Method for monitoring components of a vehicle related to exhaust emissions

Abstract

The invention relates to a method for monitoring components of a vehicle related to exhaust emissions, comprising: providing emission data comprising information about a time profile (210a) of exhaust emissions of the vehicle; providing status data comprising information about a time profile (212a, 214a, 216a) of each of a plurality of status indicators, wherein each of the plurality of status indicators is an indicator of a status of one component of the vehicle, respectively; determining, based on the emission data and the status data, correlation information about a correlation between a change of the exhaust emissions and a change of at least one of the status indicators; determining, based on the correlation information, maintenance information about at least one of the plurality of components; and providing the maintenance information.

Description

Technical Field

[0001] The present invention relates to a method for monitoring exhaust emission-related components of a vehicle, particularly with respect to maintenance needs, and to a computing unit and computer program for performing the method. Background Technology

[0002] Vehicles, especially those with internal combustion engines, produce exhaust gases or emissions, particularly nitrogen oxides. These emissions are mandated to be kept as low as possible within the framework of climate and environmental protection regulations. Summary of the Invention

[0003] According to the present invention, a method for monitoring vehicle components in terms of exhaust emissions is provided, along with a computing unit and a computer program for performing the method. Furthermore, advantageous configurations are also provided.

[0004] This invention relates to vehicle exhaust emissions and the monitoring of vehicle components involved in exhaust emissions. In particular, it also relates to identifying whether a component requires maintenance.

[0005] Therefore, diagnostic functions can be set up in principle. The diagnostic system can be located in the vehicle's controller, specifically possessing a considerable number of diagnostic functions. These functions can identify and display faults in the system or components, typically including information about the cause of the fault. The basic idea behind displaying fault information is that the fault can be eliminated or subsequently eliminated through repair, usually by replacing the relevant components or sub-components (such as sensors, actuators, or subsystems).

[0006] Especially for vehicles with internal combustion engines, current regulations stipulate the identification and display of exhaust-related or emission-related faults (e.g., via warning lights) as part of the so-called On-Board Diagnostics (OBD) system, provided that the fault causes emissions to exceed specified thresholds. In particular, information regarding the cause of the fault can be determined in a standardized form through standardized interfaces, thereby ensuring effective and cost-effective repairs for vehicle operators. It is generally stipulated that faults displayed in this way should be eliminated, and measures such as requiring regular technical inspections can be taken to prevent the operation of the faulty vehicle.

[0007] Against the backdrop of strengthened climate and environmental protection, the boundary conditions for fault elimination are constantly being redefined, as exemplified by the so-called EU7 regulations. In addition to existing on-board diagnostics requirements, these regulations, for example, introduce new requirements under the overarching concept of "OnBoard Monitoring" (OBM). A fundamental idea of ​​OBM is that, independent of the (and often existing) fault identification points defined within the OBD framework, any exceedance of prescribed emission limits must be identified and displayed, and the fault should or will be enforced to be eliminated through appropriate measures, particularly including limiting possible vehicle operation.

[0008] In this respect, this on-board monitoring departs from the fundamental idea of ​​OBD diagnostics related to component failures, focusing solely on observed emissions increases across the entire vehicle. Therefore, effectively identifying and eliminating the root causes of the failure remains essential, and may even be more significant due to the lack of fault entries in OBD.

[0009] To comply with nitrogen oxide (NOx) emission standards, a NOx sensor can be used as an additional signal transmitting component. One possible failure scenario is that a driver warning is issued and vehicle operation is restricted via the OBM system due to elevated NOx emissions measured by the sensor. However, conventional OBM systems often do not display a fault in this situation. This behavior may be due to the cumulative effects of aging vehicle components, or because individual vehicle components have not yet reached the point where a fault needs to be displayed (in the conventional OBM sense).

[0010] Therefore, repairs are required at a repair shop, but there is no information about the component causing the malfunction in the OBD system.

