Method and diagnostic unit for diagnosing a differential pressure sensor

Abstract

Method for diagnosing a differential pressure sensor (15) which determines a differential pressure (16) across an exhaust aftertreatment unit of an exhaust gas purification system of an internal combustion engine (10) arranged in an exhaust stream (11) and which is connected to the exhaust stream (11) upstream of the exhaust aftertreatment unit via a front connection (15.1) and a rear connection (15.2) is connected to the exhaust stream (11) downstream of the exhaust aftertreatment unit, wherein the diagnosis is based on a comparison of a measured time differential pressure gradient (33) of the measured differential pressure (16) as a result of a change in the exhaust pressure upstream of the exhaust aftertreatment unit with an expected time differential pressure gradient (32) of an expected differential pressure, characterized in that a correlation is determined between a course of the measured differential pressure gradient (33) with a course of the expected differential pressure gradient (32) and that a high correlation indicates an intact differential pressure sensor (15) with intact connections (15.1, 15.2) and a low correlation indicates a defective differential pressure sensor (15) or one or two defective connections (15.1, 15.2) of the differential pressure sensor (15). wherein a cross-correlation value is formed by forming a normalized cross-correlation from the course of the measured differential pressure gradient (33) and the course of the expected differential pressure gradient (32) and that the correlation between the measured and the expected course of the differential pressure gradient (33, 32) is evaluated on the basis of the cross-correlation value, wherein a damaged or interrupted rear connection (15.2) of the differential pressure sensor (15) is inferred if the cross-correlation value is above a predetermined first correlation threshold, preferably above a first correlation threshold greater than 1, particularly preferably above a first correlation threshold greater than 1.2, and / or wherein a damaged or interrupted front connection (15.1) of the differential pressure sensor (15) is inferred if the cross-correlation value is below a predetermined second correlation threshold, preferably below a second correlation threshold less than 0, particularly preferably below a second correlation threshold less than -0.2.

Description

State of the art

[0001] The invention relates to a method for diagnosing a differential pressure sensor which determines a differential pressure across an exhaust gas aftertreatment unit of an exhaust gas purification system of an internal combustion engine arranged in an exhaust gas line and which is connected with a front connection to the exhaust gas line upstream of the exhaust gas aftertreatment unit and with a rear connection to the exhaust gas line downstream of the exhaust gas aftertreatment unit, wherein the diagnosis is based on a comparison of a measured temporal differential pressure gradient of the measured differential pressure as a result of a change in the exhaust gas pressure upstream of the exhaust gas aftertreatment unit with an expected temporal differential pressure gradient of an expected differential pressure.

[0002] The invention further relates to a diagnostic unit for diagnosing a differential pressure sensor for determining a differential pressure across an exhaust gas aftertreatment unit of an exhaust gas purification system in an exhaust system of an internal combustion engine, wherein the measurement signals of the differential pressure sensor are supplied to the diagnostic unit, wherein the diagnostic unit is designed to form a measured differential pressure gradient from the measured differential pressure and to determine an expected differential pressure gradient from at least one exhaust gas volume flow through the exhaust gas aftertreatment unit and a flow resistance of the exhaust gas aftertreatment unit.

[0003] Emissions legislation, particularly in the USA and Europe, sets limits for pollutant emissions from internal combustion engines. To comply with these requirements, exhaust gas purification systems containing exhaust aftertreatment units, such as catalytic converters or particulate filters, are provided in the exhaust systems of internal combustion engines. Their correct functioning must be monitored. In addition to the emission limits, diagnostic limits are also specified; if these are exceeded, an error must be indicated. For this purpose, diagnostic functions are implemented in the vehicle. These monitor the parts and components installed to reduce emissions during vehicle operation as part of this on-board diagnostics (OBD) and indicate any malfunction that leads to the diagnostic limits being exceeded.

