DE102019212807A1 - Method for monitoring a particle filter by means of a particle sensor - Google Patents

Abstract

The invention relates to a method for monitoring a particle filter (16) by means of a particle sensor (20) which is arranged downstream of the particle filter (16) in an exhaust gas line (17) of an internal combustion engine (10), comprising: defining a first diagnostic phase (52) , comprising defining a soot emission model of a boundary particle filter arranged downstream of an internal combustion engine and determining, as a function of the soot emission model, after which period of time after a complete sensor regeneration (50) the particle sensor (20) can supply a first measurable signal when using a boundary particle filter, wherein the time period (t1) determined in this way defines the end of the first diagnostic phase (52); wherein the method further comprises: detecting output signals (40, 42) of the particle sensor, and, if a signal from the particle sensor occurs in the defined first diagnostic phase (52), determining that the particle filter (16) is damaged.

Description

The present invention relates to a method for monitoring a particle filter by means of a particle sensor as well as a computing unit and a computer program for its implementation.

State of the art

Internal combustion engines inevitably produce soot particles in the exhaust gas, for which increasingly strict limit values apply. This is why soot filters or particle filters are used in the exhaust gas aftertreatment system in diesel engines, but also in gasoline engines. The performance of a built-in particle filter in the exhaust gas flow can be monitored, for example, by suitable soot particle sensors, whose output signal is evaluated by a control unit in order to infer the soot particle content in the exhaust gas. Corresponding sensors can be arranged in front of and / or behind a particle filter for various purposes.

Such sensors can, for example, comprise two comb-like, intermeshing electrodes on a ceramic base substrate, in which an accumulation of conductive soot particles ensures that the impedance of the electrodes changes as a function of the accumulated amount of soot. With increasing particle concentration on the sensor surface, a decreasing resistance or an increasing current can be measured with a constant voltage applied between the electrodes. However, a certain minimum amount of deposited soot particles is necessary before a conductive connection and thus a measurable signal can be generated. Until this minimum amount of particles has built up on the sensor, the resistive particle sensor will not deliver a signal and then show a change in resistance, i.e. a current signal that is dependent on the applied measurement voltage.

Likewise, a resistive particle sensor can only deliver a meaningfully evaluable signal up to a certain maximum load; above this, the resistance will no longer decrease significantly or the measured current will no longer increase significantly. A particle sensor can therefore be regenerated by pyrolysis by heating it to a suitable temperature at which the deposited soot particles can be burned off. The sensor can be heated, for example, by heating elements on the back of the ceramic substrate. The heating elements can also be used to keep the sensor at a relatively constant operating temperature, since the conductivity is also influenced by the temperature. A measurement cycle for diagnostic purposes in the vehicle is usually started with such a regeneration of the sensor, and after a diagnostic result has been reached, complete regeneration is carried out again.

The future legislation on the monitoring of particle filters contains significantly lower limit values and provides that even small damage to a particle filter must be detected reliably and in good time in the on-board diagnosis (OBD). It is therefore necessary that a corresponding particle sensor must immediately provide a diagnostic result even with low soot concentrations in the exhaust gas. These limit values and response times cannot be achieved with the previous methods.

Disclosure of the invention

According to the invention, a method for monitoring a particle filter as well as a computing unit and a computer program for its implementation with the features of the independent patent claims are proposed. Advantageous refinements are the subject matter of the subclaims and the description below.

In particular, a method for monitoring a particle filter by means of a particle sensor is proposed, which is arranged in an exhaust gas tract of an internal combustion engine downstream behind the particle filter, a first diagnostic phase being defined, which defines a soot emission model of a boundary particle filter arranged after an internal combustion engine and determining in Depending on the soot emission model, after which period of time after a complete sensor regeneration the particle sensor can deliver a first measurable signal when using a boundary particle filter, the period of time thus determined defining the end of the first diagnostic phase. On this basis, output signals from the particle sensor are then recorded, and if a signal from the particle sensor occurs in the defined first diagnostic phase, it is determined that the particle filter is damaged. By defining a limit particle filter, i.e. a particle filter that is barely or no longer permissible on the basis of desired limit values and emission models, it can be determined very early in the first diagnostic phase whether a particle filter is more damaged than this limit particle filter, so that, for example, an exchange is necessary. Soot emission models are known in the art.

