial results are present:
- the adaptation of an imaginary offset has been simulated
- the active check by means of lambda modification of the downstream oxygen sensor has been simulated
- a result of the deceleration fuel cut-off check of the downstream oxygen sensor is available

Simulation of the adaption of an imaginary offset
For the simulation of an offset through the fast path (large offset), an offset greater than a threshold value is assumed. The value of this assumed offset is greater than or equal to the offset that leads to the fault threshold values being exceeded. The simulated adaptation value is incremented as in the case of the real fault. As soon as the enable conditions for the offset adaptation are fulfilled, a part of the imaginary offset is adopted step by step in the adaptation value.
To set the flag which indicates that the offset adaptation method could have found the fault, two conditions must be fulfilled:
- the simulated offset exceeds the calibratable threshold value which corresponds to the maximum absolute value of the two fault threshold values of a lean or rich offset.
- the time counter that is incremented when the conditions for the detection of a small offset using the slow adaptation path (I-part of the second control loop) are fulfilled, exceeds a calibratable threshold value. The time counter and the threshold value are the same as for the fault detection.

In order to increment the numerator for this diagnostic, the simulated check of the oxygen sensor downstream of the catalyst is necessary, too.

Simulation of the check of the oxygen sensor downstream of the catalyst
As in the case of a fault, the downstream sensor is additionally checked after a detected offset in order to enable the correct pinpointing, the check of the downstream sensor must be simulated before the condition "numerator complete" may be set. This IUMPR simulation corresponds to the simulation for the oscillation check of the downstream oxygen sensor.5

Incrementing the denominator
The denominator is incremented by one if the conditions for incrementing the General Denominator according to CCR (d) (4.3.2.) (F) (ii) are fulfilled.