Procedure for diagnosing a particulate filter

By predicting and correcting signal changes using an engine model and compensating for cross-sensitivities, the method enhances the accuracy and reliability of particle mass flow measurement in internal combustion engines, particularly for diagnosing particle filters and adapting emission maps.

DE102006018956B4Inactive Publication Date: 2024-11-21ROBERT BOSCH GMBH
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
DE102006018956
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2006-04-24
Publication Date
2024-11-21
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing particle sensors in internal combustion engines are cross-sensitive to external factors like temperature and exhaust gas velocity, leading to imprecise signal changes and delayed compensation, which affects the accuracy and dynamics of particle mass flow measurement.

Method used

The method compensates for cross-sensitivities by using an engine model to predict signal changes, incorporating variables such as exhaust gas temperature, volume flow, and gaseous components, and applies correction methods like gradient and triggering to evaluate measured and predicted signals accurately.

Benefits of technology

This approach allows for precise and timely compensation of cross-sensitivities, enabling accurate measurement of particle mass flow even under dynamic engine conditions, improving the reliability of onboard diagnosis and emission map adaptation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method for on-board diagnostics (OBD) of a particulate filter in the exhaust system of an internal combustion engine by means of at least one resistive particulate sensor arranged downstream of the particulate filter in the exhaust system (1) wherein - a predicted particle mass flow limit value (23) is determined based on a predicted particle mass flow (22) provided by an engine model (60) and based on a limit filter model (61), - the predicted particle mass flow limit value (23) is fed to an accumulation model (62), as well as - a predicted signal change (20) is determined using a conductivity model (63), which, due to the use of the limit filter model (61), represents the signal change expected for a limit particle filter, - wherein in a comparison stage (50) the predicted signal change (20) is compared with a measured signal change (11) of the particle sensor (1) and - where a defective particle filter is inferred if the measured signal change (11) of the particle sensor (1) is greater than the predicted signal change (20), wherein the measured signal change (11) of the particle sensor (1) and / or the predicted signal change (20) of the particle sensor (1) are corrected taking into account influencing factors (30) on cross-sensitivities of the particle sensor (1).
Need to check novelty before this filing date? Find Prior Art

Description

State of the art

[0001] The invention is based on a method for determining a mass of particles or a particle mass flow in an exhaust system of an internal combustion engine, wherein at least one resistive particle sensor is arranged in the exhaust system of the internal combustion engine, the measured signal change of which is compared with a predicted signal change of the particle sensor determined from an engine model.

[0002] Such a resistive particle sensor is described in DE 101 33 384 A1. The particle sensor is constructed of two interlocking electrodes, which are at least partially covered by a protective tube, which also serves as a collecting sleeve to improve particle deposition. If particles from the exhaust gas of an internal combustion engine deposit on the particle sensor, this leads to an analyzable change in the impedance of the particle sensor, from which the quantity of deposited particles and thus the quantity of particles entrained in the exhaust gas can be determined.

[0003] Particle sensors exhibit strong cross-sensitivity to external factors such as sensor temperature, exhaust temperature, or exhaust velocity. These factors influence both the deposition of particles on the sensor and the measured impedance of the sensor.

[0004] The subsequently published DE 10 2005 040 790 A1 describes a method for operating a sensor for detecting particles in an exhaust gas stream. At least one measure of the exhaust gas flow is determined at the particle sensor, and the measure of the exhaust gas flow is taken into account when evaluating the particle sensor signal provided by the particle sensor.

[0005] Other factors influencing the cross-sensitivities of the sensors are not taken into account in the described method. When compensating for the cross-sensitivities of the particulate sensor, the problem arises that the signal changes of the particulate sensor are very small. It can take a comparatively long time to collect a sufficient quantity of particles to produce a signal change that can be evaluated by the engine control unit. For this reason, timely compensation of the cross-sensitivities can sometimes not be carried out or can only be carried out very inaccurately. The dynamics of the particulate sensor and thus its output signal may be insufficient to follow the changing operating conditions of the internal combustion engine when the operating situations of the internal combustion engine change.In such a case, a quantifiable signal change cannot be assigned a single value, for example an exhaust gas volume flow, but rather a history of influencing parameters that changed during the measurement time must be attributed to it.