A second diagnostic phase can then be defined, which immediately follows the first diagnostic phase, with at least one output signal of the particle sensor being recorded and evaluated in at least a predetermined period of time within the second diagnostic phase to detect damage to the particle filter. In this second phase you can now Damages are detected that are below the damage level of the defined limit particle filter, since every increase indicates a renewed deposition of particles.

According to an exemplary embodiment, the expected soot concentration in the exhaust gas of the internal combustion engine is modeled for this purpose, and if there is an increase in the expected soot concentration, the specified time period for the detection of a sensor signal is set so that the time period with the start of the increase in the expected soot concentration begins. It can thus be determined whether or not an expected increase in the soot concentration in the exhaust gas is offset by a corresponding signal from the particle sensor, in order to monitor the functionality of the sensor on this basis.

It can be established, for example, that the particle filter is damaged if the increase (i.e. the value difference) of at least one recorded output signal in the predetermined period of time in the second diagnostic phase is above a predetermined limit value. A very small limit value (down to the value 0) can optionally be selected here or a limit value such that at least insignificant fluctuations do not lead to a diagnosis result being triggered. However, more than one output signal can also be evaluated for this purpose.

According to one embodiment, a sensor regeneration can be carried out, for example, after a predetermined number of evaluations of output signals from the particle sensor and / or in the event that the output signal from the particle sensor is above a maximum loading threshold value. A sensor regeneration can take place completely or partially. The sensor regeneration can, for example, be carried out in such a way that a residual load of the sensor with soot particles remains, which corresponds to the sensor load at the end of the first diagnostic phase. In this way, after the partial generation, the sensor is immediately back in the second diagnostic phase and can reliably measure the smallest signal increases without any time delay.

Optionally, a counter value can be stored which indicates the number of sensor signals evaluated so far in the second diagnostic phase; additionally or alternatively, a time period can also be stored which indicates the time period since a last complete and / or partial sensor regeneration. In this way, even after an interruption of the measurement cycle, e.g. by switching off the vehicle, the monitoring can be continued directly in the previous diagnostic phase when the system is restarted. This is particularly suitable for brief interruptions in the operation of the combustion engine, e.g. in hybrid vehicles.

A measure can preferably be taken, for example a warning signal can be output if it is established that the particle filter is damaged. In addition, at least some of the detected sensor signals can optionally also be output via a diagnostic interface of the control unit.

A computing unit according to the invention, for example a control unit of a motor vehicle, is set up, in particular in terms of programming, to carry out a method according to the invention.

The implementation of a method according to the invention in the form of a computer program or computer program product with program code for performing all method steps is advantageous, since this causes particularly low costs, especially if an executing control device is used for other tasks and is therefore available anyway. Suitable data carriers for providing the computer program are, in particular, magnetic, optical and electrical memories, such as hard drives, flash memories, EEPROMs, DVDs, etc. A program can also be downloaded via computer networks (Internet, intranet, etc.).

Further advantages and embodiments of the invention emerge from the description and the accompanying drawing.

The invention is shown schematically in the drawing using exemplary embodiments and is described below with reference to the drawing.

Figure list

• 1 Figure 3 shows schematically a system in which embodiments of the invention may be used;
• 2 shows the signal curve of a resistive particle sensor according to an embodiment of the invention with a damaged filter; and
• 3rd shows the signal curve of a resistive particle sensor with an undamaged filter.