[0004] To monitor the function of an exhaust gas aftertreatment unit, in particular a particulate filter, it is known to monitor the pressure drop across the exhaust gas aftertreatment unit. In the case of a particulate filter, its flow resistance increases with increasing soot deposits. This can be monitored by measuring the differential pressure across the particulate filter, and regeneration of the particulate filter can be initiated when a capacity limit is reached. The particulate filter can be located in the exhaust system of a gasoline- or diesel-powered internal combustion engine and can be a pure, uncoated particulate filter or a catalytically coated particulate filter as a combination of a particulate filter with a three-way catalyst.

[0005] In addition to monitoring the exhaust aftertreatment unit itself, monitoring of the sensors required for it must also be provided. A differential pressure sensor is usually connected to the exhaust system via two connections, one before and one after the exhaust aftertreatment unit to be monitored. The connections can be designed as simple hose lines. If one of the connections falls off or is damaged, the diagnostic result of the exhaust aftertreatment unit will be corrupted. Reliably detecting a drop in one of the differential pressure sensor connections can sometimes be very difficult because the measurement signal from the differential pressure sensor can still be within a valid value range after a drop in one connection.

[0006] US 2013 / 0 327 018 A1 discloses a method for monitoring a differential pressure sensor (32) of an exhaust aftertreatment system (20) of a vehicle used to treat an exhaust flow from an engine (24), the method comprising: enabling a monitoring algorithm of a control module (38) only under the condition that the vehicle is in a post-operational state, no inhibiting conditions are detected, and at least one previous (n-1) learned value of the differential pressure sensor (32) is less than a threshold pressure value (92); wherein the monitoring algorithm comprises: detecting a learned value (n-1, n, n+1, n+2, etc.) of the differential pressure sensor (32) for each post-operation event of the vehicle within a predefined time period after the engine has been turned off; calculating an absolute value of a rate of change between a current (n) learned value of the differential pressure sensor (32) and a previous (n - 1) learned value of the differential pressure sensor (32); comparing the absolute value of the rate of change of the learned values ​​of the differential pressure sensor (32) with the threshold pressure value (92) to determine whether the absolute value of the rate of change of the learned values ​​is greater than the threshold pressure value (92) or less than the threshold pressure value (92); and signaling a fault in the differential pressure sensor (32) if the absolute value of the rate of change of the learned values ​​is greater than the threshold pressure value (92).

[0007] DE 10 2005 034 270 A1 discloses a method and a device for monitoring a differential pressure sensor arranged parallel to an exhaust component, in particular a particulate filter. The dynamic behavior of the differential pressure signal is evaluated as a result of a predetermined change in the exhaust pressure upstream of the exhaust component. According to one embodiment, a gradient of the differential pressure signal is determined upon a change in the exhaust pressure and compared with a second threshold value. The second threshold value is formed as a function of operating parameters of the internal combustion engine and represents at least approximately an expected gradient signal. If the gradient of the pressure signal does not exceed the second threshold value, a fault in the front supply line (supply line blocked or broken) to the differential pressure sensor is assumed.In a further design variant, an overshoot of the differential pressure signal is evaluated over time.

[0008] The signal curve is thus evaluated for a single change in exhaust pressure. To increase diagnostic reliability, several such individual evaluations can be performed, and only if a repeated error message occurs can a faulty front connection be determined.

[0009] Document DE 10 2011 003 748 A1 describes a method and device for monitoring the function of a differential pressure sensor arranged parallel to a particulate filter by means of connections. An exhaust flap is provided downstream of the particulate filter in the flow direction. To monitor the connections of the differential pressure sensor, the position of the exhaust flap is changed in such a way that the exhaust back pressure upstream of the exhaust flap, and thus the pressure upstream and downstream of the particulate filter, increases. If the differential pressure sensor is correctly connected, the measured pressure difference remains constant within specified tolerances. If one of the connections is loose, the pressure change is not transmitted to the differential pressure sensor. This changes the measured pressure difference, and the error is detected.If the pressure difference exceeds a specified expected value, a defect in the rear connection can be assumed, as the pressure increase at the particulate filter outlet is not transmitted to the differential pressure sensor. If, however, the front connection is defective, the pressure change caused by the exhaust flap leads to a negative change in the measured pressure difference.