[0006] The subsequently published DE 10 2005 034 247 A1 describes a method for monitoring an exhaust emission limit of an internal combustion engine using an engine control system. The engine control system has at least one exhaust gas sensor and emits an error signal if the exhaust emission limit is exceeded. The emissions predicted for the current driving condition are determined using an engine model and compared with the signal from the exhaust gas sensor or a reference emission value derived therefrom. The exhaust gas sensor can be a collecting particle sensor. The method enables exhaust emission monitoring under operating conditions of the internal combustion engine that deviate from the standardized driving cycles used to determine the limit values.

[0007] Further methods are known from EP 1 242 723 B1 and DE 10 2004 043 122 A1. Advantages of the invention

[0008] The invention provides that the measured signal change of the particle sensor and / or the predicted signal change of the particle sensor are corrected taking into account influencing factors on cross-sensitivities of the particle sensor. In particular, when determining the predicted signal change of the particle sensor, it is possible to compensate for cross-sensitivities even during dynamic operating point changes of the internal combustion engine that occur faster than the response of the particle sensor, since the correction can also be performed for small signal changes, i.e., signal changes predicted during an operating point of the internal combustion engine that is currently being passed through.Since the predicted signal changes are determined from a characteristic field of the engine model, these measurement intervals can be determined for very short times and thus can be assigned to individual operating points of the internal combustion engine with the resulting very small predicted signal changes.

[0009] With a real particle sensor, such small signal changes may not be able to be evaluated under certain operating conditions due to the particle sensor's lack of dynamic response and fluctuations in the measured value. In this case, it is advisable to correct the cross-sensitivities in the measured signal change, for example, at the time of comparison with the predicted signal change, to account for the instantaneous cross-sensitivities. The measured and predicted signal changes can be evaluated using the well-known gradient method and the trigger method. In the case of the gradient method, the signal change corresponds to the rate of rise dI / dt of the sensor current. With the trigger method, it is monitored whether the sensor signal exceeds a specified threshold, and the time at which the threshold is exceeded is defined as the trigger time of the particle sensor.

[0010] The key parameters influencing the measurement behavior of today's particle sensors can be compensated for by taking into account the factors influencing the cross-sensitivities of the particle sensor, such as exhaust gas temperature and / or the temperature of the particle sensor and / or the exhaust gas volume flow and / or the influence of the gaseous exhaust gas components nitrogen oxides, hydrocarbons, and / or water vapor, to correct the measured signal change and / or the predicted signal change of the particle sensor. Instead of the exhaust gas volume flow, the exhaust gas velocity can also be used. However, depending on the sensor concept used, other influencing factors can also be included.

[0011] For the evaluation of the measured and predicted sensor signals, it can be provided that when a predefined current threshold is exceeded or a predefined resistance threshold is undershot, the measured signal change and / or the predicted signal change are compared with a predefined trigger threshold. When the trigger threshold is reached, the measured signal change is compared with the predicted signal change. This method enables the digital information "threshold reached / threshold not reached" to be compared at a specific time, preferably the trigger time.

[0012] In a preferred embodiment of the invention, it is provided that a particle mass flow predicted to be deposited on the particle sensor is determined from a particle mass flow predicted by the engine model in an accumulation model, taking into account the influencing variables on the cross-sensitivities temperature of the particle sensor, exhaust gas temperature and / or exhaust gas volume flow, and that a predicted deposited mass is determined from the predicted deposited particle mass flow by integration, and a predicted signal change of the particle sensor is determined via a conductivity model. The engine model outputs the particle mass flow to be expected for the current operating situation of the internal combustion engine and supplied to the particle sensor. The accumulation model determines how many of the predicted particles transported past accumulate on the particle sensor.The deposition model takes into account that the mass of particles deposited per unit of time from the predicted particle mass flow depends on the current influencing parameters of particle sensor temperature, exhaust gas temperature, and exhaust gas volume flow. Since the forecast can be carried out for very short measurement intervals, the particle mass flow deposited on the particle sensor in the current operating situation of the internal combustion engine can be determined. The integration determines the mass of particles deposited on the particle sensor over a longer period of time under different operating conditions. If the forecast is correct and the particle sensor is functioning correctly, this mass corresponds to the particle mass actually deposited on the particle sensor. The conductivity model converts the predicted deposited mass into a corresponding signal from the particle sensor, which enables a comparison with the measured signal from the particle sensor.

