D E S C R I P T I O N

On-Board-Diagnosis Method for an Exhaust Aftertreatment System and On-Board- Diagnosis System for an Exhaust Aftertreatment System

TECHNICAL FIELD

The invention relates to an on-board-diagnosis method for an exhaust aftertreatment system and an on-board-diagnosis system for an exhaust aftertreatment system according to the preambles of the independent claims.

BACKGROUND OF THE INVENTION

Present regulatory conditions in the automotive market have led to an increasing demand to improve fuel economy and reduce emissions in present vehicles. These regulatory conditions must be balanced with the demands of a consumer for high performance and quick response for a vehicle.

A diesel engine has an efficiency of up to about 52% and is thus the best converter of fossil energy. NO _x emission concentration, i.e. the emission of nitrogen oxides NO and NO ₂ , is dependent upon local oxygen atom concentration and the local temperature. Said high efficiency is however only possible at an elevated combustion temperature at which high NO _x levels are inevitable. Moreover, a suppression of NO _x formation by internal means (air/fuel ratio) has the tendency to cause an increase in particulates, known as the NO _x -particulates trade off. Furthermore, an excess of oxygen in the exhaust gas from a diesel engine prevents the use of stoichiometric 3-way-catalyst technology for reduction of NO _x as is used in gasoline engine cars from the late 80-ties.

Both carbon particulates and NO _x are typical emissions in the exhaust gas of diesel engines. Requirements for reducing such emissions increase and trigger various approaches in the art to reduce emissions. In the European patent EP 1 054 722 B1 an exhaust aftertreatment system is disclosed which combines a particulate filter collecting soot and nitrogen-oxides reduction catalysts in the exhaust tract. For removing soot NO ₂ is generated by oxidation of NO in an

oxidation catalyst. Soot which is collected in a particulate filter is oxidized by NO ₂ . Residual amounts of NO and NO ₂ in the exhaust gas are reduced to nitrogen gas in a selective-catalytic-reduction catalyst (SCR catalyst) by injecting ammonia into the SCR catalyst.

During operation all catalysts degrade due to accumulation of poisons, thermal migration of the catalyst material etc. This degradation seriously influences the operation of aftertreatment systems. Therefore it is desirable to detect the degradation of a catalyst in the aftertreatment system before the operation of the aftertreatment system fails or legal requirements cannot be fulfilled because of the degradation. This is done by the so called OBD (On-Board Diagnosis).

SUMMARY OF THE INVENTION

It is an object of the invention to provide an improved on-board-diagnosis method for an exhaust aftertreatment system. It is another object of the invention to provide an adequate improved exhaust aftertreatment system.

The objects are achieved by the features of the independent claims. The other claims and the description disclose advantageous embodiments of the invention.

An on-board-diagnosis method for an exhaust aftertreatment system of an engine is proposed comprising a diesel particulate filter unit which encompasses a particulate filter (DPF) in which soot can be oxidized by NO ₂ and constituents of the exhaust gas are deoxidized in a selective-catalytic-reduction (SCR) catalyst arranged downstream of the diesel particulate filter unit, wherein the exhaust gas flows from the diesel particulate filter unit to the selective-catalytic-reduction catalyst. According to invention an NO ₂ content upstream of the nitrogen-oxides reduction unit is estimated; at least one of an NO _x or NO ₂ content upstream of the nitrogen-oxides reduction unit is measured; a comparison between the estimated and the measured contents is performed; and based on the result of the comparison it is decided on at least one conditioning step of the diesel particulate filter unit.

Favourably, the proper operation of the diesel particulate filter unit can be reliably checked on board of a vehicle which is equipped with said aftertreatment system, improving the environmental compatibility of diesel engines. This can be done based on the NO2 level in the exhaust gas on particular locations and/or based on the efficiency of the nitrogen-oxides reduction unit.

Preferably for high load operation conditions of the engine, the following steps can be performed: measuring the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; calculating the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring an NO _x content in the exhaust gas downstream of the nitrogen-oxides reduction unit; measuring or calculating an NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; comparing the measured and the expected NO ₂ contents; and based on the result of the comparison deciding on at least one conditioning step of the diesel particulate filter unit.

Favourably the operability of the diesel particulate filter unit can be determined with high accuracy.