[0011] Status indicators can be used to describe the state of vehicle components. A component may be assigned multiple status indicators. However, in particular, one status indicator is dedicated to (only) a specific component. This state is specifically the so-called State of Health (SoH). In principle, such status indicators or corresponding signals are already available within the framework of traditional OBD. The basic concept here is to provide the so-called State of Health (SoH) indicators, calculated in or derived from the OBD function, to the repair shop or cloud storage, and, especially when using artificial intelligence methods, to infer the repair plan with the highest probability of success through analysis and evaluation.

[0012] In particular, existing and validated OBD diagnostics typically make diagnostic decisions based on threshold comparisons of characterizing parameters, and it is expected that a larger distance of the current value relative to the threshold is associated with a lower probability of failure, while being close to the threshold indicates the presence of damage or at least the expected (soon) failure.

[0013] Data analysis and evaluation can be conducted, in particular, in the so-called cloud, that is, on a central computing system independent of the vehicle. In the cloud, available information can be linked with maintenance tips, such as model-based or artificial intelligence methods, or based on on-site observations (such as the evaluation of repair shop cases) and additional measurement signals from the inspection procedures to be performed in the repair shop.

[0014] The challenge in identifying components requiring repair lies in appropriately screening and analyzing available data in the sense of component health status (SoH). Selected indicators (condition indices) typically require normalization to determine the recommended maintenance order (ranking) of the condition indices based on direct comparisons, ultimately leading to the recommended maintenance sequence for the components. Ideally, this order should also consider the emissions impact of the failure, particularly nitrogen oxides.

[0015] The method proposed within the scope of this invention provides a simple solution for utilizing exhaust emission values, which are collected, for example, by means of a sensor, specifically a nitrogen oxide sensor.

[0016] However, it has been shown that the problem with cloud-based centralized analysis and evaluation of fleet data lies in data acquisition and availability, requiring a very large number of measurements in order to robustly model the existing relationships across the entire fleet.

[0017] While it's theoretically possible to learn such data models progressively from observed field data, this makes the information needed for maintenance—spanning the entire fleet and specifically targeting faulty vehicles requiring repair—only available at a significantly later point compared to mass production. This is because failure cases must occur before they can be used for modeling or training. Only when a sufficient number of vehicles have failed in the field does a substantial amount of data containing information important for maintenance (e.g., how a vehicle with a faulty component behaves) exist.

[0018] In contrast, the method proposed in this invention contributes to the solution without relying on the cloud (at least in the sense of the aforementioned cloud solutions) by analyzing and evaluating data over time, particularly individual vehicle data (i.e., vehicle history data). The analysis and evaluation can be performed locally on the vehicle, but also externally, for example, in the so-called cloud (but based on individual vehicle data).

[0019] Therefore, the analysis, evaluation, and deduction of vehicle maintenance prompts can be carried out under fault conditions, and are independent of the time of occurrence (relative to the fleet's lifespan, i.e., the so-called fleet age), as well as the data of other vehicles in the fleet.

[0020] Another challenge lies in obtaining quantitative information about the actual emissions impact of failures and component repairs, as emissions measurements, such as those for nitrogen oxides, are typically not possible in repair shops. Quantitative detection of component failures can be achieved by detecting changes in emissions after repair compared to before repair; however, this information is only available when the corresponding failure occurs in the field. Obtaining this information early requires significant overhead costs associated with targeted, artificial degradation of components within a test fleet.

[0021] The method proposed in this invention avoids the aforementioned problems by modeling the emission impact of changes in the State of Health (SoH) index and estimating emission improvements in vehicles based on this model.

[0022] The proposed solution follows this concept: independently of fleet data, based solely on historical data (particularly measured values) of individual vehicles, especially measured NOx emissions, to deduce the appropriate sequence (ranking) of maintenance measures and conclusions regarding their expected emission improvements, or at least to determine whether any components require maintenance and / or which components require maintenance. The proposed solution is based on correlating changes in exhaust emission values ​​with changes in condition indicators.