[0013] In a further preferred embodiment of the invention, it is provided that a predicted differential signal change of the particle sensor is determined from the particle mass flow predicted by the engine model in the deposition model, taking into account the variables influencing the cross-sensitivities of the particle sensor, i.e., temperature of the particle sensor, exhaust gas temperature, and / or exhaust gas volume flow, and that the predicted signal change of the particle sensor is determined by integrating the predicted differential signal change. In this embodiment of the invention, in contrast to the previously described variant, the variable of predicted deposited particle mass flow, which is difficult to determine, is not used; instead, the deposition model directly determines a differential signal change corresponding to the current operating conditions of the internal combustion engine, taking into account the aforementioned cross-sensitivities.From this, the predicted signal change, comparable to the measured signal change, is determined through integration across multiple operating points. The advantage of this variant is that the deposition model can be calibrated by measuring the signal change of a particle sensor in the exhaust gas of an internal combustion engine operating with constant operating parameters. This can be done accordingly for different operating points of the internal combustion engine.

[0014] By considering the influence of the current temperature of the particle sensor on the conductivity of the particle sensor in the conductivity model or in a first temperature correction stage to determine the predicted signal change, or by considering the influence of the current temperature in a second temperature correction stage to determine the measured signal change, the predicted signal change can be directly compared with the measured signal change. In addition to the mass of particles deposited on the particle sensor, the conductivity of the particle sensor is essentially determined by the current temperature of the particle sensor and the temperature of the deposited particles.Either the current temperature of the particle sensor can be taken into account when determining the predicted signal change, or the predicted signal change can be determined for a specified temperature of the particle sensor, and the measured signal change can be converted to this specified temperature. In addition to the temperature of the particle sensor, it is also conceivable to consider other factors influencing the conductivity of the particle sensor.

[0015] If the method is carried out at operating points of the internal combustion engine where the particle mass or particle mass flow can be measured or predicted with high accuracy, the accuracy and reliability of the method can be significantly improved. Furthermore, it is conceivable to separately weight such favorable operating conditions in decisions derived from the measurement results, depending on the respective applications.

[0016] A preferred variant of the invention provides that a weighting factor is determined depending on a particle mass deposited on the particle sensor and / or the exhaust gas temperature and / or the temperature of the particle sensor and / or the exhaust gas volume flow and / or the predicted particle mass flow and / or a measured particle mass flow, with which the predicted deposited particle mass flow or the predicted differential signal change is corrected. This takes into account that a particle mass passing the particle sensor makes different contributions to the signal formation at different times during a measurement cycle, as can be the case particularly with evaluation according to the trigger method.A particle mass deposited immediately after the regeneration of the particle sensor contributes less to the triggering than a particle mass deposited shortly before the particle sensor is triggered.

[0017] In one application of the method, it is provided that the method is used in the on-board diagnostics (OBD) of a particulate filter in the exhaust system of an internal combustion engine. A predicted particulate mass flow limit value is determined from the particulate mass flow predicted by the engine model and a limit filter model, which limit value is fed to the deposition model. In a comparison stage, the predicted signal change of the particulate sensor is compared with the measured signal change of the particulate sensor, and a defective particulate filter is concluded if the measured signal change of the particulate sensor is greater than the predicted signal change. The particulate sensor is arranged downstream of the particulate filter to be monitored in the exhaust direction.The particle retention capacity stored in the limit filter model corresponds to the retention capacity of a particulate filter that just barely meets the OBD limit in a legally prescribed driving cycle. The predicted signal change, according to the described method, outputs a limit value for the measured signal change depending on the actual operating points of the internal combustion engine. If the measured signal change and thus the actual mass of particles deposited on the particulate sensor is below the values ​​predicted for the limit filter, it can be concluded that the particulate filter is intact. If, on the other hand, the measured sensor signal exceeds the predicted signal change, it can be concluded that the particulate filter is defective, since more particles have reached the particulate sensor than are permissible according to the particulate filter.

[0018] The reliability of the detection of a defective particulate filter can be increased by concluding that a particulate filter is defective if the measured signal change of the particulate sensor is greater than the predicted signal change in several consecutive measuring cycles.