Preferably for low load operation conditions of the engine, the following steps can be performed: calculating the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring or measuring and calculating an NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring the NO _x content in the exhaust gas downstream of the nitrogen-oxides reduction unit; and deriving the efficiency of the NO _x conversion in the nitrogen-oxides reduction unit from the calculated NO ₂ and the measured and calculated NO _x contents. Favourably, a real NO ₂ sensor is not necessary as the NO ₂ content can be calculated, this embodiment thus needs less hardware and is cost efficient. If the efficiency of the nitrogen-oxides reduction unit is not sufficient, a separate diagnosis for the nitrogen-oxides reduction unit as well as for a reducing-agent system (in case a SCR catalyst is used with injection of a

reducing agent), can be triggered. This can be done after the regeneration related to the diesel particulate filter unit or before the regeneration.

Preferably, a conditioning step can be performed if the difference between the compared NO _x conversions and/or the compared NO ₂ contents are beyond a predetermined value.

Particularly, the NO _x conversion or the compared NO ₂ contents in the exhaust gas can be derived by determining the NO ₂ content with a real sensor and/or a virtual sensor.

Deriving the NO _x conversion in the exhaust gas can be advantageously done by determining an efficiency of the nitrogen-oxides reduction unit.

Another advantage with the present invention is that one may achieve high passive regeneration and HC oxidation in the DPF system and maintaining a good NO ₂ /NO ratio for high NO _x -conversion in the SCR-system for a fresh as well as an aged system. Advantageously, the preferred arrangement allows to using a smaller SCR-catalyst, giving both cost, space and weight benefits. The diesel particulate filter unit (DPFU) may have the oxidation catalyst (DOC) upstream of the diesel particulate filter (DPF). Alternatively or additionally, the DPF may comprise a catalytic coating which oxidizes exhaust gas components and which can replace or support the DOC. An advantage with this embodiment is that one will still further save space, cost and weight.

Preferably, a conditioning step can be performed if the differences between the compared NO _x conversions and/or the compared NO ₂ contents are beyond a predetermined value.

The NO _x conversion and/or the NO ₂ contents can be derived by determining the NO ₂ content in the exhaust gas with a real sensor and/or a virtual sensor. Alternatively or additionally, NO _x conversion can be derived from the efficiency of the nitrogen-oxides reduction unit without using a real NO ₂ sensor.

Preferably a first conditioning step can be performed by heating an oxidation stage in the diesel particulate filter unit to at least 35O ⁰ C, preferable to a temperature between 35O ⁰ C and 450 ⁰ C.

Favourably, a second conditioning step can be performed by heating an oxidation stage in the diesel particulate filter unit to at least 450 ⁰ C, preferable to a temperature between 450 ⁰ C and 550 ⁰ C.

Advantageously, a third conditioning step is performed by heating an oxidation stage in the diesel particulate filter unit to at least 550 ⁰ C, preferable to a temperature between 550°C and 65O ⁰ C.

Particularly, the second conditioning step can be performed after the first conditioning step and the third conditioning step can be performed after the second conditioning step. After each conditioning step it is decided if a further conditioning step has to be performed or not. The thermal load to the particulate filter unit is favourably reduced to the bare necessary heating steps and temperatures.

If all conditioning steps have been unsuccessful, after performing all conditioning steps unsuccessfully, an alarm can be set.

According to another aspect of the invention, an exhaust aftertreatment system of an engine is proposed comprising a diesel particulate filter unit which encompasses a particulate filter in which soot can be oxidized by NO ₂ and constituents of the exhaust gas are deoxidized in a nitrogen-oxides reduction unit arranged downstream of the diesel particulate filter unit, wherein the exhaust gas flows from the diesel particulate filter unit to the nitrogen-oxides reduction unit, wherein the operability of at least the diesel particulate filter unit can be determined by calculating the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring or calculating an NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring the NO _x content in the exhaust gas downstream of the nitrogen-oxides

reduction unit; calculating the NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; and deriving the actual and the expected NO _x conversion from the calculated NO ₂ and the measured and calculated NO _x contents.

Preferably an NO ₂ sensor can be provided upstream of the selective-catalytic- reduction catalyst and downstream of the diesel particulate filter unit. Additionally or alternatively, the NO ₂ sensor can be provided downstream of the nitrogen- oxides reduction unit. The NO ₂ sensor can be a real sensor implemented as hardware or a virtual sensor implemented as software where the NO ₂ content is calculated based on appropriate operation parameters of the engine and the exhaust gas aftertreatment system.

