DE102014209810B4 - Method and device for detecting soot and ash loading of a particle filter - Google Patents

Abstract

Method for detecting a soot and ash load (23, 24) of a particle filter (13) as a component of an exhaust gas purification system in the exhaust line (11) of an internal combustion engine (10), wherein to monitor the particle filter (13) a differential pressure (19) between the inlet and outlet of the particle filter (13) is measured and this is evaluated in a diagnostic unit (18), characterized in that to determine the soot and ash load (23, 24) the temporal gradient of the measured differential pressure (19) on the particle filter (13) is correlated with an expected temporal gradient of a differential pressure (19) of an intact and unloaded reference particle filter and this correlation is subjected to different temporal trend analyses,
wherein a cross-correlation factor (21) is formed by forming a standardized cross-correlation from the gradient of the measured differential pressure (19) across the particle filter (13) and the gradient of the expected differential pressure (19) across the reference particle filter,
wherein the cross-correlation factor (21) is subjected to long-term averaging or filtering and the trend of the cross-correlation factor (21) is observed over several driving cycles and the ash loading (24) of the particle filter (13) is derived based on a characteristic curve stored in a diagnostic unit (18).

Description

State of the art

The invention relates to a method for detecting a soot and ash load of a particle filter as a component of an exhaust gas purification system in the exhaust system of an internal combustion engine, wherein a differential pressure between the inlet and outlet of the particle filter is measured to monitor the particle filter and this is evaluated in a diagnostic unit.

The invention further relates to a device, in particular a diagnostic unit, for carrying out the method according to the invention.

Emissions legislation, particularly in the USA and Europe, sets limits for the emission of particle mass and also particle number and concentration. In addition to the emission limits, diagnostic limits are also specified, and if these are exceeded, an error must be displayed. For this purpose, diagnostic functions are implemented in the vehicle which monitor the parts and components installed to reduce emissions during vehicle operation as part of this on-board diagnosis (OBD) and display a malfunction which leads to the diagnostic limits being exceeded.

The DE 11 2008 000 223 T5 discloses a method for monitoring the functionality of an exhaust aftertreatment system of a motor vehicle. This is done by analyzing measured values recorded by a sensor. The sensor measures the size of a parameter related to the exhaust gases flowing out of an exhaust aftertreatment device such as a catalyst or filter, which is part of the exhaust aftertreatment system.

The DE 100 59 683 B4 relates to a method for monitoring a particle filter system in the exhaust system of an internal combustion engine, wherein the loading state of the particle filter is determined using a first method and a second method, and wherein the quotient is formed from the loading states determined using these two methods,
characterized in that
the first method is independent of the sulphur content of the fuel and the second method is dependent on the sulphur content of the fuel, and
from the quotient the sulphur content of the fuel used for the internal combustion engine is derived, which influences the ageing of the particulate filter system.

The soot particles emitted by an engine, particularly a diesel engine, can be efficiently removed from the exhaust gas using a diesel particulate filter (DPF). Currently, a so-called wall-flow diesel particulate filter (DPF) is the state of the art. Due to its channels that are closed on one side and the porous filter material, up to 99% of the soot can be separated from the surfaces of its filter walls. One disadvantage is that the filter must be thermally regenerated from time to time, as such a filter becomes clogged with soot particles over time. This involves using measures inside or outside the engine to increase the temperature (> 600° C), and the accumulated soot in the filter is burned off when there is a certain excess of oxygen in the exhaust gas, as otherwise the exhaust back pressure would rise too much. In order to be able to regenerate the filter in a timely manner but not too frequently, suitable soot load detection is required.

In addition to the soot particles, ash residues also build up in the particle filter over its service life. These can come from non-flammable additives in the fuel or engine oil. After a long period of operation, these ash deposits lead to an increase in the exhaust back pressure of the particle filter. The ash cannot be removed by regeneration as described above. If a filter is heavily loaded with ash, this can have repercussions on the engine's combustion due to the increasing back pressure. This is why a high ash load must be detected and displayed.