[0010] According to a variant of the invention, the temporal progression of the measured differential pressure after a change in the position of the exhaust flap is also evaluated. If the changed pressure is delayed on one side of the differential pressure sensor, for example, due to a blocked or jammed connection, this error can also be detected.

[0011] The method requires that an exhaust flap is provided in the exhaust duct after the particulate filter, which is not the case in most exhaust systems, so the method cannot be applied here.

[0012] DE 10 2007 000 892 B4 discloses a diagnostic device that can detect a blocked connection of a pressure differential measuring device. The device measures a pressure differential across an exhaust gas purification device. The pressure differential is determined at a steady-state operating point. A second steady-state operating point is then set, which is expected to produce a large change in the pressure differential compared to the first steady-state operating point. If the connecting line is blocked, the measured differential pressure converges only slowly toward a stable final value. If the time until such a final value is reached is too long, it is concluded that a connection of the device is inhibited (blocked).

[0013] The procedure does not allow the detection of a dropped connection.

[0014] It is therefore an object of the invention to provide a method with which a differential pressure sensor can be reliably monitored for various possible causes of errors.

[0015] It is a further object of the invention to provide a diagnostic unit for carrying out the method. Disclosure of the invention

[0016] The object of the invention relating to the method is achieved in that a correlation between a course of the measured differential pressure gradient and a course of the expected differential pressure gradient is determined and that in the case of a high correlation, an intact differential pressure sensor with intact connections is concluded and in the case of a low correlation, a defective differential pressure sensor or one or two defective connections of the differential pressure sensor is concluded.

[0017] It may be provided that the correlation is determined and evaluated over at least two or more than two changes in the exhaust gas pressure.

[0018] By evaluating the differential pressure gradients, the method is independent of the measured absolute differential pressures, which can still be within a valid value range even if, for example, the differential pressure sensor connection is disconnected. By correlating the measured differential pressure gradient with the expected differential pressure gradient, the two gradients can be compared over a longer period of time, significantly reducing the risk of misdiagnosis compared to the evaluation of a single differential pressure change. This is especially true if multiple changes in the exhaust pressure, and thus in the differential pressure, occur within the evaluation period in which the correlation is determined.

[0019] A simple and reliable determination of a correlation between the measured and expected differential pressure gradients can be achieved by forming a cross-correlation value from the measured differential pressure gradient and the expected differential pressure gradient using a normalized cross-correlation. The correlation between the measured and expected differential pressure gradients is then evaluated using the cross-correlation value. For this purpose, the cross-correlation value (KKF) is calculated, for example, according to the following relationship: KKF=∑(d(Δp(k))*d(Δp*(k))) / ∑(d(Δp*(k))*d(Δp*(k))) where d(Δp (k) ) / dk the measured differential pressure gradient and d(Δp* (k)) / dk represents the expected differential pressure gradient across an intact exhaust aftertreatment unit. The resulting normalized cross-correlation value is independent of the signal levels of the gradients and takes high or low values ​​for insufficient correlation and values ​​around 1 for good correlation.

[0020] According to a particularly preferred embodiment of the invention, it can be provided that a damaged or interrupted rear connection of the differential pressure sensor is inferred if the cross-correlation value is above a predetermined first correlation threshold, preferably above a first correlation threshold greater than 1, particularly preferably above a first correlation threshold greater than 1.2, and / or that a damaged or interrupted front connection of the differential pressure sensor is inferred if the cross-correlation value is below a predetermined second correlation threshold, preferably below a second correlation threshold less than 0, particularly preferably below a second correlation threshold less than -0.2. The diagnosis thus enables the detection and precise localization of the error.If the rear connection of the differential pressure sensor breaks or becomes damaged, the temporal gradients of the differential pressure signal increase significantly, as the differential pressure sensor now not only detects the differential pressure across the exhaust aftertreatment unit, but also measures it against the ambient pressure. The differential pressure sensor thus also detects the pressure drop across exhaust components that are located downstream of the exhaust aftertreatment unit in the exhaust system, for example, in the form of a silencer, in the direction of exhaust flow. The normalized cross-correlation then results in a value significantly greater than 1.0, which is above the range of values ​​for the cross-correlation value that occurs with an intact and correctly connected differential pressure sensor and an intact exhaust aftertreatment unit.If the front connection of the differential pressure sensor falls off or becomes damaged, the differential pressure sensor will measure negative differential pressures and negative differential pressure gradients over a wide range. The normalized cross-correlation then shows a value significantly less than 0. Thus, the obtained cross-correlation value lies outside the value ranges that would occur with an intact differential pressure sensor over an intact exhaust aftertreatment unit or with a removed exhaust aftertreatment unit.