According to a further aspect of the invention, a computer program is proposed which is storable on a computer readable medium, comprising a program code for use in an on-board-diagnosis method for an exhaust aftertreatment system of an engine comprising a diesel particulate filter unit which encompasses a particulate filter in which soot can be oxidized by NO ₂ and constituents of the exhaust gas are deoxidized in a nitrogen-oxides reduction unit arranged downstream of the diesel particulate filter unit, wherein the exhaust gas flows from the diesel particulate filter unit to the selective-catalytic-reduction catalyst, comprising at least the steps of calculating the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring or calculating an NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; measuring the NO _x content in the exhaust gas downstream of the nitrogen-oxides reduction unit; calculating the NO _x content in the exhaust gas upstream of the nitrogen-oxides reduction unit and downstream of the diesel particulate filter unit; and deriving the actual and the expected NO _x conversion from the calculated NO ₂ and the measured and calculated NO _x contents.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention together with the above-mentioned and other objects and advantages may best be understood from the following detailed description of the

embodiments, but not restricted to the embodiments, wherein is shown schematically:

Fig. 1 a first embodiment of an exhaust aftertreatment system according to the invention;

Fig. 2a, 2b preferred locations where NO _x and NO ₂ levels can be determined; Fig. 3 a flowchart of a first preferred diagnosis method using a real NO ₂ sensor and virtual NO _x and NO ₂ sensors according to the invention; and Fig. 4 a flowchart of a second preferred diagnosis method using virtual NO _x and NO ₂ sensors according to the invention.

In the drawings, equal or similar elements are referred to by equal reference numerals. The drawings are merely schematic representations, not intended to portray specific parameters of the invention. Moreover, the drawings are intended to depict only typical embodiments of the invention and therefore should not be considered as limiting the scope of the invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS OF THE INVENTION

According to a first aspect of the invention a preferred exhaust gas after treatment system 12 depicted in Fig. 1 comprises a diesel particulate filter unit (DPFU) 60 arranged downstream of a diesel engine 12 and a NO _x reducing unit 70 such as preferably a selective-catalytic-reduction (SCR) arrangement arranged downstream of said DPFU 60, wherein an injector 62 is provided for feeding reducing agent such as ammonia or urea into the exhaust gas and arranged downstream of said DPF 64 and upstream said SCR catalyst. The DPFU 60 comprises an oxidation catalyst stage (DOCS) 20, e.g. an oxidation catalyst (DOC) 22 and a diesel particulate filter unit (DPFU) 60 which is arranged downstream of the DOC 22. Optionally, the DPF 64 can exhibit an oxidizing catalytic coating which can replace the DOC 22 as oxidation stage 20 or which can at least support the DOC 22.

Between the DPFU 60 and the SCR catalyst a sensing unit 40 is provided for sensing the amount of NO ₂ contained in the exhaust entering the SCR catalyst. The sensing unit 40 comprises a Nθ ₂ -sensitive sensor 50 arranged in the exhaust line 14 downstream of the DPFU 60 and upstream of the SCR catalyst and a control unit 42 connected to the sensor 50 via data line 48. Optionally a device 44 can be coupled to the control unit 42 to calculate the amount of NO ₂ entering the SCR catalyst depending on parameters 66, such as operating parameters of the engine 12 and/or on operating parameters of one or more catalysts 20, 64, 70 arranged in the exhaust aftertreatment system 10, providing a virtual sensor instead of a real NO ₂ sensor 50.

The DOCS 20, i.e. the DOC 22 and/or the catalytic coating of the DPF 64, is preferably used to generate a sufficient amount of NO ₂ for passive oxidation of soot trapped in the DPF 64 according to the reaction

(R1) NO+NO ₂ →NO ₂ .

The main function of the DPF 64 is to trap particulate matter such as soot and ashes contained in the exhaust gas. A typical vehicular exhaust aftertreatment system 10 requires one to several 100 000 km driving to fill the DPF 64 with ashes, and the DPF 64 can be emptied from ash by demounting the DPF 64 at service. To fill the DPF 64 with soot requires only one to several 1000 km driving. However, the soot can be oxidized to CO ₂ which can be done during operation of the vehicle.

For some applications it may be beneficial to coat the DPF 64 with a catalytically active material including the properties of an oxidation catalyst into the DPF 64. For proper function of the DPF 64 it is recommended to control the amount of soot trapped in the DPF 64. Regeneration of the DPF 64 may be accomplished in various ways known in the art. Preferably, NO ₂ can be used for passive oxidation of the trapped soot according to the reaction is

(R2) 2NO ₂ +C→2NO+CO ₂ .