[0023] Provides emissions data and status data. Emissions data includes information on the time trend of vehicle exhaust emissions (i.e., specifically the specific values ​​of exhaust emissions); exhaust emissions or their values ​​can be detected or have been detected, especially by means of sensors. Status data includes information on the time trend of each of several status indicators (i.e., specifically the specific values ​​of the status indicators); each of the several status indicators is an indicator of a state of a component of the vehicle. As mentioned earlier, multiple different status indicators can also be set for the same component. Status indicators can be adaptive values ​​for mixture adaptation and deviations of the adaptive values ​​from the expected linear relationship. Other indicators can be adjustment deviations or adaptive values ​​of sensors (such as air mass flow meters), or correction factors for the injected fuel quantity, which should be used to identify potential maintenance needs for fuel components. Especially in gasoline systems, probes and catalytic converters have a significant impact on emissions, and their degradation is indicated by the corresponding OBD diagnostic criteria.

[0024] Subsequently, correlation information is determined based on emission and status data. This correlation information pertains to the relationship between changes in exhaust emissions and changes in at least one of the status indicators. Maintenance information is then determined based on this correlation information for at least one of a plurality of components, and this maintenance information is provided so that, for example, the affected component can be repaired. The maintenance information may include: whether any of the components require maintenance and / or which of the components requires maintenance; it may also consider the prioritization of maintenance needs for the components, i.e., information on which component should be repaired next and, if necessary, which component should be repaired next, and then information is provided accordingly.

[0025] Therefore, the status indicator is evaluated in terms of its impact on exhaust emission values. The status indicator (SoH) can in particular be an internal controller signal, which is also used, for example, as a conventional OBD signal. However, it should be noted that not all status indicators used in conventional OBD must be used within the scope of the method of this invention. Other status indicators not used in conventional OBD can also be considered within the scope of the method of this invention.

[0026] Here, two implementation methods are specifically considered, which can be used individually or in combination.

[0027] In one implementation, the correlation information includes checking whether changes in exhaust emissions occur simultaneously with changes in at least one of the state indicators, within a pre-defined standard range. Specifically, changes exceeding a pre-defined minimum value are considered changes in exhaust emissions (or their values). Therefore, such changes should particularly involve step changes. This could, for example, apply to the average values ​​within a time window before and after the change. This also applies to changes in at least one of the state indicators (or their values).

[0028] "Changes occurring simultaneously within a pre-defined standard range" specifically includes situations where these changes occur within a pre-defined relative time interval. This leads to the consideration that changes in exhaust gases and state indicators, or changes in corresponding components, may manifest with a time delay, but still originate from the same cause.

[0029] Correlation (correlation information) provides a preferred calculation scheme for calculating information about whether such a change has occurred and, if so, for which state indicators this change has occurred. This is because causality can be inferred from the simultaneity of events (which, if necessary, can be ensured by appropriately selecting state indicators).

[0030] In another implementation, the correlation information includes a model regarding the increase in emissions caused by the fault, determined based on emission data and status data, respectively, over a pre-given time period. The correlation model here provides the dependency between exhaust emissions and at least one of the status indicators, preferably all of all multiple status indicators. Modeling can be implemented, for example, using an online algorithm for adaptively adjusting a linear model. Therefore, maintenance information is determined based on this correlation model. The pre-given time period here specifically includes a pre-given duration up to each current time point; each current time point can be equivalent to the current calculation time point or the current measurement time point of the most recent measurement. The time period shifts so that older data is eventually no longer considered.

[0031] Therefore, this analysis and evaluation, particularly during vehicle operation, is conducted continuously over time by adaptively mapping the evolution of exhaust emissions with aging (and in the event of potential failures) into the model. Both continuous deterioration and sudden failures lead to changes in condition indicators and exhaust emissions. In failure conditions, condition indicators and exhaust emissions change abruptly without observing intermediate values. In cases of continuous aging, condition indicators and exhaust emissions increase (not necessarily linearly, but continuously).

[0032] In summary, in both implementations, gradual changes in condition indicators and exhaust emissions can be expected as partial damage occurs (i.e., particularly when a failure occurs without the cause being identified by OBD). At this point, multiple condition indicators may change simultaneously (depending on the choice of condition indicator). With continued aging, the correlation between increased exhaust emissions and condition indicators may not be able to be determined by parallel changes over time.

[0033] The computing unit according to the invention, such as a controller of a motor vehicle or a computer, is configured, particularly by means of programming techniques, to perform the method according to the invention.

