GB2607300A

GB2607300A - A method for determining an active regeneration process of a gasoline particulate filter of an exhaust system, as well as an exhaust system

Info

Publication number: GB2607300A
Application number: GB2107790.4A
Authority: GB
Inventors: Smiroldo Rigel; Hofen Thomas; Gianunzio Michael
Original assignee: Daimler AG
Current assignee: Mercedes Benz Group AG
Priority date: 2021-06-01
Filing date: 2021-06-01
Publication date: 2022-12-07
Also published as: GB202107790D0

Abstract

A method for determining an active regeneration process 22 of a gasoline particulate filter 14 of an exhaust system 12 of a motor vehicle 10 comprises capturing a current exhaust gas temperature T at the gasoline particulate filter by a temperature sensor 16, transmitting the exhaust gas temperature to an electronic computing device 18, determining a current operating mode 24 of a combustion engine 26 of the vehicle, transmitting the operating mode to the electronic computing device, and determining whether the active regeneration process is happening depending on the transmitted exhaust gas temperature and the transmitted current operating mode by using a machine learning algorithm 20 of the electronic computing device. An exhaust system for a motor vehicle has at least one gasoline particulate filter, at least one temperature sensor and at least one electronic computing device with a machine learning algorithm configured to carry out the method.

Description

A METHOD FOR DETERMINING AN ACTIVE REGENERATION PROCESS OF A GASOLINE PARTICULATE FILTER OF AN EXHAUST SYSTEM, AS WELL AS AN EXHAUST SYSTEM

FIELD OF THE INVENTION

[0001] The invention relates to the field of automobiles. More specifically, the invention relates to a method for determining an active regeneration process of a gasoline particulate filter of an exhaust system of a motor vehicle.

BACKGROUND INFORMATION

[0002] Conventional monitoring strategies cannot sufficiently determine whether an observed temperature increase in a gasoline particulate filter is related to the desired increase of exhaust gas temperature during an active regeneration event or is caused by an unrelated change in vehicle operating conditions, for example, a driver demand change, road load change uphill and downhill.

[0003] WO 2019072886 Al discloses a method and a device for determining a concentration in a combustion exhaust gas of an internal combustion engine such as, for example, a diesel particulate filter. However, there is a need in the art to differentiate the root cause for the temperature increase, and to use that root to allow the monitor to only run, when active regeneration is in progress.

SUMMARY OF THE INVENTION

[0004] It is an object of the invention to provide a method as well as an exhaust system, by which a robust determining of an active regeneration process of a gasoline particulate filter of an exhaust system of a motor vehicle is realized.

[0005] This object is solved by the independent claims. Advantageous forms are presented in the dependent claims.

[0006] One aspect of the invention relates to a method for determining an active regeneration process of a gasoline particulate filter of an exhaust system of a motor vehicle. A current exhaust gas temperature is captured at the gasoline particulate filter by a temperature sensor of the exhaust system. The exhaust gas temperature is transmitted to an electronic computing device of the exhaust system. A current operating mode of a combustion engine of the motor vehicle is determined. The current operating mode is transmitted to the electronic computing device. The active regeneration process is determined depending on the transmitted exhaust gas temperature and the transmitted current operating mode by using a machine learning algorithm of the electronic computing device.

[0007] Therefore, a robust way is presented for determining the active regeneration process of a gasoline particulate filter of an exhaust system of a motor vehicle. In particular, the method provides a way to differentiate the root cause for the temperature increase and can thus be used to allow the monitor to only run, when the active regeneration is in progress.

[0008] According to an aspect of the invention, the method uses machine learning to find a different approach to identify the separation. Machine learning technologies analyze the incoming signals to the engine controller and can differentiate between different modes of the combustion engine operation and its effect on the gasoline particulate filter inlet temperature, which means, it can differentiate between the active regeneration and a driving mode. The machine learning-based methods are also used to decide to activate/deactivate the monitor at the appropriate time.

[0009] In case of false passing, meaning a defective system is not being recognized, vehicle sales approval may be withheld during on-board diagnostic certification, for example, if found during the required on-board diagnostic demonstration prior to approval. If false passing is determined after sales approval, vehicles in customer hands would need to be recalled and repaired at company's expense.