Particle filters are not yet in series use in gasoline engines. Due to the tightened emissions legislation, particularly for gasoline direct injection engines, both internal engine measures and exhaust aftertreatment measures are being discussed by almost every vehicle manufacturer. For gasoline systems, exhaust configurations with a three-way catalyst installed close to the engine and a downstream, uncoated gasoline particle filter as well as coated particle filters (so-called 4-way catalysts = 3-way catalyst + particle filter) installed close to the engine are being discussed.

It is obvious to use the methods used in diesel systems to diagnose the particulate filter or to diagnose the soot load of the particulate filter, ie measuring the pressure increase using pressure sensors or measuring the particle mass behind the particulate filter using a particle sensor.

From the DE 10 2010 002 691 A1 For example, a method and a device for diagnosing a particle filter as part of an exhaust gas purification system in the exhaust system of an internal combustion engine is known, whereby a differential pressure between the inlet and outlet of the particle filter is measured to monitor the particle filter and this is evaluated in a diagnostic unit. The differential pressure across the particle filter is determined from two differential pressure measurements or two absolute pressure measurements. This improves on-board diagnostics and, in addition to determining soot loading, can also detect when the particle filter has been manipulated or even removed.

The DE 11 2009 001 451 T5 describes a method for compensating measurement inaccuracies in a filter, the method comprising: after a cleaning process on the filter, specifically a particle filter of a diesel internal combustion engine, with which it is intended to remove removable solid particles, in particular soot particles, from the filter, determining an estimated value of the load with removable solid particles remaining in the filter after the cleaning process, and determining a first dissimilar value by quantitatively detecting a difference between i) the estimated load with removable solid particles remaining in the filter after the cleaning process and ii) a reference value. Basically, the DE 11 2009 001 451 T5 a method by which a distinction can be made between soot and ash loading of a particle filter, whereby the determined value of the non-removable particle loading (including ash) is used as a correction value in order to diagnose the most accurate soot loading possible.

The problem with gasoline-powered vehicles is that the differential pressure at the particle filter is significantly lower than is the case with diesel vehicles. The reason for this is the significantly lower exhaust gas mass flow in gasoline engines and the different design of the particle filter hardware in gasoline vehicles due to the lower raw soot mass emissions. A direct correlation between absolute differential pressure and soot and ash loading is therefore not possible with gasoline engines.

It is therefore an object of the invention to provide a method with which a detection of the soot and ash loading of a particle filter is made possible, in particular in gasoline-powered vehicles, with the aim of initiating a regeneration of the particle filter or of detecting and indicating possible impairments caused by excessive ash loading in good time.

It is a further object of the invention to provide a corresponding device, in particular a diagnostic unit, for carrying out the method.

Disclosure of the invention

The problem concerning the method is solved by the features of claims 1 to 11.

According to the invention, the diagnostic method provides that, in order to determine the soot and ash loading, the temporal gradient of the measured differential pressure on the particle filter is correlated with an expected temporal gradient of a differential pressure of an intact and unloaded reference particle filter and this correlation is subjected to various temporal trend analyses. Compared to the prior art, the method has the advantage that a loading diagnosis of the particle filter and a distinction between ash and soot loading can be made even with very small absolute pressure differences.

Preferably, as provided in a variant of the method according to the invention, the expected value of the differential pressure of the reference particle filter is determined as a model depending on current operating variables. These are usually available in a higher-level engine control system, so that the expected value for the current differential pressure of the reference particle filter can be calculated from them with little application effort.

Advantageously, the temporal gradient of the differential pressure across the reference particle filter is calculated from a volume flow and/or its temporal gradient and a flow resistance of the reference particle filter. The flow resistance can be calculated in the diagnostic unit as a fixed Value can be stored or stored in a map storage unit depending on one or more parameters.

It is particularly advantageous in the case of high back pressures, i.e. in the case of a high soot or ash load and comparatively high exhaust gas mass flows, if, in addition to the linear influence of the volume flow and/or its temporal gradients, quadratic influences of the volume flow and/or its temporal gradients are also taken into account for the calculation of the temporal gradient of the differential pressure across the reference particle filter. This can improve the accuracy of the loading diagnosis.

In a further preferred method variant, it can be provided that the measured differential pressure across the particle filter and/or the expected differential pressure across the reference particle filter and/or the volume flow are low-pass filtered to determine the model differential pressure. This allows signal fluctuations caused by interference to be suppressed for the diagnosis, which increases the quality of the diagnosis.