[0019] In a preferred embodiment of the invention, the particle mass flow predicted by the engine model, after multiplication by a predetermined vector of retention factors for the particle filter, is fed to the deposition model as a vector of predicted particle mass flows, a vector of predicted signal changes is calculated therefrom, and after comparing the vector of predicted signal changes with the measured signal change, a retention factor for the particle filter is determined. Based on the determined retention factor, it can be concluded whether a defective particle filter is present. The values ​​formed by multiplication by the vector differ from one another in that they deviate to varying degrees from the best estimate of the predicted particle mass flow from the engine model.By comparing the individual values ​​of predicted signal changes obtained from the vector with the measured signal change, the best individual values ​​in the vector of assumed retention factors can be determined, on the basis of which the statement about the degree of damage to the particulate filter is made. The actual retention factor can be determined by interpolating the neighboring retention factors that most closely predict the measured signal change. Using this method, it is not necessary to record all relevant influencing variables on the cross-sensitivities of the particulate sensor, such as exhaust gas volume flow, temperature of the particulate sensor, or particulate mass flow, over time for the integration interval; only a significantly smaller vector of the predicted signal change needs to be recorded.

[0020] In a further application of the method, it is provided that the method is used to adapt an emissions map of the engine model to determine the predicted mass of particles or the predicted particle mass flow. The predicted signal change of the particle sensor is compared with the measured signal change. If there is a deviation between the predicted signal change and the measured signal change, the emissions map of the engine model is corrected. Based on the predicted mass of particles or the predicted particle mass flow, a loading forecast of a particle filter can be carried out. The emissions map is corrected by increasing or decreasing the particle concentrations or particle mass flows stored there.This allows drifts over the lifetime of the internal combustion engine, for example due to injector drifts, to be detected and compensated.

[0021] A simple adaptation of the emission map, which can be implemented with little computational effort, can be achieved by correcting the engine model by multiplying all values ​​of the emission map of the engine model by an adaptation factor, whereby the adaptation factor is determined from the ratio between the measured signal change and the predicted signal change.

[0022] The accuracy of the emissions map adjustment can be improved by correcting the engine model by correcting values ​​of the engine model's emissions map at engine operating points that were reached in a previous period since the start of a measurement cycle. In this way, different correction factors can be incorporated into the emissions map for different engine operating conditions.

[0023] A further increase in accuracy can be achieved by correcting the engine model by correcting values ​​of the engine model's emission map at operating points of the internal combustion engine by correcting them with a function determined in a search algorithm that corrects the engine model until the measured signal change and the predicted signal change agree.

[0024] If the particulate mass flow predicted by the engine model is multiplied by a predefined vector of adaptation factors and fed into the deposition model as a vector of predicted particulate mass flows, a vector of predicted signal changes is calculated from this, and an adaptation factor for the emissions map stored in the engine model is determined after comparing the vector of predicted signal changes with the measured signal change, then the adaptation of the emissions map can be carried out with minimal memory requirements, since the operating points approached within an integration interval, along with the associated exhaust gas volume flows, temperatures, and particulate mass flows, no longer need to be recorded over time. It is sufficient to record the vector of the predicted signal change, which requires significantly less memory.This is particularly noticeable when adapting the emission map of the engine model, where the emission map is varied until the predicted and measured signal changes match, since this iterative process requires a particularly high amount of memory. Drawings

[0025] The invention is explained in more detail below with reference to the exemplary embodiments illustrated in the figures. They show: Fig. 1 an output signal of a resistive particle sensor, Fig. 2 a flow diagram for monitoring a particle filter, Fig. 3 a simplified flow diagram for monitoring the particulate filter, Fig. 4 a flow chart for monitoring the particulate filter using a trigger method, Fig. 5 a flow chart for adapting an engine model, Fig. 6 a flow diagram with temperature correction of a measured sensor signal, Fig. 7 another flow chart for monitoring the particulate filter using the trigger method, Fig. 8 a flow chart for correcting a predicted signal change of the particle sensor. Description of the embodiments