[0010] In case of false failing, which means a non-defective system is incorrectly determined as failed and the malfunction indication light is subsequently turned on, a correctly functioning vehicle needs to be serviced up to the extent that the exhaust system, including the gasoline particulate filter, is replaced prematurely. Once a certain amount of vehicles in customer hand exceed a threshold determined by the regulatory agency, all vehicles need to be recalled for repair at company's expenses. Fines can be imposed against the manufacturer.

[0011] In an embodiment the current exhaust gas temperature is captured at an inlet of the gasoline particulate filter.

[0012] In another embodiment, a beginning of the determining process of the gasoline particulate filter is determined by the machine learning algorithm.

[0013] In particular, the method is a computer-implemented method. Therefore, a computer program product with program code means is also presented according to this invention.

[0014] Another aspect of the invention relates to an exhaust system for a motor vehicle, comprising at least one gasoline particulate filter, at least one temperature sensor and at least one electronic computing device equipped with a machine learning module, wherein the exhaust system is configured to perform a method according to the preceding aspect. In particular, the method is performed by the exhaust system.

[0015] Another aspect of the invention relates to a motor vehicle comprising the exhaust system according to the preceding aspects.

[0016] Advantageous forms of configurations of the method are to be regarded as advantageous forms of the exhaust system as well as the motor vehicle. The exhaust system and the motor vehicle comprise means for performing the method.

[0017] Further advantages, features, and details of the invention derive from the following description of a preferred embodiment as well as from the drawing. The features and feature combinations previously mentioned in the description as well as the features and feature combinations mentioned in the following description of the figure and/or shown in the figure alone can be employed not only in the respectively indicated combination but also in any other combination or taken alone without leaving the scope of the invention.

BRIEF DESCRIPTION OF THE DRAWING

[0018] The novel features and characteristic of the disclosure are set forth in the appended claims. The accompanying drawing, which is incorporated in and constitutes a part of this disclosure, illustrates an exemplary embodiment and together with the description, serves to explain the disclosed principles. The same numbers are used throughout the figure to reference like features and components. Some embodiments of system and/or methods in accordance with embodiments of the present subject matter are now described below, by way of example only, and with reference to the accompanying figure.

[0019] The drawing shows in a perspective top view an embodiment of a motor vehicle comprising an embodiment of an exhaust system.

[0020] In the figure the same elements or elements having the same function are indicated by the same reference signs.

DETAILED DESCRIPTION

[0021] In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration". Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

[0022] While the disclosure is susceptible to various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawing and will be described in detail below. It should be understood, however, that it is not intended to limit the disclosure to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternatives falling within the scope of the disclosure.

[0023] The terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion so that a setup, device or method that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus preceded by "comprises" or "comprise" does not or do not, without more constraints, preclude the existence of other elements or additional elements in the system or method.

[0024] In the following detailed description of the embodiment of the disclosure, reference is made to the accompanying drawing that forms part hereof, and in which is shown by way of illustration a specific embodiment in which the disclosure may be practiced. This embodiment is described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be taken in a limiting sense.

[0025] Fig. 1 shows a schematic top view of an embodiment of a motor vehicle 10 comprising an exhaust system 12. The exhaust system 12 comprises at least one gasoline particulate filter 14, at least one temperature sensor 16 and at least one electronic computing device 18 equipped with a machine learning module 20.

[0026] According to an embodiment of a method for determining an active regeneration process 22 of the gasoline particulate filter 14 of the exhaust system 12 a capturing of the current exhaust gas temperature (T) at the gasoline particulate filter 14 is performed by the temperature sensor 16. The exhaust gas temperature T is transmitted to the electronic computing device 18. A current operating mode 24 of a combustion engine 26 of the motor vehicle is determined. The current operating mode 24 is transmitted to the electronic computing device 18. The active regeneration process 22 is determined depending on the transmitted exhaust gas temperature (T) and the transmitted current operating mode 24 by using the machine learning module 20.

[0027] According to an embodiment, the current exhaust gas temperature (T) is captured at an inlet 28 of the gasoline particulate filter 14. In another embodiment, a beginning of the determining process of the gasoline particulate filter 14 is determined by the machine learning module 20.

[0028] In particular, Fig. 1 shows a machine learning module based concept for in-vehicle monitoring (OBD, on-board diagnostic) of the activity of an actively commanded combustion engine operating state with the purpose of regenerating a gasoline particulate filter 14.

[0029] The gasoline particulate filter 14 mechanically captures the soot particles and thus prevents their emission into the environment. To maintain its efficiency and flow rate during operation of the vehicle lifetime, the accumulated ashes need to be burned off in a process called regeneration. A certain temperature of the gasoline particulate filter 14 needs to be reached and maintained for some time for this regeneration to be successful.