[0034] It is also advantageous to implement the method of the invention in the form of a computer program for performing all the method steps or a computer program product having program code, because doing so incurs particularly low costs, especially if the execution controller is also used for other tasks and therefore always exists. Finally, the invention provides a machine-readable storage medium on which the computer program as described above is stored. Suitable storage media or data carriers for providing the computer program are, in particular, magnetic storage, optical storage, and electronic storage, such as hard disks, flash memory, EEPROM, DVDs, etc. The program can also be downloaded via a computer network (Internet, intranet, etc.). Such downloading can be wired or wireless (e.g., via WLAN networks, 3G, 4G, 5G, or 6G connections, etc.).

[0035] Further advantages and configurations of the present invention can be derived from the specification and drawings.

[0036] The present invention is illustrated schematically with reference to the accompanying drawings and is described below with reference to the drawings. Attached Figure Description

[0037] Figure 1 A vehicle in which the method of the present invention can be performed is illustrated schematically.

[0038] Figure 2a , 2b The process of the method of the present invention in a preferred embodiment is illustrated schematically.

[0039] Figure 3 The different trends of pressure and other parameters in the fluid supply system are illustrated to explain the different diagnostic results of the method of the present invention in a preferred embodiment.

[0040] Figure 4a , 4b The process of the method of the present invention in another preferred embodiment is illustrated schematically. Detailed Implementation

[0041] Figure 1 A vehicle 100 in which the methods of the present invention can be performed is schematically illustrated. The vehicle 100 has an internal combustion engine 110 and, exemplary, two components 112, 114, which are used, for example, for exhaust aftertreatment or otherwise affect the exhaust emissions of the vehicle 100. Furthermore, the vehicle 100 has a nitrogen oxide sensor 120 by means of which nitrogen oxide emissions, as an example of exhaust emissions, can be measured.

[0042] In addition, the vehicle 100 has a computing unit 130 configured as a controller, such as an engine controller. This controller can, for example, obtain the measurement value of the nitrogen oxide sensor 120, and can also obtain other measurements or other data, based on which the status indicators of components 112 and 114 can be determined or calculated.

[0043] For this purpose, a central computing unit 150 is exemplarily shown, which may, for example, represent a so-called cloud. The proposed method for monitoring vehicle exhaust emission-related components can be executed, in particular, in computing unit 130, but also on central computing unit 150, which obtains relevant data from computing unit 130, for example, via appropriate data transmission. However, regardless of where the method or the specific necessary calculations are performed, data relating to the individual vehicle 100 is invoked.

[0044] exist Figure 2a , 2bThe figure shows a graph illustrating two different implementations of the proposed method. Figure 2a and Figure 2b The upper graph shows the change of exhaust emission 202 with respect to time 200, and the lower graph shows the change of condition index 204 with respect to time 200.

[0045] Specifically, in Figure 2a The upper curve graph shows the time trend of exhaust gas emissions 210a, in Figure 2b The graph above shows the time trend of exhaust emissions 210b. Specifically, it may involve nitrogen oxide emissions. Figure 2a The graph below shows the time trends of three different state indicators: 212a, 214a, and 216a. Figure 2b The graph below shows the time trends of three different state indicators: 212b, 214b, and 216b.

[0046] It should be noted that the trends shown in the various graphs, namely exhaust emissions and status indicators, or the values ​​they are based on, were measured at (at least approximately) the same point in time. The corresponding assignments can be made, for example, using the timestamps of the individual measurements.

[0047] These trends, or the data they are based on, can be detected over a longer period of time. Since vehicles typically move and thus travel a distance, time trends are also correlated with location trends.

[0048] Figure 2a , 2b The status indicators in both can be the same, but for the purpose of illustration, data from different vehicles or different time periods are involved, which can also be seen from the slightly different specific trends.

[0049] To monitor components, both implementations now provide emissions data and status data. Emissions data includes information about the temporal trends of vehicle exhaust emissions, such as, for example, in... Figure 2a , 2b As shown in the graph above. State data includes information about the time trend of each of several, for example, three state indicators, such as... Figure 2a , 2b As shown in the graph below. Each status indicator represents a status of a component of the vehicle.