[0026] Fig. Figure 1 shows the output signal of a resistive particle sensor 1 (not shown here), such as is used in the exhaust duct of an internal combustion engine to determine the loading status of a diesel particulate filter or to monitor its function for on-board diagnostics. The particle sensor 1 has two comb-like, intermeshing electrodes exposed to the exhaust gas flow. If soot particles are deposited, the conductivity between the electrodes increases, and when an electrical voltage is applied to the particle sensor 1, a current can be measured. The deposited particle mass can be determined from the current level. Fig. 1, a measured sensor signal 10 is plotted along a time axis 40 on a current axis 41. As the sensor load increases, the measured sensor signal 10 increases. Fluctuations in the measured sensor signal occur due to cross-sensitivity of the particle sensor 1 with temperature. In addition to the absolute level of the current through the particle sensor 1, which is detected by the measured sensor signal 10, a signal change 11 measured over a defined period of time can also be used to evaluate the load of the particle sensor 1. After the particle sensor 1 has been burned off, loading initially does not lead to a measurable current flow because the deposited particles do not form a continuous current path between the electrodes. Therefore, a trigger threshold 42 is usually set, above which the measured sensor signal 10 is evaluated.This trigger threshold 42 is reached at a trigger time 43 after the particle sensor 1 has been burned free and represents a measure of the deposition rate on the particle sensor 1. This variant of the evaluation of the output signal of the particle sensor 1 is called the trigger method.

[0027] Fig. Figure 2 shows a flow chart for monitoring the diesel particulate filter using particulate sensor 1. In an engine model 60, a predicted particulate mass flow 22 is determined based on current operating parameters of the internal combustion engine, which is then fed to a limit filter model 61. The limit filter model 61 models which portion of the predicted particulate mass flow 22 is retained in a limit-value particulate filter and which portion penetrates it and leaves as a predicted particulate mass flow limit value 23. A deposition model 62 is used to model the deposition behavior of the particles present in the exhaust gas on the particulate sensor 1. An exhaust gas volume flow 31, an exhaust gas temperature 32, and a temperature 33 of the particulate sensor 1 are fed to the deposition model as influencing variables 30 on the cross-sensitivities. Instead of the exhaust gas volume flow 31, an exhaust gas velocity at the location of the particulate sensor 1 can also be used.From the influencing variables 30 and the predicted particle mass flow limit value 23, the deposition model 62 determines a predicted deposited particle mass flow 24, which is integrated in an integration stage 51 to a predicted deposited mass 25. The predicted deposited mass 25 is converted into a predicted signal change 20 in a conductivity model 63, taking into account the temperature 33. The conductivity model 63 takes into account, on the one hand, the current paths occurring due to the formation of the particle coating and the resulting electrical resistance of the arrangement, as well as the temperature dependence of the specific electrical resistance of the layer material.The predicted signal change 20, which represents the signal change to be expected for a borderline particulate filter due to the use of the boundary filter model 61, is compared with the measured signal change 11 in a comparison stage 50. If the measured signal change 11 is higher than the predicted signal change 20, a defective particulate filter is determined, and a diagnostic signal 70 is generated. It can be provided that the diagnostic signal 70 is only generated after several consecutive diagnoses of a defective particulate filter in order to avoid misdiagnoses.

[0028] In Fig. 3 is a opposite Fig. 2 shows a simplified flow chart for diagnosing the particulate filter. Here, too, the predicted particulate mass flow 22 determined in the engine model 60 is fed to the limit filter model 61, the predicted particulate mass flow limit value 23 is determined, and this is fed to the deposition model 62. However, in the deposition model 62, a predicted differential signal change 26 is directly determined, which is integrated in the integration stage 51 to form a predicted integral signal change 21. In a first temperature correction stage 52, the predicted signal change 20 is determined from the predicted integral signal change 21, which is compared in the comparison stage 50 with the signal change 11 measured at the particulate sensor 1.

[0029] The use of a trigger method to assess the suitability of the particulate filter is Fig. 4. The sensor signal 10 measured at the particle sensor 1 is fed to a threshold value stage 54, which evaluates the extent to which a predetermined threshold value is reached by the measured sensor signal 10. If the threshold value is reached, the threshold value stage 54 sends a time signal 72 to the comparison stage 50, in which the comparison is made between the predicted signal change 20 and the measured sensor signal 10, which corresponds to the threshold value. If the predicted signal change 20—this can also be the predicted signal change 20 since the last burn-off—is smaller than the threshold value, it is concluded that a particle filter is defective. In this embodiment, the comparison in the comparison stage 50 is performed only once, after each burn-off of the particle sensor 1, thus shortening the program sequence.