Embodiment (s) of the invention

1 Figure 12 schematically shows a system in which exemplary embodiments of the invention may be applied. An internal combustion engine 10 , which can be designed as a diesel engine, receives combustion air via an air supply 11 fed. The exhaust gas of the internal combustion engine 10 is via an exhaust system 17th dissipated in the at least one particle filter 16 is arranged. Furthermore, in the exhaust system 17th various sensors can be arranged, their signals to an engine control 14th be supplied, such as a temperature sensor 15th , a particle sensor placed in front of the filter 12th and an exhaust gas probe designed as a lambda probe (not shown). Downstream after the particle filter there is a resistive particle sensor 20th trained particle sensor arranged. Also the particle filter 16 can with the engine control 14th or another control unit.

When a measuring voltage is applied, the particle sensor is able to output a signal that reflects the load on the sensor with soot particles. The signal can be output as a current or as a resistance of the particle sensor element. Further components of the exhaust gas tract and the exhaust gas aftertreatment system, such as catalytic converters, are not shown in this figure.

The engine control 14th determines, for example, on the basis of the data supplied to it, an amount of fuel that is controlled via fuel metering 13th the internal combustion engine 10 can be fed. The engine control 14th can, for example, further process and evaluate the sensor signals, for example by means of a suitable software module and / or control the heating of the respective elements for filter and / or sensor regeneration, as will be described below.

The devices shown are used to monitor the particle emissions of the internal combustion engine 10 (On-board diagnosis) and a prognosis of the loading of the particle filter are possible. The particle filter can be, for example, a wall-flow filter or a bypass filter.

While particle filters are often used in diesel engines, gasoline engines are also increasingly being fitted with particle filters. The present invention can be used both for monitoring diesel particulate filters and petrol particulate filters or comparable filter systems.

In order to achieve a more sensitive detection of defects and damage to a particle filter, a two-phase evaluation of the signal from a particle sensor can be used according to a first embodiment of the invention.

A first diagnostic phase is defined in that the expected particle load is determined for a theoretical particle filter which, according to the limit values to be applied, is, for example, still considered to be intact. This modeled particle filter is used as a borderline case (borderline particle filter). On the basis of this boundary particle filter with marginal damage, it is then determined after which period of time on the particle sensor used 20th soot would be deposited, but the particle load would be just too low to lead to a measurable sensor signal. _{In this way, a period t 1} is obtained as the limit of the first diagnosis phase, which indicates the time after which a first sensor signal would be measurable in the case of a limit particle filter. A model of the particle emissions of the internal combustion engine can be used, among other things, to calculate the duration t ₁ 10 can be used depending on the operating point, which in turn can be constructed from modeled and / or recorded data. The termination of a complete sensor regeneration can be used as the starting time t _{0 of a measurement cycle or a first diagnosis phase.}

If, in a diagnostic operation, a measurable signal at the particle sensor already occurs in this first phase, that is to say before the end of the time period t ₁ 20th can be detected, this indicates that the particle load is even higher than with the predefined limit particle filter. From this it can then be concluded immediately that the measured particle filter should be considered defective because it has a poorer filter behavior than the limit particle filter. Usually, increases in current during the first phase indicate a high level of damage to the particle filter.

In the case of a marginally intact particle filter or an even better filter, on the other hand, the first phase from time t ₀ to t _{1 should run} without a measurable sensor signal.

After the period t _{1 has} elapsed, a second diagnostic phase begins. In this phase, diagnostic results can be obtained in that the sensor signal is recorded over a period of time Δt, for example the current output by the particle sensor. A model of the soot emissions of the engine can be used in order to determine an increase Δs in the particle emissions to be expected in this period. If at least some of these emissions are not retained by the particle filter and are recorded by the particle sensor, a measurable sensor signal ΔI can be expected at these times in the second diagnostic phase, since the minimum sensor load for signal generation is already present at the end of the first diagnostic phase at time t ₁ . The measure of the change in current .DELTA.I is preferably also a measure of the damage to the DPF, so that greater damage can be assumed in the case of a high signal change.