For an efficient passive regeneration it is necessary to establish the exhaust gas temperature above a critical limit, preferably above 25O ⁰ C, and to provide an adequate amount of NO ₂ . The amount of NO ₂ in the exhaust gas fed into the DPF 64 can be increased by the DOCS 20 by oxidation of NO to NO ₂ .

Depending in the engine 12 emissions of soot and nitrogen oxides NO, NO ₂ , generally referred to as NO _x , the passive oxidation of soot can keep the soot level in the DPF 64 low at exhaust temperatures above 250 ⁰ C. For some engine emissions the ratio of NO _x /soot is too low for oxidizing the soot by NO ₂ . Alternative to passive oxidation of soot it can be oxidized by oxygen at high temperatures, preferably at about 600 ⁰ C. This can be achieved by either providing a burner (not shown) in the exhaust aftertreatment system 10 or by adding fuel to the exhaust gas which is burnt on an oxidation catalyst (not shown) upstream of the DPF 64. Activation of the burner or adding fuel is done in a regeneration phase which has a typical duration of a few minutes and which can last as long as 30 min if necessary.

Downstream of the DPF 64 and upstream of the nitrogen-oxides reduction unit 70, by way of example an SCR catalyst, the exhaust gas contains one or more constituents as NO and NO ₂ , which can be deoxidized in the SCR catalyst.

The main task of the SCR catalyst is to reduce NO _x , i.e. NO and NO ₂ , with a reductant to nitrogen gas N ₂ and water H ₂ O. On the SCR catalyst ammonia NH ₃ reacts with NO _x to form nitrogen. Usually on vehicles urea is injected into the exhaust gas and by the exhaust gas temperature urea is thermolyzed or hydrolyzed into NH ₃ in the exhaust gas and the SCR catalyst. The reductant, e.g. NH ₃ or urea, is added to the exhaust gas upstream of the SCR catalyst, for instance by the injector 62 (indicated by a broad arrow upstream of the SCR catalyst). The efficiency of the SCR catalyst is strongly dependent on the exhaust gas temperature, the space velocity of the exhaust gas and the NO ₂ /NO ratio in the exhaust gas which enters the SCR catalyst.

Depending on the kind of NO _x there are three principal chemical reactions possible:

(R3) 4NO + 4NH ₃ + O ₂ → 4N ₂ + 6H ₂ O (R4) NO + NO ₂ + 2NH ₃ →2N ₂ + 3H ₂ O (R5) 6NO ₂ + 8NH ₃ →7N ₂ + 12H ₂ O

The reaction (R4) has the highest efficiency and is efficient from exhaust temperatures below 200 ⁰ C and above. Reaction (a) becomes efficient at 300 ⁰ C and for reaction (c) the efficiency is lower than reaction (a) on vanadium based SCR-catalyst while it is on zeolite-based catalyst more efficient than reaction (a) but not as efficient as reaction (b). Further, on zeolite-based catalyst an unfavourable competitive reaction to reaction (c) exist which is generating the greenhouse gas N ₂ O:

(R6) 4NO2 + 4NH ₃ → 2N ₂ O + 2N ₂ + 6H ₂ O.

The NO ₂ formation in the DOCS 20 will depend on the exhaust gas mass flow and the temperature of the DOCS 20. Besides the flow and temperature dependency, the DOC 22 and/or the catalytic coating in the DPF 64 adsorbs sulphur (S), which can be contained in the exhaust gas, at lower temperatures and releases the sulphur at temperatures above 350 ⁰ C. If driving conditions let the DOCS 20 adsorb a lot of sulphur, the NO ₂ formation will be poisoned. The NO ₂ content after the DPF 64 will also depend on the condition of the DPF 64.

Sulphur is the main source to deactivate NO ₂ formation on the DOC 22 and on the catalytic coating of the DPF 64. Sulphur sticks to the catalyst at lower temperatures, typically below 400 ⁰ C and is released at higher temperatures (> 400 ⁰ C). The actual temperatures for sulphur adsorption and desorption depend on the particular catalyst formulation.