The preferred method provides that a standardized cross-correlation factor is formed by forming a cross-correlation from the gradient of the measured differential pressure across the particle filter and the gradient of the expected differential pressure across the reference particle filter. This standardized cross-correlation factor is independent of the signal levels of the gradients and takes low values for insufficient correlation and high values for good correlation. This distinction can be used, for example, to detect a missing or defective particle filter, as described in a parallel application by the applicant.

The preferred loading diagnostic procedure provides in a further step that the cross-correlation factor is subjected to long-term averaging or filtering, usually over several tens of thousands of kilometers of driving distance, and the trend of the cross-correlation factor is observed over all relevant driving cycles and the ash loading of the particle filter is derived using a characteristic curve stored in a diagnostic unit. An increasing ash loading of the particle filter over a long driving distance leads to the flow resistance of the particle filter and thus also the cross-correlation factor increasing steadily with increasing driving distance, with short-term fluctuations being suppressed by this long-term averaging or filtering.

If the long-term averaging or filtering is designed in such a way that medium-term increases in the values due to the soot loading, which decrease again after regeneration, are not taken into account, the increase in the correlation values can be clearly attributed to an irreversible ash loading.

The method can also provide for a medium-term evaluation of the cross-correlation factor and the soot load of the particle filter to be derived using a further characteristic curve stored in a diagnostic unit, whereby this is corrected by proportions of the ash load that was previously determined. This medium-term averaging or filtering only takes into account corresponding increases in the correlation values over a few driving cycles, corresponding to a few thousand kilometers of driving distance, which corresponds to a typical soot load. After the regeneration process, the cross-correlation value decreases again.

It is particularly advantageous if, depending on the determined ash load, measures are taken to compensate for increasing back pressure on a cylinder filling of the internal combustion engine and/or, if a limit value for the ash load is exceeded, a warning display is activated, for example in order to replace the particle filter. The measures to compensate for the back pressure can in particular stabilize the engine combustion again.

The diagnostic procedure works particularly reliably if the diagnosis is carried out when certain dynamic criteria are reached and/or exceeded, particularly in the gradient of the expected differential pressure. The gradients of the exhaust gas mass flow, the exhaust gas volume flow, the engine speed or variables derived from them are also possible.

The previously described process variants also work reliably if the differential pressure and its temporal gradient are determined from the signals of two differential pressure sensors and/or two Absolute pressure sensors, which are arranged upstream and downstream of the particulate filter in the exhaust system, determine the pressure.

A particularly preferred application of the method, as described above, provides for the use of the method in a gasoline-powered internal combustion engine in which the exhaust system has at least one separate catalyst and a particle filter or a catalyst-particle filter combination or a catalytically coated particle filter, so-called 4-way catalysts, on which differential pressure sensors can also be installed on the catalyst housing. In these engines in particular, the volume flow is rather low, so that only small pressure differences can occur across such a gasoline particle filter, as described at the beginning, so that by applying the method according to the invention with its variants, a safe and reliable load diagnosis of soot and ash particles is made possible in particular here, so that appropriate regeneration strategies for the filter can be used and the expected future legal requirements can therefore also be met.

The problem relating to the device is solved in that the diagnostic unit has devices for carrying out the method with the features described above and in particular comprises calculation units for determining a cross-correlation factor from a temporal gradient of a measured differential pressure across the particle filter and a gradient of a model-determined differential pressure of a reference particle filter, calculation units for averaging or filtering the cross-correlation factor and comparator units for comparing the averaged or filtered cross-correlation factor with at least one characteristic curve that can be stored in the diagnostic unit. The functionality can be implemented in the diagnostic unit using software. The diagnostic unit can be designed as a separate unit or as an integral component of a higher-level engine control system.

• 1 beispielhaft ein technisches Umfeld für die Erfindung,
• 2 in schematischer Darstellung eine weitere Variante des technischen Umfeldes, in der das Verfahren angewendet werden kann,
• 3 eine schematische Darstellung zum Aufbau eines Partikelfilters,
• 4 in einem Verlaufsdiagramm schematisch den Verlauf des Wertes für einen Kreuzkorrelationsfaktor in Abhängigkeit einer Fahrstrecke des Fahrzeugs.