2 shows a signal curve over time of the output signal of a resistive particle sensor 20th behind a marginally damaged particle filter (ie a particle filter that corresponds to the boundary particle filter) together with the corresponding soot concentration in the exhaust gas flow 17th . The top line 30th represents the soot concentration, while the lower line 40 the sensor signal shows. The emissions can be calculated using a suitable model of the engine, taking the operating conditions into account. At the beginning of the curve shown (after reaching the end of the dew point) the sensor is completely regenerated 50 instead of. This regeneration phase 50 can take place when the engine is started or independently. During a regeneration phase, the particle sensor can be heated to a higher temperature at which the deposited soot can be burned off. The temperature required for this depends, among other things, on the gas components present. The regeneration phase 50 can from a control unit connected to the sensor 14th initiated and controlled. This can be an engine control unit, for example 14th act, for example via a bus system with the sensor 20th and other elements can be connected, but also another suitable control unit can take over these tasks. After the sensor regeneration is complete 50 the measuring cycle begins for diagnostic purposes.

As in 2 can be seen, the sensor signal increases 40 in the first diagnostic phase 52 By definition, it does not occur until time t ₁ despite the corresponding soot concentration, since the diagnosis phases are determined on this basis. As soon as a measurable signal can be expected for a limit particle filter, the first diagnostic phase ends 52 (Time t ₁ ) and the second diagnostic phase 53 begins. With at least one modeled increase Δs _1, Δs ₂ , ... in the soot concentration 30th As can be seen in the short marked sections of the upper curve, the associated sensor signals ΔI ₁ , ΔI ₂ , ΔI ₃ , of the particle filter 20th detected. For a limit particle filter, a measurable sensor signal can also be expected here, i.e. an increase in current. The time over which the increase in the sensor signal is measured is the duration Δt. This diagnosis duration Δt is variable and depends on the magnitude and duration of the positive change in the soot concentration. The diagnosis is started in each case when the positive gradient of the soot concentration exceeds a predetermined limit value and ends when the positive gradient of the soot concentration falls below a predetermined limit value. The probability of a positive change in the soot concentration and a correspondingly evaluable diagnostic result increases if the starting soot concentration is low, for example during the transition from engine idling / overrun / engine stop phase to strong acceleration of the vehicle.

The signal increase .DELTA.I in this period can then be in a corresponding control unit 14th are evaluated and either only evaluated as a qualitative indication of the presence of a certain damage or are evaluated quantitatively, whereby, for example, an association between the amount of current increase ΔI for a given increase in soot concentration Δs and the amount of damage to the particle filter can be found or specified can be, for example as the ratio ΔI / Δs. Subsequently, with the next increase Δs _{2 in} the modeled soot concentration, a further diagnostic result can be obtained in the same way, that is, again by comparing the sensor signal ΔI ₂ with the respectively expected soot concentration Δs ₂ .

Alternatively, the soot concentration can also be monitored or modeled without continuous monitoring 30th the sensor signal 40 are detected in a similar manner, with an increase in the sensor signal then as well 40 to a new particle load and thus to damage to the particle filter 20th indicates. In this case, the sensor signals can be recorded, for example, at regular time intervals or continuously, without necessarily using changes in the soot concentration as a trigger.

Preferably in the second diagnostic phase 53 several sensor signals ΔI ₁ , ΔI ₂ , ΔI ₃ , ... or diagnostic results 54 , 55 , 56 , 57 can be evaluated without the sensor 20th is regenerated between the individual diagnostic results. The particle sensor can therefore be in the previously described second diagnostic phase 53 remain and continue to accumulate soot. The associated sensor signal 40 will continue to rise accordingly in absolute values, at least until the maximum measurable load is reached. Usually, an approximately linear increase in relation to the particle loading will be expected. Depending on the type of sensor, the load and other conditions such as the temperature, a different course can also be present.