When low sulphur diesel fuel is used, which is now generally available in Europe and USA, it will take several hours or a day of engine operation without reaching 400 ⁰ C to give a noticeable decrease in NO ₂ formation in the DOC 20 and/or the

- i i - coated DPF 64. Such driving is unusual with heavy duty vehicles but can occur. However, sulphur poisoning of the DOC 22 and/or the coated DPF 64 can occur after shorter times if the driver gets fuel with higher sulphur contents, e.g. when driving in markets without low-sulphur fuel or fuelling high sulphur fuel by mistake. It's then important to detect such a poisoning and make a desulphation of the DOCS 22. Sulphur is removed from the DOC 22 and/or the coated DPF 64 by heating the catalysts to above 400 ⁰ C for more than 5 minutes, which can be done by injecting fuel into the exhaust or by activating a burner. Another source of sulphur is the lubricant oil.

Some conditions on some catalytic materials can cause a reversible degradation of the DOCS 20 in a manner that can it be reconditioned when heated to high temperatures above e.g. 500°C for a predetermined time period, e.g. several minutes.

The desulphatisation temperature does not degrade the SCR-catalyst and during desulphatisation the SCR-catalyst gets a temperature where it works very efficient and the influence of NO ₂ /NO ratio is low.

The description of the virtual sensor is a map or physical model of the NO ₂ formation in the DOC 22 and optionally in the DPF 64 if it's coated and on the NO ₂ consumption in the DPF 64. The sulphur dependency of the NO ₂ will not be included in the model since this invention is a way of handling the sulphur effect on NO ₂ (and it's hard to model also due to unknown variations of sulphur content in the fuel (low-sulphur fuel could be any thing below 10ppm in Europe for example).

According to the invention, an NO _x conversion is used for on-board-diagnosis of the correct function of the DOCS 20, i.e. the DOC 22 and/or the oxidizing catalytic coating of the DPF 64, if the DPF 64 is provided with such a coating. The NO _x conversion is derived from temperature, exhaust gas mass flow and NO ₂ levels in the exhaust gas. The NO ₂ sensor can be a real, physical sensor 50 or a virtual sensor wherein the NO ₂ level is calculated based on an appropriate model described below.

A virtual NO _x sensor is a rather complex model and consists preferably of following sub-models which are given in quotes:

"Engine-out NOy": The amount of NO _x at the outlet of the engine 12 can be estimated by a sensor or a model with following inputs for example: load or fuel amount, timing for fuel injection, engine speed, intake air pressure, intake air temperature, EGR (EGR = exhaust gas recycling) amount and intake air humidity. These are parameters of the engine 12 and sensed values. There are several ways to build the model. It can be map-based where all or at least some of the relevant parameters are, or can be, corrected by correction factors laid down in the map. It can also be a model built on a neural network as base.

"Exhaust gas flow": The exhaust gas flow can be measured, or derived from the measured air intake flow and the fuel amount, or from the calculated air intake flow from engine speed, intake air pressure, intake air temperature, EGR amount and volumetric efficiency of the engine.

"Exhaust gas flow in oxidation catalyst": The exhaust gas flow in the DOCS 20 can be measured or calculated.

"Temperature in catalyst": The temperature can e.g. be measured upstream of the DOCS 20. By applying an appropriate signal filter the measured value together with the exhaust gas flow into the DOCS 20 as a parameter can represent the actual catalyst temperature. Alternatively the temperature can be calculated by using a simple heat balance.

"Sulphur in oxidation catalyst": The sulphur content in the DOCS 20 is preferably calculated. For instance the calculation can be derived from the parameters in parentheses: (sulphur content in catalyst)=(sulphur content in catalyst a second before) + (sulphur adsorbed from exhaust during a second) - (sulphur desorbed during a second). The parameter "sulphur adsorbed from exhaust during a second" is the sulphur content in the fuel and lubrication oil consumed during the said second multiplied with a factor, wherein the factor is between 0 and 1 and has a temperature dependency which can e.g. be derived from a map containing

temperature dependent values of the factor. The parameter "sulphur desorbed during a second" is the sulphur content in the DOCS 20 one second before multiplied with another temperature dependent factor which can be derived in the same way as the first factor described above.