[0021] The expected differential pressure gradient can be determined by determining the profile of the expected differential pressure from at least an exhaust gas volume flow through the exhaust aftertreatment unit and a flow resistance of the exhaust aftertreatment unit. The expected differential pressure can be simply determined as the product of the existing exhaust gas volume flow and the flow resistance. The temporal change in the exhaust gas volume flow, assuming that the flow resistance remains at least approximately constant over the duration of the diagnosis, results in a profile of the expected differential pressure and thus of the expected differential pressure gradient. The assumed flow resistance can, for example, be designed for a borderline exhaust aftertreatment unit with a maximum possible flow resistance. This can reliably prevent misdiagnoses.

[0022] In a further preferred method variant, the measured differential pressure across the exhaust aftertreatment unit and / or the expected differential pressure across the exhaust aftertreatment unit and / or the exhaust gas volume flow can be low-pass filtered to determine the expected differential pressure. This allows interference-related signal fluctuations to be suppressed for diagnosis, which increases diagnostic quality.

[0023] A correlation between the changes in the measured and expected differential pressures can only be evaluated in sufficiently dynamic operating situations of the internal combustion engine. Therefore, it can be specified that the diagnosis is only performed if the differential pressure gradient, preferably the expected differential pressure gradient, or a measure correlating with the differential pressure gradient exceeds a specified value. Measures correlating with the differential pressure gradient can include, for example, the gradients of the exhaust gas mass flow, the exhaust gas volume flow, the engine speed, or variables derived from them.

[0024] If the differential pressure sensor is intended to determine the differential pressure across a particulate filter, the function of the differential pressure sensor and thus that of the particulate filter can be reliably monitored. Monitoring particulate filters using a differential pressure sensor is the standard procedure today.

[0025] The method therefore has a wide range of applications, whereby the diagnosis of the differential pressure sensor can be carried out cost-effectively without additional components, based solely on the data already available.

[0026] The method can preferably be used in a gasoline-powered internal combustion engine in which the exhaust system has at least a separate catalytic converter and a particulate filter, or a catalytic converter-particulate filter combination, or a catalytically coated particulate filter. In gasoline engines, the flow rate is relatively low compared to diesel engines, so only minor pressure differences can occur across a gasoline particulate filter used in such engines. The high detection accuracy of the method enables reliable detection of a defect in the connections of the differential pressure sensor even under these unfavorable conditions.

[0027] The object of the invention relating to the diagnostic unit is achieved in that the diagnostic unit has an electrical circuit arrangement or a program sequence for carrying out the method according to one of claims 1 to 7. In particular, the implementation by a program sequence can be carried out cost-effectively.

[0028] If the diagnostic unit is intended to be a separate component or an integral part of a higher-level engine control system, the differential pressure sensor diagnosis can be performed in a specially designed unit or without additional components, thus resulting in a cost-effective solution. The engine control system typically already has all the data necessary to perform the described procedure.