[0030] In addition to the application in the on-board diagnosis of particulate filters, the method according to the invention can also be advantageously used for adapting the engine model 60 for more precise determination of the predicted particulate mass flow 22. This application is described in Fig. 5. In the engine model 60, the predicted particle mass flow 22 is determined and fed to the deposition model 62, in which it is converted into the predicted differential signal change 26, taking into account the influencing variables 30 on the cross-sensitivities exhaust gas volume flow 31, exhaust gas temperature 32, and temperature 33 of the particle sensor 1. The predicted integral signal change 21 is determined by integration in the integration stage 51, from which the predicted signal change 20 is determined in the first temperature correction stage 52, taking into account the temperature 33.In the comparison stage 50, the predicted signal change 20 is compared with the measured signal change 11 of the particle sensor 1. Depending on the result of the comparison, an adaptation factor 71 is determined in an adaptation stage 55. This adaptation factor is incorporated into the engine model 60 and is designed such that in the next run of the method, the difference between the predicted signal change 20 and the measured signal change 11 of the particle sensor 1 is reduced, so that the engine model 60 thus better models the real conditions. In a simple embodiment, the adaptation factor 71 can be the ratio of the measured signal change 11 and the predicted signal change 20. However, it can also be provided to change only those operating points of an emissions map in the engine model 60 that were approached in the previous driving cycle.

[0031] A particularly good adaptation of the emissions map in the engine model 60 is achieved if the influencing variables 30 on the cross-sensitivity resulting from the operating parameters of the internal combustion engine are stored in a time-resolved manner, and the adaptation is continued until the predicted signal change 20 and the measured signal change 11 match. However, this can require a lot of storage space. This can be avoided by using the engine model 60 to determine not just a single value of the predicted particulate mass flow 22, but rather a vector of such values, each of which has a fixed ratio to the predicted particulate mass flow 22. The following calculations are each performed with all values ​​of the vector. In the comparison stage, it is determined which value from the vector best matches the measured value. Interpolation between values ​​from the vector can also be performed.This determines the ratio at which the measured and predicted values ​​best match. The advantage of this implementation is that the memory requirement is determined only by the number of elements in the vector.

[0032] In the embodiments of the invention presented so far, the temperature dependence of the resistance of the particle sensor 1 is taken into account in the predicted values ​​to enable comparison with measured values. In the embodiment according to Fig. 6, the temperature correction is applied to the measured signal. As in Fig. 4, the predicted differential signal change 26 is provided based on the operating parameters of the internal combustion engine and the emissions map stored in the engine model 60 after processing in the limit filter model 61 and the deposition model 62. From this, the predicted signal change 20, in this case not corrected by the temperature 33, is generated in the integration stage 51. The measured signal change 11 provided by the particle sensor 1 is converted in a second temperature correction stage 53, which takes into account the temperature 33 of the particle sensor 1, into a measured temperature-corrected signal change 13, which is compared with the predicted signal change 20 in the comparison stage 50. In this case, the comparison therefore takes place with signals characteristic of a selectable temperature.

[0033] In Fig. Figure 7 shows an embodiment of the onboard diagnostics according to the triggering method in which the temperature correction is performed on the measured sensor signal 10. As shown in Fig. 6, the predicted signal change 20 is generated based on the engine model 60. The particle sensor 1 generates the measured sensor signal 10, which is fed to the second temperature correction stage 53, which generates a measured temperature-corrected sensor signal 12, which is fed to the threshold value stage 54. If the measured temperature-corrected sensor signal 12 reaches the specified threshold value, the time signal 72 is generated, which causes the comparison stage 50 to compare the measured temperature-corrected sensor signal 12 with the predicted signal change 20 for the entire measurement cycle. If the measured temperature-corrected sensor signal 12 is higher than the predicted signal change 20, a defective particle filter is concluded, and the diagnostic signal 70 is triggered.