In this case, for example, the number n of diagnostic results can be specified that is to be evaluated before the sensor is regenerated. This can be done, for example, by a simple counter that is increased by one with each diagnostic evaluation and with a sensor regeneration 58 is reset. Such a counter (or another implementation of a predetermined number of diagnostic results) can also be stored in a non-volatile memory element, so that a vehicle is in the second phase, for example 53 can be switched off and the next time the vehicle is started, the diagnosis is carried out immediately in the second phase 53 can be continued correctly without performing a sensor regeneration beforehand. In addition or as an alternative, the modeled soot load can also be used 30th in the exhaust gas flow and the time period t ₁ determined therefrom up to the end of the first phase are stored together with the time already elapsed in this phase when a vehicle is parked, so that in this case the diagnosis phase is continued 52 , 53 becomes possible immediately after the start. Special parameters can also be incorporated into the modeling, which take into account, for example, the restart of the engine and its effect on the soot load.

When it is determined that a sensor regeneration 58 should take place because, for example, the specified number of diagnostic events has already been recorded or the current of the sensor signal 40 is above a predetermined limit value, a complete or partial regeneration 58 of the sensor element. For example, partial regeneration can only take place until conductive connections are made on the sensor 20th have burned off, but a residual charge remains on the sensor. This means that the sensor after the partial regeneration 58 can preferably still be loaded to such an extent that a measurable increase in the sensor signal is immediately evident in the event of a further particle loading 40 occurs. With a partial regeneration measured in this way, the sensor is immediately in a further second diagnostic phase 53 , in which the further evaluation of one or more diagnostic results can then take place as described above. However, a more extensive partial regeneration up to a complete regeneration, ie a burn-off of the entire soot load on the sensor, can also take place.

The different triggers for regeneration of the sensor element 20th such as exceeding a limit value for the measured current 40 and / or the achievement of a maximum number n of diagnostic results can also be combined with one another as desired. For example, according to one embodiment, the detection of a current above the predefined limit value can always be a regeneration 58 trigger. As a limit value, for example, the current that corresponds to a soot load on the sensor can be set accordingly 20th above the measuring limit. Different regeneration cycles can also be defined for different triggering events so that, for example, if the current is too high, a complete regeneration can take place, while only a partial regeneration is carried out after a maximum number of diagnostic results. However, the same regeneration process can also be used for all triggers. In addition, a sensor regeneration can optionally also be triggered by further events, such as switching off a vehicle, exceeding a limit value in a modeled soot load, and / or a sensor regeneration that takes place simultaneously with a filter regeneration. In one possible embodiment, it could be specified that after a predetermined number of runs of the second diagnostic phase with corresponding partial regenerations of the filter 16 a full regeneration is carried out, after which a new first diagnostic phase 52 entry. This can be done in the control unit 14th can be saved in which diagnosis phase 52 , 53 the sensor system is currently located. Likewise, it could also be specified that a sensor regeneration takes place at least when, or only when, a regeneration of the filter is also carried out 16 is triggered as long as the current 40 or the sensor load is below the maximum load limit.

In 3rd is an exemplary sensor signal for comparison 42 if there is an intact particle filter 16 shown, again being the sensor current 42 and the soot concentration 30th are plotted over time. The course of the soot concentration in this example corresponds to that in 2 shown, and also the diagnostic phases defined using a limit particle filter 52 , 53 and regeneration phases 50 , 58 are correspondingly identical. In this case of an undamaged or hardly damaged particle filter, the signal 42 in the first diagnostic phase 52 identical to the case 40 be a limit particle filter, namely in both cases in the first diagnostic phase 52 no signal should appear. In the first phase, only a statement about whether a particle filter is worse than a limit particle filter is therefore meaningful.

In contrast to the example in 2 but is now also in the second diagnostic phase 53 no sensor signal or at least only a very small sensor signal (eg below a threshold value that is valid as a relevance limit) can be measured.