[0029] The invention is explained in more detail below using an exemplary embodiment illustrated in the figures. It shows: Fig. 1 shows a schematic representation of a variant of the technical environment in which the invention can be used, Fig. 2 in a first diagram, a measured differential pressure curve and an expected differential pressure curve across a particle filter with an intact differential pressure sensor, Fig. 3 in a second graph, the measured differential pressure curve and the expected differential pressure curve across a particle filter with an open rear connection of the differential pressure sensor, Fig. 4 in a third diagram, the measured differential pressure curve and the expected differential pressure curve across a particle filter with an open front connection of the differential pressure sensor, Fig. 5 in a fourth diagram, the measured differential pressure curve and the expected differential pressure curve for a removed or defective particulate filter, Fig. 6 in a fifth diagram, the course of a measured differential pressure gradient and the course of an expected differential pressure gradient with an open rear connection of the differential pressure sensor, Fig. 7 in a sixth diagram, the curve of the measured differential pressure gradient and the curve of the expected differential pressure gradient with an open front connection of the differential pressure sensor and Fig. 8 in a seventh trend diagram the trends of cross-correlation values ​​determined from the differential pressure gradients for different errors of the differential pressure sensor.

[0030] Fig. 1 shows a schematic representation of an embodiment variant of the technical environment in which the invention can be used. The representation is limited to the components necessary for the description of the invention. Shown as an example is an internal combustion engine 10 designed as a gasoline engine, wherein the exhaust gas of the internal combustion engine 10 is discharged via an exhaust system 11. A multi-stage exhaust gas purification system is provided in the exhaust system 11. For this purpose, a catalyst 12 designed as a three-way catalyst and a particulate filter 13 are arranged as exhaust gas aftertreatment units along the exhaust system 11 in the flow direction of an exhaust gas stream 14. A silencer 17 is provided downstream of the particulate filter 13.

[0031] For diagnosing the particulate filter 13, a differential pressure sensor 15 is connected, which can be used to determine the pressure difference (differential pressure 16) between a filter inlet and a filter outlet of the particulate filter 13. For this purpose, the differential pressure sensor 15 is connected to the exhaust system 11 upstream of the particulate filter 13 via a front connection 15.1 and to the exhaust system 11 downstream of the particulate filter 13 via a rear connection 15.2. An output signal from the differential pressure sensor 15 is fed to a diagnostic unit 18, in which, as part of an on-board diagnostic (OBD) system, a diagnosis of the differential pressure sensor 15 can be performed for an open or damaged front or rear connection 15.1, 15.2. This diagnostic unit 18 can be part of a higher-level engine control unit (ECU).

[0032] In the illustrated embodiment, connections 15.1, 15.2 are designed as hose lines. If one of these connections 15.1, 15.2 falls off or is damaged, for example, by a marten bite, the differential pressure 16 is determined incorrectly, resulting in an incorrect diagnosis of the particulate filter 13. The measured differential pressures 16 remain within a normal range, particularly if the rear connection 15.2 is damaged or has fallen off, so it is hardly possible to distinguish them from normal operation based on the measured absolute differential pressures 16.

[0033] Fig. 2 shows in a first diagram a measured differential pressure curve 23 and an expected differential pressure curve 22 over the Fig. 1 with an intact and correctly connected differential pressure sensor 15. For this purpose, the differential pressure curves 22, 23, as measured or expected for an intact particulate filter 13, are plotted against a time axis 20 and a differential pressure axis 21. The expected differential pressure curve 22 is formed as the product of a respective existing exhaust gas volume flow and a flow resistance of a reference particulate filter. What is characteristic here is that only slight signal height differences and phase differences occur between the expected differential pressure curve 22 and the measured differential pressure curve 23, thus indicating a high correlation between the differential pressure curves 22, 23. This high correlation signals an intact and correctly connected differential pressure sensor.

[0034] Fig. 3 shows, in a second curve diagram, the measured differential pressure curve 23 and the expected differential pressure curve 22 across the particulate filter 13 with an open rear connection 15.2 of the differential pressure sensor 15. The rear connection 15.2 has faultily detached itself from the exhaust system 11 or the differential pressure sensor 15. In the event of this fault, the measured differential pressure 16 and the temporal gradients of the differential pressure signal increase significantly, since the differential pressure sensor 15 not only detects the differential pressure 16 across the particulate filter 13, but also measures it against the ambient pressure and thus, in comparison to a correctly connected differential pressure sensor 15, also detects the pressure drop across the silencer 17 or other exhaust gas components (not shown) that are arranged downstream of the particulate filter 13.