"NO? formation in catalyst": The NO ₂ formation in the DOCS 20 can be derived from interpolating in a 3-D based on the parameters exhaust gas flow, temperature in catalyst and sulphur content. It can also be calculated using a physical model with sulphur content, temperature, exhaust gas flow and oxygen concentration as input parameters. The model can be e.g. a specific NO ₂ formation rate which is k-i ¹ CN _O -C _C E and an NO ₂ decomposition rate which is k ₂ C _NO2 , where ki and k ₂ are temperature dependent and sulphur-content dependent parameters and C is the concentration of NO, NO ₂ and O ₂ , respectively. The specific rate is integrated over the catalyst volume. If there is a wide range of the HC content in the engine's working area or if an HC-injector is used, then the HC level is also an input parameter to the model, e.g. as a denominator for the specific rates (1 + K _a C _H c)-

"NO? out from the particulate filter": The amount NO ₂ which is released from the DPF 64 is the difference between the amount of NO ₂ fed into the DPF 64, NO ₂ formed in the DPF 64 (which is zero if no catalytic layer is provided in the DPF 64 for NO ₂ generation) and NO ₂ consumed by soot in the DPF 64. NO ₂ formed in the DPF 64 can be calculated in the same manner as the NO ₂ formed in the DOCS 20 (see above), preferably a physical model. NO ₂ consumed by soot in the DPF 64 is proportional to the amount of soot in the DPF 64 and can be expressed as a specific rate k3-C _N o2-C _S oot- Again, k ₃ is a temperature dependent parameter and C the respective concentration of NO ₂ and soot.

"Soot load in particulate filter": The soot load in DPF 64 can be derived from a measured pressure drop over the DPF 64 and/or by applying a model: (soot in the DPF 64 at a current time) = (soot in the DPF 64 at a time before the current time) + (soot emitted by the engine during the current time) - (soot burnt by NO ₂ during the current time). Soot burnt by NO ₂ during the current time is given by the "NO ₂ out from particulate filter" model, soot emitted by the engine during the current time is given from a soot sensor or a similar model as the "Engine-out NO _x " model.

The usage of a pressure drop for calculation of a soot amount in the DPF 64 can introduce some errors due to the fact that the soot characteristic is changing with time. Therefore it is preferred to use a model for calculating the soot load and use the pressure drop as a qualitative check of the model.

Figs. 2a and 2b depict preferred example locations where the NO ₂ and NO _x levels can be estimated either by measurement or calculation.

Fig. 2a corresponds to an arrangement preferably used at high loads of the engine 12 (Fig. 1), Fig. 2b corresponds to an arrangement preferably used al low loads of the engine 12 (Fig. 1).

From estimating NO ₂ and NO _x contents in the exhaust gas at different locations an actual measured and estimated conversion of NO ₂ in the DPFU 60 and conversion NO _x in the SCR catalyst can be derived. At high loads (Fig. 2a) it is preferred to measure the NO ₂ content in the exhaust gas upstream of the nitrogen-oxides reduction unit 70 and to calculate, i.e. estimate an expected NO ₂ content upstream of the nitrogen-oxides reduction unit 70. Additionally, the NO _x content upstream of the nitrogen-oxides reduction unit 70 can be measured or calculated. Downstream of the nitrogen-oxides reduction unit 70 the NO _x content is measured and calculated. A difference between the measured and the calculated contents indicates that a problem with the NO oxidation in the DPFU 60 has occurred.

At low loads (Fig. 2b) it is preferred to calculate the NO ₂ content in the exhaust gas yielding an estimated NO ₂ content upstream of the nitrogen-oxides reduction unit 70 and to calculate and/or to measure an expected NO _x content upstream of the nitrogen-oxides reduction unit 70. Downstream of the nitrogen-oxides reduction unit 70 the NO _x content is measured and calculated.

One or more temperatures sensors (not shown) are provided at convenient locations for determining the catalysts temperatures.

The NO _x -conversion is determined based on these values and on the temperature, exhaust gas massflow and the estimated NO ₂ content.

Fig. 3 illustrates a first embodiment of a preferred method of a preferred NO ₂ error handler for on-board diagnosis of the DPFU 60 according to the invention. An NO ₂ content measured with a real sensor and an estimated NO ₂ content calculated with the virtual sensor model are compared in step 102. By way of example, the nitrogen-oxides reduction unit 70 is a SCR catalyst. By comparing the NO _x - and NO ₂ contents upstream and downstream of the SCR catalyst, the efficiency of the SCR catalyst can also be determined

The comparison is based on the values of inputs 104 to 112 which comprise: input 104: temperature signals from one or more temperature sensors; exhaust gas mass flow; input 106: virtual NO ₂ signal downstream of DPF 64; input 108: NO ₂ signal measured by a real sensor 50 downstream of DPF 60 and upstream of SCR catalyst. If no reductant is injected into the SCR catalyst, the sensor 50 can be located downstream of the SCR catalyst; input 110: NO _x sensor signal downstream of SCR catalyst; input 112: Real or virtual NO _x signal upstream of SCR catalyst (can be upstream DPFU 60);

If the difference between the estimated virtual NO ₂ -sensor signal and the measured real NO ₂ -sensor signals is within predetermined limits, then the system is working properly and the NO ₂ error handler is finished and can jump back to step 102 for a next on-board diagnosis procedure. This can be done periodically with a predetermined time delay or can be done continuously.