[0050] from Figure 2a , 2bThe basic concepts of the two implementation methods can be seen from the first point. These can also be described as two paths used to determine correlations. As mentioned earlier, the two implementation methods or paths can also be used in parallel or together to better monitor components and ultimately determine their potential maintenance needs.

[0051] In both implementations, correlation information is determined based on emission data and state data to determine the correlation between changes in exhaust emissions and changes in at least one, but preferably all, state indicators.

[0052] exist Figure 2a In the upper graph, specifically in the time trend chart 210a of exhaust emissions, three points show significant changes in the trend, i.e., step changes. Each of these three changes is temporally correlated with a change in one of the condition indicators, indicated by double arrows. This demonstrates that this data analysis method can identify the correlation between condition indicators and exhaust emissions, and in particular, can identify specific condition indicators and thus pinpoint the causes of failures in specific components, thereby identifying those that may require maintenance.

[0053] exist Figure 2b In the graph above, specifically the time trend of exhaust emissions (210b), slight fluctuations are visible, but no significant changes are observed. The same applies to the condition indicators. While this doesn't directly establish a causal relationship between the condition indicators or related components and the changes in exhaust emissions, a certain correlation exists, as can be seen from the overall upward trend. Therefore, further analysis is possible, as detailed below.

[0054] With the help of Figure 3 First, a detailed explanation of the basis Figure 2a The implementation method or path. (and) Figure 2a Similarly, the upper graph shows the change in exhaust emissions 302 with respect to time 300. The lower graph shows the time trends 312 and 314 for two different state indicators.

[0055] This section illustrates one possible implementation of the basic concept: determining the contribution of component localized damage to the condition index by analyzing and evaluating exhaust emissions and condition indicators. Simultaneous changes can be determined by analyzing and evaluating the emission and condition data from this trip, or the average emission and condition values ​​before and after a significant change in exhaust emissions. If the average values ​​of the two test measurement periods 320 and 322 are significantly different for both exhaust emissions and the condition index (here, condition index 312), then simultaneous changes are identified. As detailed above, this correlation is a preferred method for determining the co-significant changes in condition parameters and emissions.

[0056] Changes in exhaust emissions are specifically attributed here to state indicators that exhibit the largest average variation. Alternatively, for example, the combination of variations can be stored as a representation of a component, and the variation can be attributed to that component. Instead of the average value, the variation in exhaust emissions can be attributed using the correlation between exhaust emissions and sample values ​​of state indicators as a criterion. In the simplest (and preferably) case, exhaust emission values ​​and state indicator values ​​obtained for the journey under appropriate conditions are used as sample values.

[0057] The described approach allows for continuous assessment of the presence of simultaneous changes in exhaust emissions and status indicators (using appropriately selected simultaneous change quality criteria, as explained above) through a series of events. In the current case, for example, the incremental portion 330 of exhaust emissions is assigned to a status indicator or identified component as a total contribution to the increase in emissions. In the simplest case, it is assigned to the status indicator with the highest correlation (or more generally: the maximum value of the simultaneously changing quality criteria). In more complex cases, screening criteria consisting of multiple quality criteria can be defined, particularly when multiple component failures cause changes in multiple status indicators.

[0058] The unique advantage of analysis and evaluation based on this approach lies in utilizing additional information about the temporal correlation between changes in state indicators and exhaust emissions, compared to the absolute values ​​of state indicators. Therefore, unlike schemes that analyze and evaluate state indicators at maintenance time points, this approach supplements and analyzes information about the temporal trend of exhaust emission changes. In contrast, models based on state indicators at maintenance time points only consider the final state of the state indicators, completely ignoring the information derived from the temporal trend of exhaust emissions results (measured by vehicle sensors) associated with the state indicators.

[0059] Now, with the help of Figure 4a , 4b Detailed explanation based on Figure 2b The implementation method or path is described below. Here, the changes in exhaust emission 402 with respect to state index 404 are shown in graphs. Here, circles indicate specific pairs (numerical pairs) of values ​​for exhaust emission and state index, which were detected (at least approximately) at the same time point.