This indicates that the upstream particle filter 16 works properly and the particles so far out of the exhaust stream 17th be filtered that at the downstream particle sensor 20th no significant amounts of soot are deposited. As in the previous example, a standard partial regeneration can then optionally be performed after a predetermined number n of diagnostic results 58 of the sensor. Alternatively, it is also possible to have such a regular regeneration 58 partially suspend, especially if based on the sensor signals (or the absence of signal increases) it can be concluded that there is no soot load above the limit value of the first phase 52 also exists. In this case, a partial or complete sensor regeneration could, for example, only then 58 be carried out when the measured sensor current 42 rises above the specified limit value. For example, only partial regeneration then 58 of the sensor performed when the measured sensor current 42 rises above the specified limit value, the same (partially loaded) initial state is always achieved.

It is also possible to provide regular partial regenerations of the sensor after a predetermined number of diagnostic results and / or to only carry out these partial regenerations if a measurable sensor current occurred at least once in the previous diagnostic phase, which indicates incipient damage or insufficient filter effect.

After a complete sensor regeneration, the diagnostic method can go back to the beginning of the first diagnostic phase, regardless of the reason for the regeneration 52 be reset, so that again after a predetermined period of time t ₁ , which corresponds to the measurable loading of a limit particle filter, the second diagnostic phase 53 begins.

The number of diagnostic results in the second diagnostic phase, i.e. the number of detected sensor signal increases and their optional evaluation, can be used in each of the repeated second phases 53 be the same or differ due to certain conditions as stated above. There should be at least one diagnostic result 54 can be recorded, but significantly more results can be obtained 55 , 56 , 57 in a second diagnostic phase 53 before a sensor regeneration is carried out, e.g. 5 to 10 results. It can be assumed that the current does not increase sharply, so that the sensor can be operated for a long time in the second diagnostic phase, even with the operation of a limit particle filter. In general, there does not have to be an upper limit for the number of diagnoses carried out per diagnosis phase as long as the sensor is still able to output a meaningful signal and is below its maximum measurable load.

Each of these diagnostic results achieved 54 , 55 , 56 , 57 can be evaluated individually, and / or information from several diagnostic results can also be evaluated, e.g. in the form of an average value from several signal increase values ΔI ₁ , ΔI ₂ , ΔI ₃ , ... from several diagnosis periods Δt ₁ , Δt ₂ , Δt ₃ , ... or by considering the highest current increase max (Δl ₁ , ..., Δl _n ) during a second diagnostic phase 53 . Other evaluation options are also possible. It is also possible that certain diagnostic results are excluded from the evaluation, for example if using a temperature sensor 15th an exhaust gas temperature outside the appropriate operating conditions has been determined. Part of the diagnostic results from a second phase 53 could continue to be used. It is also conceivable to use diagnosis results from several diagnosis phases, in particular from several second diagnosis phases taking place one after the other with partial regeneration carried out in between 58 to evaluate together and, for example, to gain further information about the filter status from an increase in the individual sensor signal increases over time.

In the event of a detected filter defect or an at least momentarily inadequate filter function of the particle filter 16 a corresponding signal can be generated by the control unit. This signal can, for example, be passed on via interfaces in order to enable diagnosis on a connected device and / or can lead to a visible or audible warning signal for the driver. For example, a warning lamp or warning display can indicate that a filter replacement or at least a check is required. Optionally, the diagnosis results obtained can be partially or completely stored so that, for example, in a workshop, by reading out these values via an interface of the control unit, further information about the degree of damage can be obtained. It can be determined how many diagnostic results a measurable sensor signal has to produce in order to lead to the determination of filter damage and a corresponding warning signal, so that, for example, short-term fluctuations do not have to trigger a false alarm.

A suitable particle sensor can be used for all embodiments 20th immediately behind the particle filter to be monitored 16 be arranged, or there can be further elements of the exhaust gas aftertreatment system (for example one or more catalytic converters) between the filter and the sensor. Further catalytic converters or other exhaust gas aftertreatment elements can also be suitably arranged upstream or downstream of the particle filter system. The particle filter can also be combined with a catalytic converter element (for example in an SCR-F system, selective catalytic reduction on filter). Several particle filters can also be used 16 with individually assigned particle sensors 20th exist, or if there are several particle filters, only one particle sensor can be used 20th can be used for monitoring. Diagnostic information obtained from the particle sensor can also be combined with further diagnostic results in a control unit 14th can be combined, for example with a measured pressure or counter pressure in front of or behind a filter system.