[0035] Fig. Figure 4 shows, in a third graph, the measured differential pressure curve 23 and the expected differential pressure curve 22 across the particulate filter 13 with an open front connection of the differential pressure sensor 15. The front connection 15.1 has faultily separated from the exhaust system 11 or the differential pressure sensor 15. With this fault, the differential pressure sensor 15 measures negative differential pressures 16 with negative differential pressure gradients over a wide range. In addition to the differing signal levels, this results in a phase shift between the expected and measured differential pressure curves 22, 23.

[0036] Fig. 5 shows, in a fourth diagram, the measured differential pressure curve 23 and the expected differential pressure curve 22 for a removed or defective particulate filter 15. In this case, the flow resistance between the connections 15.1, 15.2 becomes very small, so that the measured differential pressure curve 23 and its gradient are significantly smaller than those of the expected differential pressure curve 22.

[0037] Fig. 6 shows, in a fifth curve diagram, the curve of a measured differential pressure gradient 33 and the curve of an expected differential pressure gradient 32 with an open rear connection 15.2 of the differential pressure sensor 15. The curves are plotted over a longer period of time with several changes in the differential pressure 16 applied across the particle filter 13 against a second time axis 30 and a differential pressure gradient axis 31.

[0038] Due to the drop of the rear connection 15.2 from the exhaust line 11 or the differential pressure sensor 15, this measures, as already Fig. 3, against ambient pressure. The measured differential pressure gradients 33 are thus significantly higher than the expected differential pressure gradient 32 if the changes in the differential pressure 16 are sufficiently large. Accordingly, there is little or no correlation between the curves of the measured differential pressure gradient 33 and the expected differential pressure gradient 32. According to the invention, the error is detected when there is a lack of correlation between the curves of the measured and expected differential pressure gradients 33, 32.

[0039] Fig. 7 shows, in a sixth diagram, the curve of the measured differential pressure gradient 33 and the curve of the expected differential pressure gradient 32 with an open front connection 15.1 of the differential pressure sensor 15. The curves are again plotted against the second time axis 30 and the differential pressure gradient axis 31.

[0040] By the fall or damage of the front connector 15.1, as this Fig. 4, predominantly negative differential pressures 16 and differential pressure gradients 33 are measured, resulting in a phase shift between the curves of the differential pressure gradients 32, 33. Accordingly, even with this error, there is no or only a slight correlation between the curves of the measured differential pressure gradient 33 and the expected differential pressure gradient 32, and the error can be detected according to the invention due to the lack of correlation.

[0041] The diagnostic method according to the invention is therefore based on monitoring the differential pressure sensor 15 and its connections 15.1, 15.2 by evaluating the correlation of the temporal gradient of the measured differential pressure 16 across the particulate filter 13 with an expected temporal gradient of the differential pressure 16 across an intact particulate 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.

[0042] The measured differential pressure signal is first low-pass filtered to suppress noise. Subsequently, the curve of the measured differential pressure gradient 33 d(Δp( k) / dk is determined as the temporal gradient of the signal from pressure sensor 15, where k represents the k-th measurement. For the same period, a corresponding curve of the expected differential pressure gradient 32 d(Δp* (k) ) / dk is determined by calculating the temporal course of the expected differential pressure gradient 32 across an intact particle filter 13 from an exhaust gas volume flow or its temporal gradient and a flow resistance of an intact particle filter 13, the reference particle filter. The expected differential pressure gradient 32 or the exhaust gas volume flow entering there can optionally also be low-pass filtered.

[0043] Subsequently, a standardized cross-correlation of the measured differential pressure gradient 33 and the expected differential pressure gradient 32 is used to determine the extent to which they are similar. For this purpose, a cross-correlation value (KKF) is calculated according to the following relationship: KKF=∑(d(Δp(k))*d(Δp*(k))) / ∑(d(Δp*(k))*d(Δp*(k))) where d(Δp( k ) / dk the in the Fig. 6 and Fig. 7 shown measured differential pressure gradient 33 and d(Δp* (k) ) / dk represent the expected differential pressure gradient 32 at different times, denoted by the index k. The expected differential pressure ΔP* (k) is calculated from the product of exhaust gas volume flow and the flow resistance R* of the intact reference particulate filter.