[0060] In continuous path analysis, particularly during vehicle operation, the evaluation is performed over time by adaptively mapping the evolution of exhaust emissions with aging (and in the event of a fault) into the model (correlation model). Both continuous deterioration and sudden faults lead to changes in state indicators and exhaust emissions. In fault conditions, state indicators and exhaust emissions change abruptly without intermediate values. In continuous aging conditions, state indicators and exhaust emissions increase, not necessarily linearly, but continuously.

[0061] Figure 4a This situation is illustrated. 440 represents a numerical pair from an earlier time period, and 442 represents a numerical pair from a later time period, where the condition indicators and exhaust emissions show a step change. This indicates a fault. The model created based on these numerical pairs 440 and 442 is labeled 452. 444 represents a numerical pair representing an alternative scenario with continuous changes, where the condition indicators and exhaust emissions change only slightly. This indicates continuous aging. The model created based on these numerical pairs 440 and 444 is labeled 454.

[0062] Modeling exhaust emissions as a whole with respect to state indicators during vehicle operation offers the possibility, in both cases, of inferring achievable emission reductions through maintenance from the dependence of exhaust emissions on state indicators. Here, considering the extent to which recent experiences are incorporated into the model (the so-called forgetting factor; referred to above as the pre-given time period), in aging vehicles, maintenance of parts is expected to achieve only a partial emission reduction compared to new vehicles. Therefore, this model is a local approximation of the emission effect.

[0063] The emission contributions of each state indicator derived from the model can be directly interpreted as a ranking, that is, providing information on "which state indicators can contribute to emission reduction and how much".

[0064] The advantage of the proposed approach is that it does not require normalization of the input (state indicators), that is, it does not require prior knowledge of the relationship between state indicators and exhaust emissions (which may be yet to be determined during development), and especially does not require knowledge of the state indicator threshold when exhaust emissions exceed the OBM threshold.

[0065] When using models, a simple starting point might be to choose state indicators such that the emission contribution derived from those state indicators can be used to deduce component maintenance. This selection is typically based on expert knowledge, and supplementarily, more complex models in the cloud based on fleet information can provide information about the correlation between state indicators and components. In particular, even with careful selection of state indicators, it may be unavoidable in individual cases that a single component affects multiple state indicators, and therefore it may be necessary to map the exhaust emissions derived from those state indicators to components through transformations (e.g., from fleet information).

[0066] As a simple modeling method, the least squares method can be used, which can be implemented recursively to save resources. The advantages of recursive least squares are robustness, statistical interpretability, and resource conservation, which are also reflected in the requirement for the amount of available data. Furthermore, by pre-linearizing the state indicators, nonlinear relationships can also be depicted.

[0067] The following section will elaborate on the possibilities of modeling, specifically for discontinuous paths (i.e., changes based on faults) and continuous paths (i.e., changes based on aging).

[0068] In discontinuous path scenarios, different approaches can be used to assess temporally correlated changes in state indicators and emissions, all based on the same core idea. By analyzing and evaluating state indicators and emissions before and after a hypothetical failure event (emission jump), a quality standard is derived, its value being a quantitative measure of temporal overlap. Besides the previously mentioned correlation (cross-correlation) and statistical tests (single-sample or two-sample tests) used to calculate emissions and state indicators separately, a simple probability-based assessment approach can also be employed.

[0069] For example, a Pearson correlation coefficient r can be used here, where individual values ​​i of variables x (state indicators) and y (exhaust emissions) are used, preferably one value per trip. Here, n is the total number of trips considered, and the underlined x and y represent the arithmetic mean.

[0070] Another possibility is the so-called one-sample t-test, used to calculate the empirical probability of a significant change (the difference between the values ​​before and after the failure event is not equal to zero). Instead of performing a one-sample t-test on the difference, a two-sample t-test can be performed on the values ​​of the two test sets.

[0071] Similarly, one could consider, for example, the so-called Yur coefficient, in which the correlation between the premise (changes in state indicators) and the conclusion (changes in exhaust emissions) can be determined.