The invention is explained in more detail below using an embodiment shown in the figures. It shows:

• 1 an example of a technical environment for the invention,
• 2 in schematic representation another variant of the technical environment in which the method can be applied,
• 3 a schematic representation of the structure of a particle filter,
• 4 in a trend diagram schematically the course of the value for a cross-correlation factor as a function of the distance traveled by the vehicle.

1 shows a schematic of the technical environment in which the method according to the invention can be used. An internal combustion engine 10 is shown as an example, which is designed as a gasoline engine, wherein the exhaust gas of the internal combustion engine is discharged via an exhaust line 11 in which an exhaust gas purification system is arranged, which in the example shown is designed in several stages. In the flow direction of the exhaust gas (exhaust gas flow 14) in the example shown, a catalyst 12 is initially provided, which can be designed as a three-way catalyst, followed by a particle filter 13. Furthermore, exhaust gas probes and other sensors are usually arranged in the exhaust line 11, although these are not shown in this schematic principle drawing, and whose signals are fed to an engine control unit (Electronic Control Unit ECU).

In order to diagnose the soot or ash load of the particle filter 13, a differential pressure sensor 15 is provided according to the state of the art, with which the pressure difference (differential pressure 19) between the filter inlet and filter outlet of the particle filter 13 can be determined. The output signal of the differential pressure sensor 15 is fed to a diagnostic unit 18, in which a diagnosis of the soot or ash load and a necessary initiation of the regeneration of the particle filter 13 can be carried out as part of an on-board diagnosis (OBD). This diagnostic unit 18 can be part of the higher-level engine control unit (ECU).

2 shows an alternative technical environment. Instead of measuring the differential pressure 19 above the particle filter 13 using the differential pressure sensor 15, the differential pressure 19 is measured against the ambient pressure before and after the particle filter 13. Two independent differential pressure sensors 16, 17 are provided for both differential pressure measurements, the signals of which are sent to the diagnostic device 18 for evaluation. The differential pressure sensors 16, 17 are connected to the exhaust system 11 by hose connections or pipes.

In a variant not shown here, the differential pressure 19 across the particle filter 13 can be determined by means of an absolute pressure sensor before and after the particle filter 13. In principle, both pressure sensor types can also be used together, i.e. a differential pressure sensor 16 before the particle filter 13 and an absolute pressure sensor after the particle filter 13 or an absolute pressure sensor before the particle filter 13 and a differential pressure sensor 17 after the particle filter 13.

3 shows schematically a section of a filter element of a 1 and 3 shown intact particle filter 13. The filter element is made of a porous ceramic substrate, which forms the filter walls 13.1 of alternately closed inlet and outlet channels 13.2, 13.3. The channels 13.2, 13.3 are closed on one side by an impermeable closure 13.4, which is also made of the porous ceramic substrate. The arrows indicate the exhaust gas flow 14 through the filter element. The exhaust gas penetrates the porous filter walls 13.1 on its way from the inlet channels 13.2, which are open to an inlet of the particle filter 13, to the outlet channels 13.3, which are open to an outlet. Soot particles 13.5 and ash particles 13.6 carried in the exhaust gas are filtered out in the pores of the filter walls 13.1.

The filter element is designed by a suitable choice of the porous ceramic substrate in such a way that the filter walls 13.1 offer the lowest possible flow resistance to the exhaust gas flow 14 and at the same time have a high filtering effect for particles carried in the exhaust gas, in particular soot particles 13.5 and ash particles 13.6. A defect in the particle filter 13 can occur if at least some of the closures 13.4 are missing and the affected inlet and outlet channels 13.2, 13.3 are open on both sides. The defect can be caused by material defects or manipulation, for example. The exhaust gas can then flow through the filter element unfiltered without passing through the filter walls 13.1.

The catalyst 12 and the particle filter 13 can also be combined as a so-called four-way catalyst (FWC), which is a catalytically coated particle filter 13. The only requirement for the inventive method is that a differential pressure sensor 15 is installed above the particle filter 13 or the coated particle filter or two absolute pressure sensors are installed upstream and downstream of the particle filter.