[0044] To assess whether the differential pressure sensor 15 is properly connected or functioning properly, the output value of the standardized cross-correlation, the cross-correlation value KKF, is compared with threshold values ​​stored in the diagnostic unit 18 or a higher-level engine control unit.

[0045] Fig.Figure 8 shows, in a seventh graph, the curves of cross-correlation values ​​determined from the differential pressure gradients 32, 33 for various errors of the differential pressure sensor 15. For this purpose, a first cross-correlation value curve 42, a second cross-correlation value curve 43, a third cross-correlation value curve 44, and a fourth cross-correlation value curve 45 are plotted against a third time axis 40 and a correlation value axis 41. Furthermore, a first cross-correlation value range 46 and a second cross-correlation value range 47 are marked by double arrows opposite the correlation value axis 41.

[0046] The second cross-correlation value curve 43 is obtained for an intact particulate filter 13 and a correctly connected and functioning differential pressure sensor 15. Under these conditions, the second cross-correlation value curve 43 can lie within the first cross-correlation value range 46.

[0047] The third cross-correlation value curve 44 is measured with a removed particulate filter 13 and a correctly connected and functioning differential pressure sensor 15. The second cross-correlation value curve 43 can lie within the second cross-correlation value range 47.

[0048] The first cross-correlation value curve 42 is obtained when the rear connection 15.2 of the differential pressure sensor 15 is loose or damaged. The obtained correlation values ​​are then significantly above the first cross-correlation value range 46, which allows the error to be reliably detected and localized to the rear connection 15.2.

[0049] The fourth cross-correlation value curve 45 is obtained when the front connection 15.1 of the differential pressure sensor 15 is loose or damaged. The obtained correlation values ​​are significantly below the first cross-correlation value range 46 and also below the second cross-correlation value range 47. A fault at the front connection 15.1 of the differential pressure sensor 15 can thus be reliably detected and distinguished from a defective rear connection 15.2 or a removed particulate filter 13.

[0050] For a high correlation of the curves of the expected and measured differential pressure gradients 32, 33, as is the case with the intact differential pressure sensor 15 and particle filter 13, a cross-correlation value in the range of 1 results. To detect a defect in the rear connection 15.2, it is therefore checked whether a cross-correlation value significantly greater than 1, in particular greater than 1.2, is obtained.

[0051] A removed particulate filter 13 leads to low cross-correlation values, particularly cross-correlation values ​​between 0.5 and -0.2. To detect a defect in the front connector 15.1, it is therefore checked whether a cross-correlation value significantly less than 0, particularly less than -0.2, is obtained.

[0052] The method works particularly reliably when a certain degree of dynamic excitation is present, i.e., when the differential pressure gradients 32, 33 exceed a certain level. Therefore, cross-correlation evaluation is only advantageous when certain dynamic criteria are met. Possible criteria for this are the gradients of the exhaust gas mass flow, the exhaust gas volume flow, the speed of the internal combustion engine 10, or variables derived from them. Ideally, the expected differential pressure gradient is used directly to evaluate sufficient dynamics.

[0053] 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 relative to the atmospheric pressure, or two absolute pressure sensors upstream and downstream of the particulate filter 13. The diagnostic procedure also allows the front and rear connections 15.1, 15.2 to be checked, which are then each connected to separate pressure sensors.

[0054] The advantage of this method is that the correlation can be determined over a longer period of time across multiple changes in the exhaust gas volume flow and thus in the differential pressure 16. This significantly increases diagnostic reliability compared to an evaluation of the measured differential pressure gradient 33 for a single pressure change and also compared to repeated individual measurements.

[0055] In an advantageous embodiment, the diagnostic method is stored as software in the diagnostic unit 18 and can be used in particular in gasoline engines with gasoline particle filters, but also in diesel engines.