[0072] In the case of continuous paths, a basic approach can be to use the known recursive least squares (RLS) method to identify the parameters of the linear model used to obtain the state indices online. Here, to obtain the state indices, the contribution of each state index to the observed increase in emissions can be calculated. The specific calculation process is as follows: Here, the model output variable y is based on the input h and the parameter vector. The calculation is as follows: This is the basic scheme of the RLS method. The estimated value of the current emission y (in this application, a measurement of exhaust emissions) is obtained as a weighted linear combination of the input variable state indices (vector h). The estimator provides the parameters currently used for weighting.

[0073] If new state index values ​​exist, this is addressed by recalculating matrix P and vector P in each update step. To perform online updates. This is done by pre-setting a forgetting factor. This method assigns higher weights to more recent state indicator values. The advantage of this recursive approach is that it does not require storing large amounts of data and the computational cost remains constant over time.

[0074] When applying the RLS method to the proposed support maintenance measures (so-called precise positioning) as suggested, it can be set up using vectors. The value is used to calculate the contribution of each state indicator to the increase in emissions yi. To do this, the change in the state indicator relative to a reference value (a fault-free vehicle) is proportionally plotted against a vector, where necessary. The corresponding parameters are multiplied. When exhaust emissions change linearly according to state indicators, the sum of the contributions to yi will be equivalent to the total increase in emissions. When there is a non-linear relationship between exhaust emissions and state indicators, the calculated value yi can vary depending on the forgetting factor. The forgetting factor controls, in a desired way, the extent to which past emission increases that occurred concurrently with increases in state indicators are still considered when obtaining the output variable yi.

[0075] Figure 4b The relationship is illustrated. For the numerical pairs shown, from the known normal functional state on the left to the current state on the right, 460 represents the change (increase) in the state index, and 462 represents the increase (change, yi) in emissions. Different models (here, lines) are represented by 470, 472, and 474. Due to the forgetting factor, the forgetting factor increases from 470 through 472 to 474, therefore earlier numerical pairs are considered less frequently.

Claims

1. A method for monitoring vehicle components (112, 114) in relation to exhaust emissions, comprising: Provide emissions data, which includes information on the time trends (210a, 210b) of the vehicle's exhaust emissions; Provide status data, which includes information on the time trend (212a, 214a, 216a, 212b, 214b, 216b) of each of a plurality of status indicators, wherein each of the plurality of status indicators is an indicator of a state of a component of the vehicle. Based on the emission data and the state data, determine correlation information regarding the correlation between changes in exhaust gas emissions and changes in at least one of the state indicators; Based on the correlation information, maintenance information is determined for at least one of the multiple components; as well as Provide repair information.

2. The method according to claim 1, wherein, Determining relevance information includes: Check whether the change in exhaust emissions (330) occurs simultaneously with a change in at least one of the stated state indicators, within the range of a pre-defined standard. The relevant information includes the following: whether such a change has occurred, and if so, for which state indicators the change has occurred.

3. The method according to claim 2, wherein, The pre-defined criteria include: changes in exhaust emissions and changes in at least one of the state indicators occur within a pre-defined time interval relative to each other.

4. The method according to any one of the preceding claims, wherein, The correlation information includes models (452, 454), which are determined based on emission data and state data within a pre-given time period. The correlation model indicates the dependence of exhaust emissions on at least one of the state indicators, preferably on all of the multiple state indicators.

5. The method according to claim 4, wherein, The pre-given time period includes the pre-given duration up to each current time point.

6. The method according to any one of the preceding claims, wherein, Repair information for at least one of the multiple components includes at least one of the following: Information regarding whether any of the multiple components require repair. Information about which of the multiple components needs repair; Information regarding the prioritization of repair needs for multiple components.

7. The method according to any one of the preceding claims, wherein, Exhaust gas emission values ​​include nitrogen oxide emission values.

8. The method according to any one of the preceding claims, wherein, The vehicle's exhaust emissions are detected or have been detected by means of a sensor (120), which is particularly located on or inside the vehicle.

9. A computing unit (130) configured to perform all method steps of the method according to any one of the preceding claims.

10. A computer program, when executed on a computing unit, causes the computing unit to perform all the method steps of the method according to any one of claims 1 to 8.

11. A machine-readable storage medium having a computer program as claimed in claim 10 stored thereon.

Images