The loading detection according to the invention is based on monitoring the particle filter 13 by correlating the temporal gradient of the measured differential pressure 19 across the particle filter 13 loaded with soot and ash to the expected temporal gradient of an empty particle filter 13. The expected value is determined from a model depending on current operating variables of the internal combustion engine 10. The essential steps of the diagnosis are explained below.

The measured differential pressure signal is first low-pass filtered to suppress the noise. The temporal gradient d(Δp _(k) )/dk of the signal is then determined, where k means the kth measurement. In parallel, a corresponding reference value d(Δp* _(k) )/dk is determined by calculating a temporal progression or the gradient of a pressure difference of an intact filter from an exhaust gas volume flow or its temporal gradient and the flow resistance of the intact or unloaded filter, the reference filter. This value or the volume flow entering it can optionally also be low-pass filtered.

wobei d(Δp_(k))/dk der gemessene Differenzdruckgradient und d(Δp*_(k))/dk der Referenzdruckgradient oder modellierte Druckgradient 26 darstellt. Der Referenzdruck p*_(k) berechnet sich aus dem Produkt von Abgasvolumenstrom und dem Strömungswiderstand R* des intakten, unbeladenen Referenz-Partikelfilters.Then, a standardized cross-correlation of the temporal gradients from the measured differential pressure Δp _(k) to the reference differential pressure Δp* _(k) is used to determine the extent to which the gradients of the current measured value and the reference value are similar. For this purpose, a cross-correlation factor 21 KKF (see 4 ) is formed according to the following relationship: $$\text{KKF}=\mathrm{\Sigma }\left(\text{d}\left(\Delta {\text{p}}_{\left(\text{k}\right)}\right)*\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)\right)/\mathrm{\Sigma }\left(\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)*\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)\right)$$

where d(Δp _(k) )/dk is the measured differential pressure gradient and d(Δp* _(k) )/dk is the reference pressure gradient or modeled pressure gradient 26. The reference pressure p* _(k) is calculated from the product of the exhaust gas volume flow and the flow resistance R* of the intact, unloaded reference particulate filter.

To assess whether the particle filter is present or installed correctly or is functioning properly, the initial value of the standardized cross-correlation, the cross-correlation factor 21 KKF, is compared with a threshold value previously determined and stored in the control unit or in the diagnostic unit 18. If the result is below the threshold value, which corresponds to only a slight or non-existent correlation, the particle filter 13 has been removed or is defective. If the result is above the threshold value, which corresponds to a good correlation, the particle filter 13 is present or intact.

This method section corresponds to a diagnostic method for detecting a missing or defective particle filter 13, as described in a parallel application of the applicant.

The method works particularly reliably when a certain dynamic excitation is present, i.e. when the differential pressure gradients exceed a certain level. Therefore, the cross-correlation is only evaluated when certain dynamic criteria are met. The gradients of the exhaust gas mass flow, the exhaust gas volume flow, the speed or variables derived from them are possible for this. Ideally, the gradient of the differential pressure reference value is used directly for this.

In an alternative arrangement, the differential pressure and its temporal gradient can also be determined from the signals of two differential pressure sensors, each measuring the differential pressure compared to the air pressure, or two absolute pressure sensors upstream and downstream of the particle filter 13, as shown schematically in 2 is shown.

mit d(ΔV_(k))/dk als Gradient vom Abgasvolumenstrom und dem Strömungswiderstand R* des intakten, unbeladenen Referenz-Partikelfilters ermittelt werden, sondern alternativ auch unter Berücksichtigung eines quadratischen Einflusses des Volumenstromgradienten nach der Beziehung $$\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)/\text{dk}=\left(\text{R1*}\times \text{d}\left(\Delta {\text{V}}_{\left(\text{k}\right)}\right)\text{/dk}\right)+\left(\text{R2*}\times {\left(\text{d}\left(\Delta {\text{V}}_{\left(\text{k}\right)}\right)\text{/dk}\right)}^2\right)$$