If the difference is beyond predetermined limits, e.g. if the real NO ₂ signal is only a fraction of the virtual NO ₂ signal, for instance the real signal is e.g. 50% of the virtual signal or less, the NO ₂ error handler initiates a sulphur regeneration of the DPFU 60 in step 114, i.e. sulphur is removed from the DPFU 60. The DOCS 20 is regenerated in order to remove sulphur from the DOCS 20. If the DPF 64 is comprises alternatively or additionally an oxidizing catalytic coating, the DPF 64 is regenerated alternatively or additionally. The regeneration is done e.g. by control of the temperature and flow of exhaust gas from and/or fuel into the engine 12.

Preferably the conditioning is done at a temperature of at least 350 ⁰ C, preferably at a temperature between 350 ⁰ C and 450 ⁰ C.

After the regeneration the two signals are compared again in step 116. If the signals are within the predetermined limits, e.g. if they are nearly equal, the NO ₂ handler stores a fault code (step 118). The fault code indicates that a successful sulphur regeneration has been performed and that the DPFU 60, particularly the DOC 22, had been deactivated by sulphur oxidation. Then the NO ₂ error handler is finished and can jump back to step 102 for a next on-board diagnosis procedure.

If the two signals still deviate after the sulphur regeneration an oxidation-catalyst- activity conditioning is initiated in step 120. The regeneration can be done as before, e.g. by control of the temperature and flow of exhaust gas from and/or fuel into the engine 12. Preferably the conditioning is done at a temperature of at least 450°C, preferably at a temperature between 450 ⁰ C and 550°C.

Subsequently the two signals are compared in step 122 in the same manner as after the regeneration of the DPFU 60. If the signals are within the predetermined limits, e.g. if they are nearly equal, the NO ₂ handler stores a fault code (step 124). The fault code indicates that a successful sulphur regeneration has been performed and that the DPFU 60, particularly the DOC 22, had been deactivated by sulphur oxidation. Then the NO ₂ error handler is finished and can jump back to step 102 for a next on-board diagnosis procedure.

If the two signals still deviate, a soot regeneration of the DPF 64 is performed in step 126, e.g. a burner is activated and/or fuel is injected to oxidize the soot a high temperatures. The two signals are again compared in step 128. If the two signals are now within predetermined limits, the NO ₂ error handler sends a signal to a soot error handler indicating that there could have been more soot in the DPFU 60 than expected (step 130). Then the NO ₂ error handler is finished and can jump back to step 102 for a next comparison.

If the virtual and the real NO ₂ signals still deviate a fault code of a malfunctional DPFU 60 is set (step 132) and if necessary, e.g. if the NO ₂ is so low that the low

level influences the NO _x emission of the aftertreatment system 10, an alarm is set. An alarm is favourably activating a MIL (M I L=malf unction indicator light).

In a second, not depicted embodiment of the invention the first comparison between the two NO ₂ signals is also compared (if possible) with the efficiency of the SCR catalyst. This can be reasonably done when the SCR catalyst is in a state where there is a measurable difference in efficiency as a function of NO ₂ /NO ratio. If the efficiency is as expected from the virtual NO ₂ signal a fault code is set on the real NO ₂ sensor. The real efficiency can for example be expressed as a conversion 1-[lnput 110]/[lnput 112]. The expected conversion (i.e. efficiencies) can then be found in a map with temperature, mass flow, NO _x signal upstream the SCR catalyst and NO ₂ signals upstream the SCR catalyst.