Claims

[1] Method for diagnosing a differential pressure sensor (15) which determines a differential pressure (16) across an exhaust gas aftertreatment unit of an exhaust gas purification system of an internal combustion engine (10) arranged in an exhaust system (11) and which is connected to the exhaust system (11) upstream of the exhaust gas aftertreatment unit by a front connection (15.1) and to the exhaust system (11) downstream of the exhaust gas aftertreatment unit by a rear connection (15.2), wherein the diagnosis is based on a comparison of a measured temporal differential pressure gradient (33) of the measured differential pressure (16) as a result of a change in the exhaust gas pressure upstream of the exhaust gas aftertreatment unit with an expected temporal differential pressure gradient (32) of an expected differential pressure, characterized bythat a correlation is determined between a course of the measured differential pressure gradient (33) and a course of the expected differential pressure gradient (32), and that in the case of a high correlation, an intact differential pressure sensor (15) with intact connections (15.1, 15.2) is concluded, and in the case of a low correlation, a defective differential pressure sensor (15) or one or two defective connections (15.1, 15.2) of the differential pressure sensor (15) is concluded, wherein a cross-correlation value is formed by forming a standardized cross-correlation from the course of the measured differential pressure gradient (33) and the course of the expected differential pressure gradient (32), and that the correlation between the measured and the expected course of the differential pressure gradient (33, 32) is evaluated on the basis of the cross-correlation value, wherein a damaged or interrupted rear connection (15.2) of the differential pressure sensor (15) is inferred if the cross-correlation value is above a predetermined first correlation threshold, preferably above a first correlation threshold greater than 1, particularly preferably above a first correlation threshold greater than 1.2, and / or that a damaged or interrupted front connection (15.1) of the differential pressure sensor (15) is inferred if the cross-correlation value is below a predetermined second correlation threshold, preferably below a second correlation threshold less than 0, particularly preferably below a second correlation threshold less than -0.2 [2] Method according to claim 1, characterized by that the correlation is determined and evaluated over at least two or more than two changes in exhaust pressure. [3] Method according to one of claims 1 to 2, characterized bythat the course of the expected differential pressure is determined at least from an exhaust gas volume flow through the exhaust gas aftertreatment unit and a flow resistance of the exhaust gas aftertreatment unit. [4] Method according to one of claims 1 to 3, characterized by that the measured differential pressure (16) across the exhaust gas aftertreatment unit and / or the expected differential pressure across the exhaust gas aftertreatment unit and / or the exhaust gas volume flow are low-pass filtered to determine the expected differential pressure. [5] Method according to one of claims 1 to 4, characterized by that the diagnosis is only carried out if the differential pressure gradient (32, 33), preferably the expected differential pressure gradient (32), or a measure correlating with the differential pressure gradient (32, 33) exceeds a predetermined value. [6] Method according to one of claims 1 to 5, characterized bythat the differential pressure sensor (15) determines the differential pressure (16) via a particle filter (13). [7] Use of the method according to one of claims 1 to 6 in a gasoline-powered internal combustion engine (10), in which the exhaust system has at least one separate catalyst (12) and a particle filter (13) or a catalyst-particle filter combination or a catalytically coated particle filter. [8] Diagnostic unit (18) for diagnosing a differential pressure sensor (15) for determining a differential pressure (16) across an exhaust gas aftertreatment unit of an exhaust gas purification system in an exhaust system (11) of an internal combustion engine (10), wherein the measurement signals of the differential pressure sensor (15) are supplied to the diagnostic unit (18), wherein the diagnostic unit (18) is designed to form a measured differential pressure gradient (33) from the measured differential pressure (16) and to determine an expected differential pressure gradient (32) from at least one exhaust gas volume flow through the exhaust gas aftertreatment unit and a flow resistance of the exhaust gas aftertreatment unit, characterized by that the diagnostic unit (18) has an electrical circuit arrangement or a program sequence for carrying out the method according to one of claims 1 to 7. [9] Diagnostic unit (18) according to claim 8, characterized bythat the diagnostic unit (18) is a separate unit or an integral part of a higher-level engine control system.

Images