mit einem linearen Strömungswiderstandsanteil R1 und einem quadratischen Strömungswiderstandsanteil R2. Dieser quadratische Einfluss wird insbesondere bei hohen Gegendrücken, d.h. bei hoher Ruß- oder Aschebeladung 23, 24 (siehe 4) und bei hohen Abgasmasseströmen relevant.In contrast, the reference value d(Δp* _(k) )/dk can optionally be determined not only via the linear relationship as described above $$\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)/\text{dk}=\text{R*}\times \text{d}\left(\Delta {\text{V}}_{\left(\text{k}\right)}\right)/\text{dk}$$

with d(ΔV _(k) )/dk as the gradient of the exhaust gas volume flow and the flow resistance R* of the intact, unloaded reference particle filter, but alternatively also taking into account a quadratic influence of the volume flow gradient according to the relationship $$\text{d}\left(\Delta {\text{p*}}_{\left(\text{k}\right)}\right)/\text{dk}=\left(\text{R1*}\times \text{d}\left(\Delta {\text{V}}_{\left(\text{k}\right)}\right)\text{/dk}\right)+\left(\text{R2*}\times {\left(\text{d}\left(\Delta {\text{V}}_{\left(\text{k}\right)}\right)\text{/dk}\right)}^2\right)$$

with a linear flow resistance component R1 and a quadratic flow resistance component R2. This quadratic influence is particularly noticeable at high counter pressures, ie at high soot or ash loading 23, 24 (see 4 ) and relevant for high exhaust gas mass flows.

Based on the standardized cross-correlation factor 21 KKF determined in (1), in a further process step the standardized cross-correlation factor 21 KKF is subjected to various trend analyses with different averaging methods for the clear detection of soot loading 23 and/or ash loading 24.

As in 4 As shown, an increasing ash load 24 of the particle filter 13 over a distance 22 leads to an increase in the flow resistance of the particle filter 13 and thus also in the standardized cross-correlation factor 21 KKF. The increase due to the ash load 24 usually occurs very slowly over a very long running time of the engine, which corresponds to a distance 22 of several tens of thousands of kilometers, and is not reversible. The increase due to the soot load 23, on the other hand, occurs more quickly, e.g. over several thousand kilometers, and is also reversible, since the soot particles 13.5 are burned during filter regeneration. This gives a typical course of the cross-correlation factor 21 over the distance 22, as shown by way of example in the course diagram 20 in 4 is shown.

The evaluation procedure is based on the fact that the standardized cross-correlation factor 21 KKF is analyzed over many driving cycles in the diagnostic unit 18. For this purpose, a trend of the KKF values is observed over all relevant driving cycles via a strong averaging or filtering. Based on the long-term average of the standardized KKF values, an ash load 24 of the particle filter 13 is derived via a characteristic curve that is stored in the diagnostic unit 18. Depending on this ash load characteristic value, suitable strategies can then be applied in order to, for example, determine the influence of the increasing back pressure on the cylinder filling of the engine and also on the modeling of the same in the engine control unit. advises and/or in the diagnostic unit 18. In addition, depending on the ash loading characteristic value, the driver can be informed with the request to drive to the workshop if the influence of the ash loading 24 exceeds a critical value, ie the particle filter is excessively clogged. The averaging or filtering of the KKF values for determining the ash loading is preferably designed in such a way that the rather medium-term increases in the values as a result of the soot loading 23 have no significant influence on the determined ash loading value.

In parallel to this long-term averaging or filtering of the KKF values, a corresponding medium-term trend analysis of the KKF values is carried out using a comparatively weaker averaging or filtering of the KKF values over several driving cycles. The resulting value is corrected by the proportion of the ash loading correlation from the long-term average mentioned above. The soot loading 23 of the particle filter 13 is derived from the remaining proportion using a further characteristic curve, which is also stored in the diagnostic unit 18. Suitable soot regeneration strategies can then be applied depending on this soot loading characteristic value.

An alternative approach to the method provides that the flow resistance R itself can be determined from the gradients of the volume flow and the gradient of the measured differential pressure signal. From the flow resistance R determined in this way, the soot and ash loading 23, 24 of the particle filter 13 can also be determined via a suitable long-term evaluation.