[0034] To take into account a further cross-sensitivity of the particle sensor 1, it can be provided that the accumulation of particles on a particle sensor 1 that has just been burned off is weighted differently in the forecast than the accumulation on a particle sensor 1 that is already covered with particles. Such an embodiment is described in the Fig. 8. As already mentioned in Fig.7, the predicted differential signal change 26 is determined based on the engine model 60. From the predicted signal change 20 since the last burn-off, which represents the loading of the particle sensor 1, this loading is evaluated in a weighting stage 57 and a weighting factor 73 is output, which takes into account the different deposit behavior with different loading of the particle sensor 1. In a multiplication stage 56, the predicted differential signal change 26 is multiplied by the weighting factor 73 and the weighted predicted differential signal change 27 is calculated, from which the predicted signal change 20 since the last burn-off is calculated in the integration stage 51. This is compared in the comparison stage 50 with the measured signal change 11 of the particle sensor 1.In an extension of the embodiment, additional factors influencing the weighting factor 73, such as the exhaust gas volume flow 31, the exhaust gas temperature 32, the temperature 33 of the particle sensor 1, or a particle mass flow in the exhaust duct, can be taken into account. In a further embodiment, the weighting factor can be applied to the predicted particle mass flow 22 instead of to the predicted differential signal change 26.

Claims

[1] Method for on-board diagnosis (OBD) of a particle filter in the exhaust system of an internal combustion engine by means of at least one resistive particle sensor (1) arranged in the exhaust system downstream of the particle filter, wherein - a predicted particle mass flow limit value (23) is determined on the basis of a predicted particle mass flow (22) provided by an engine model (60) and on the basis of a limit filter model (61), - the predicted particle mass flow limit value (23) is fed into a deposition model (62), and - a predicted signal change (20) is determined on the basis of a conductivity model (63), which represents the signal change to be expected for a borderline particle filter due to the use of the boundary filter model (61), - wherein in a comparison stage (50) the predicted signal change (20) is compared with a measured signal change (11) of the particle sensor (1) and - wherein a defective particle filter is concluded if the measured signal change (11) of the particle sensor (1) is greater than the predicted signal change (20), wherein the measured signal change (11) of the particle sensor (1) and / or the predicted signal change (20) of the particle sensor (1) are corrected taking into account influencing variables (30) on cross-sensitivities of the particle sensor (1). [2] Method according to claim 1, characterized by that the exhaust gas temperature (32) and / or the current temperature (33) of the particle sensor (1) and / or the exhaust gas volume flow (31) and / or the influence of the gaseous exhaust gas components nitrogen oxides, hydrocarbons and / or water vapor are taken into account as influencing variables (30). [3] Method according to one of the preceding claims, characterized bythat the influence of the current temperature (33) of the particle sensor (1) on the conductivity of the particle sensor (1) is taken into account in the conductivity model (63) or in a first temperature correction stage (52) for determining the predicted signal change (20) or that the influence of the current temperature (33) of the particle sensor (1) is taken into account in a second temperature correction (53) for determining the measured signal change (11). [4] Method according to one of the preceding claims, characterized by that the method is carried out at operating points of the internal combustion engine at which the mass of the particles or the particle mass flow can be measured or predicted with high accuracy. [5] Method according to one of the preceding claims, characterized bythat a weighting factor is determined as a function of a particle mass deposited on the particle sensor (1) and / or the exhaust gas temperature (32) and / or the current temperature (33) of the particle sensor (1) and / or the exhaust gas volume flow (31) and / or the predicted particle mass flow (22) and / or a measured particle mass flow, with which the predicted deposited particle mass flow (24) is corrected. [6] Method according to one of the preceding claims, characterized by that a defective particle filter is concluded if the measured signal change (11) of the particle sensor (1) is greater than the predicted signal change (20) in several consecutive measuring cycles.

Citation Information

Patent Citations

  • sensor for detecting particles and method for checking its function

    DE10133384A1

  • Sensor element for particle sensors and methods for its manufacture

    DE102004043122A1

  • monitoring of emission limit values

    DE102005034247A1

  • Integrated particle sensor operation, e.g. resistive particle sensor, for use in motor vehicle, involves determining exhaust gas flow rate in sensor, and considering cross sensitivity of sensor in comparison with different gas flows

    DE102005040790A1

  • Diagnostic method for an exhaust system of a combustion engine

    EP1242723B1