Referring now to Fig. 4 a third embodiment of the preferred NO ₂ error handler is presented. This flow chart corresponds to an arrangement depicted in Fig. 2b related to low engine loads. By way of example, the nitrogen-oxides reduction unit 70 is a SCR catalyst. Deviating from the first embodiment in Fig. 3 there is no real NO ₂ sensor 50 and the estimated NO ₂ content is calculated instead with the virtual sensor model. A measured and an estimated NO _x conversion of the SCR catalyst is compared in step 202 based on the estimated NO ₂ content and the estimated temperature signal on the SCR catalyst, preferably based on temperature, exhaust gas mass flow and NO ₂ level downstream of the DPFU 60 and upstream of the SCR catalyst. By comparing the NO _x - and NO ₂ contents upstream and downstream of the SCR catalyst, the efficiency of the SCR catalyst can be determined.

The comparison is based on the values of inputs 204 to 212 which comprise: input 204: temperature signal of one or more temperature sensors, exhaust gas mass flow; input 206: virtual NO ₂ signal downstream of DPF 64 and upstream of the SCR catalyst; input 208: NO _x sensor signal downstream of the SCR catalyst; expected based on the NO ₂ -sensor signal of the virtual sensor;

input 210: real or virtual NO _x sensor signal upstream of the SCR catalyst (can be upstream DPFU 60); input 212: virtual NO _x signal downstream of SCR catalyst (preferably based on inputs 204, 206, and 210).

If the deviation of the real estimated efficiency of the SCR catalyst (based on the virtual sensor signal) from the expected efficiency is above a predetermined threshold the same procedure is done as in Fig. 3. If the difference between the two efficiency values of the example SCR catalyst is within predetermined limits, then the NO ₂ -handler can jump back to step 202 and is ready for the next onboard diagnosis procedure. This can be done periodically with a time delay of continuously.

If the difference is beyond predetermined limits, then the NO2 error handler initiates a sulphur regeneration of the DPFU 60 in step 214. A condition beyond predetermined limits is e.g. if the real efficiency value (based on measured signals) is only a fraction of the estimated value (based on calculated signals). The DOCS 20 is regenerated in order to remove sulphur from the oxidizing catalytic material of the DOC 20. If the DPF 64 comprises, alternatively or additionally, an oxidizing catalytic coating, the DPF 64 is regenerated alternatively or additionally. The regeneration is done e.g. by control of the temperature and flow of exhaust gas from and/or fuel into the engine 12. Preferably the conditioning is done at a temperature of at least 350 ⁰ C ₁ preferably at a temperature between 350 ⁰ C and 450 ⁰ C.

After the regeneration the two signals are compared again in step 216. If the signals are within the predetermined limits, e.g. if they are nearly equal, the NO ₂ handler stores a fault code (step 218). The fault code indicates that a successful sulphur regeneration has been performed and that the DPFU 60, particularly the DPF 64, had been deactivated by sulphur oxidation. Then the NO ₂ error handler is finished and can jump back to step 202 for the next on-board diagnosis procedure.

If the two signals still deviate after the sulphur regeneration an oxidation-catalyst- activity conditioning is initiated in step 220. The regeneration can be done as

before, e.g. by control of the temperature and flow of exhaust gas from and/or fuel into the engine 12. Preferably the conditioning is done at a temperature of at least 45O ⁰ C ₁ preferably at a temperature between 45O ⁰ C and 550 ⁰ C.

Subsequently the two signals are compared again in step 222 in the same manner as after the regeneration of the DPFU 60. If the signals are within the predetermined limits, e.g. if they are nearly equal, the NO ₂ handler stores a fault code (step 224). The fault code indicates that a successful sulphur regeneration has been performed and that the DPFU 60, particularly the DOCS 20, had been deactivated by sulphur oxidation. Then the NO2 error handler is finished and can jump back to step 202 for the next on-board diagnosis procedure.

If the two signals still deviate, a soot regeneration of the DPF 64 is performed in step 226. The two signals are again compared in step 228. If the two signals are now within predetermined limits, the NO ₂ error handler sends a signal to a soot error handler indicating that there could have been more soot in the DPFU 60 than expected (step 230). Then the NO ₂ error handler is finished and can jump back to step 202 for the next on-board diagnosis procedure.

If the estimated and the measured NO ₂ signals still deviate a fault code of a malfunctional DPFU 60 is set (step 232) and if necessary, e.g. if the NO ₂ is so low that the low level influences the NO _x emission of the aftertreatment system 10, a separate diagnosis for the SCR catalyst and a reducing agent system coupled to the SCR catalyst is initiated.

In a forth embodiment (not shown) some or all SCR catalyst diagnosis and a reducing agent system coupled to the SCR catalyst is performed before starting sulphur regeneration.