In an advantageous embodiment, the diagnostic procedure is stored as software in the diagnostic unit 18 and can be used in particular in gasoline engines with future gasoline particle filters, but also in principle in diesel engines. This means that the future emissions regulations according to EU6b (2014) and EU6c (2017) can be met, especially in gasoline engines.

Claims (11)

Method for detecting a soot and ash load (23, 24) of a particle filter (13) as a component of an exhaust gas purification system in the exhaust line (11) of an internal combustion engine (10), wherein, in order to monitor the particle filter (13), a differential pressure (19) between the inlet and outlet of the particle filter (13) is measured and this is evaluated in a diagnostic unit (18), characterized in that , in order to determine the soot and ash load (23, 24), the temporal gradient of the measured differential pressure (19) on the particle filter (13) is correlated with an expected temporal gradient of a differential pressure (19) of an intact and unloaded reference particle filter and this correlation is subjected to different temporal trend analyses, wherein, by forming a standardized cross-correlation from the gradient of the measured differential pressure (19) across the particle filter (13) and the gradient of the expected differential pressure (19) across the reference particle filter, a A cross-correlation factor (21) is formed, wherein the cross-correlation factor (21) is subjected to long-term averaging or filtering and the trend of the cross-correlation factor (21) is observed over several driving cycles and the ash loading (24) of the particle filter (13) is derived on the basis of a characteristic curve stored in a diagnostic unit (18). procedure according to Claim 1 , characterized in that the expected value of the differential pressure (19) of the reference particle filter is determined as a model depending on current operating variables. Method according to one of the Claims 1 or 2 , characterized in that the temporal gradient of the differential pressure (19) across the reference particle filter is calculated from a volume flow and/or its temporal gradient and a flow resistance of the reference particle filter. procedure according to claim 3 , characterized in that in addition to the linear influence of the volume flow and/or its temporal gradients, quadratic influences of the volume flow and/or its temporal gradients are also taken into account for the calculation of the temporal gradient of the differential pressure (19) across the reference particle filter. Method according to one of the Claims 1 until 4 , characterized in that the measured differential pressure (19) across the particle filter (13) and/or the expected differential pressure (19) across the reference particle filter and/or the volume flow for determining the model differential pressure (19) are low-pass filtered. procedure according to claim 5 , characterized in that the long-term averaging or filtering is designed in such a way that medium-term increases in the values due to the soot loading (23), but values which decrease again after regeneration, are not taken into account. Method according to one of the Claims 5 until 6 , characterized in that a medium-term evaluation of the cross-correlation factor (21) is carried out and the soot load (23) of the particle filter (13) is derived from a further characteristic curve stored in a diagnostic unit (18), said soot load being determined by proportions of the ash load (24) which are claim 5 determined, be corrected. Method according to one of the Claims 1 until 7 , characterized in that , depending on the determined ash load (24), measures are taken to compensate for an increasing counterpressure on a cylinder filling of the internal combustion engine (10) and/or, if a limit value for the ash load (24) is exceeded, a warning display is activated. Method according to one of the Claims 1 until 8 , characterized in that the diagnosis is carried out when certain dynamic criteria are reached and/or exceeded, in particular in the gradient of the expected differential pressure (19). Method according to one of the Claims 1 until 9 , characterized in that the differential pressure (19) and its temporal gradient are determined from the signals of two differential pressure sensors and/or two absolute pressure sensors which are arranged upstream and downstream of the particle filter (13) in the exhaust system (11). Device, in particular a diagnostic unit (18), for detecting a soot and ash load (23, 24) of a particle filter (13) as a component of an exhaust gas purification system in the exhaust system (11) of an internal combustion engine (10), wherein for monitoring the particle filter (13) a differential pressure (19) between the inlet and outlet of the particle filter (13) can be measured and evaluated in this diagnostic unit (18), characterized in that the diagnostic unit (18) is set up to carry out a method according to the claims 1 until 10 and in particular comprises calculation units for determining a cross-correlation factor (21) from a temporal gradient of a measured differential pressure (19) across the particle filter (13) and a gradient of a model-determined differential pressure (19) of a reference particle filter, calculation units for averaging or filtering the cross-correlation factor (21) and comparator units for comparing the averaged or filtered cross-correlation factor (21) with at least one characteristic curve that can be stored in the diagnostic unit (18).

Images