Claims (14)

Method for monitoring a particle filter (16) by means of a particle sensor (20), which is shown in FIG an exhaust line (17) of an internal combustion engine (10) is arranged downstream behind the particulate filter (16), comprising: defining a first diagnostic phase (52), comprising: establishing a soot emission model of a boundary particulate filter arranged downstream of an internal combustion engine; - Determination, as a function of the soot emission model, after which period of time after a complete sensor regeneration (50) the particle sensor (20) can deliver a first measurable signal when using a boundary particle filter, the time period (t ₁ ) thus determined marking the end of the first diagnostic phase ( 52) defined; wherein the method further comprises: detecting output signals (40, 42) of the particle sensor, and, if a signal from the particle sensor occurs in the defined first diagnostic phase (52), determining that the particle filter (16) is damaged. Procedure according to Claim 1 , further comprising: defining a second diagnostic phase (53), which immediately follows the first diagnostic phase (52), and detecting and evaluating at least one output signal (ΔI _1, ΔI ₂ , ΔI ₃ , ...) of the particle sensor (20) in at least a predetermined period of time (Δt _1, Δt ₂ , Δt ₃ , ...) within the second diagnostic phase (53) to detect damage to the particle filter (16). Procedure according to Claim 2 , further comprising: modeling the expected soot concentration (30) in the exhaust gas of the internal combustion engine, and if there is an increase in the expected soot concentration (Δs ₁ , Δs ₂ , Δs ₃ , ...), determining the predetermined time period (Δt _1, Δt ₂ , Δt ₃ , ...) so that the period begins with the beginning of the increase in the expected soot concentration. Procedure according to Claim 2 or 3rd , whereby it is established that the particle filter (16) is damaged if the increase (ΔI _1, ΔI ₂ , ΔI ₃ , ...) of at least one output signal detected in the specified period (Δt _1, Δt ₂ , Δt ₃ , ...) ) is above a predetermined limit value in the second diagnostic phase (53). Method according to one of the Claims 2 to 4th , further comprising: carrying out a sensor regeneration (58) after a predetermined number (n) of evaluations of output signals (ΔI _1, ΔI ₂ , ΔI ₃ , ...) of the particle sensor. Method according to one of the Claims 2 to 5 , further comprising: executing a sensor regeneration (58) if the output signal of the particle sensor (40, 42) is above a maximum loading threshold value. Method according to one of the Claims 5 or 6th , the sensor regeneration (50, 58) being carried out in such a way that a residual load of the particle sensor (16) with soot particles remains, which _{corresponds to the sensor load at the end (t 1} ) of the first diagnostic phase (52). Method according to one of the Claims 2 to 7th , further comprising: storing a counter value which indicates the number of previously evaluated sensor signals (ΔI ₁ , ΔI ₂ , ΔI ₃ , ...) in a second diagnostic phase (53). A method according to any one of the preceding claims, further comprising: Saving a period of time which indicates the period of time since a last complete and / or partial sensor regeneration (50, 58). A method according to any one of the preceding claims, further comprising: Output of a warning signal if it is determined that the particle filter (16) is damaged. Method according to one of the preceding claims, further comprising: outputting at least some of the detected sensor signals (ΔI _1, ΔI ₂ , ΔI ₃ , ...) via a diagnostic interface. Computing unit (14) which is set up to carry out all method steps of a method according to one of the preceding claims. Computer program that causes a computing unit (14) to perform all method steps of a method according to one of the Claims 1 to 11 to be carried out when it is executed on the processing unit. Machine-readable storage medium with a computer program stored thereon Claim 13 .

Images