[0030] The regeneration process is divided into several operation modes. There is a service regeneration mode known which is performed at the dealership during the service, wherein the motor vehicle 10 is in a standstill mode. Furthermore, a passive regeneration is known while driving the motor vehicle 10 without additionally active control strategies in the engine controller. The active regeneration process 22 is performed while driving the motor vehicle 10 and with active change in commanded engine operation mode with the goal of a temperature increase of the exhaust gas for the purpose of the regeneration of the gasoline particulate filter 14.

[0031] Based on the available sensor information in the exhaust system 10, the invention determines a specific increase of the temperature at the inlet 28 of the gasoline particulate filter 14 based on a commanded engine operating state for the purpose of the active regeneration process 22. To allow for robust monitoring, in particular a reliable determination of the state of the monitored system, without allowing for false positives or false negatives, sufficient separation of the evaluated system signals is required. The separation indicates the difference of certain signal values of a performing system versus a non-performing system. For many monitors, separation cannot be achieved across all engine operation states, therefore, enabled conditions can be identified, allowing the monitor to only run when all environmental conditions allow for robust pass/fail determination. These monitors are considered non-continuous.

[0032] Based on a different combustion process resulting in different temperature profiles, in diesel engines it is possible to find the required separation with conventional monitoring concepts. In gasoline engines, however, which is in particular the combustion engine 26, the difference of the measured temperature of the exhaust gas flow can be caused by changing operating mode due to driver input, for example, acceleration or hill climbing, at the same time that the active regeneration process 22 commands a change in engine operation mode to allow for a temperature increase of the exhaust gas. From a monitoring perspective, the contribution to the temperature increase cannot be clearly distinguished just from the information provided by the temperature sensor 16 at the inlet 28 of the gasoline particulate filter 14.

[0033] With the machine learning module 20, a solution/model to analyze the signals available during engine operation to allow for a robust monitor enablement and a reliable pass/fail determination. Undetected patterns and relationships between several provided signals are identified and then used in a simplified model capable to run on an engine controller during regular vehicle operation. Since the functionality of the monitoring system has to be ensured along the degradation of the components up to the regulated lifetime, provision to take this aging in consideration into account is important.

[0034] The configuration of this methodology may require two parts. First, a given list of context signals, and secondly the kernel to be used for each conditional model in a Gibbs scheme. Both are chosen a priori using Bayesian optimization over all signal and kernel variants with an expected improvement acquisition function aimed to minimize the mean absolute error of the predicted signal to ground truth. According to this embodiment, for a given list of contact signals, a circular buffer with length 60 seconds records the historical value for each signal. 60 seconds is chosen heuristically, being more than double the expected forecast length. Higher buffer values would work, lower buffer values would still work with reduced efficiency. This value is just exemplary and does not restrict the invention.

[0035] A joint distribution over all signals in this buffer is estimated. This joint distribution is constructed using Gibbs sampling, which iterates over many conditional distributions to sample from high probability value areas of the joint distribution. These samples act as the representation of the joint distribution. The relevant conditional distributions, of every signal conditioned on every other signal, is formed using Gaussian process regression with some given kernel.

[0036] The circular buffer is extended 10 seconds into the future, for example only, for a total of 70 seconds. The value of the signals in these future 10 seconds is broadcasted to simply be the value they held at the 60 seconds mark, in particular held constant, as an initialization. Using one of the conditional models described above, the value for just one signal over these additional 10 seconds is then predicted from the other signals, resulting in an estimate of that signal's value over the 10 second future if the other sample values were accurate. Due to initially being held constant, those values are not accurate at

S

initialization. This is then iterated over every signal multiple time, finally resulting in a set of predicted signal values for all signals which are both consistent with each other and consistent with a recorded 60 seconds history. Thus, this Gibbs sampling scheme results in a high probability estimate for the value of every signal over the next 10 seconds. This process is used for showing the estimated future value of all signals from every other signal in this iterative way, as the "Gibbs scheme".

[0037] Using the Gibbs scheme, the future value of all signals is estimated with the regeneration signal set to "on". This is not a predicted estimate, the future regeneration cycle is to be on and using the Gibbs scheme the value of all other variables in that future are estimated. The same procedure is repeated with the regeneration signal "off". The values of the filtered temperature in each of the high probability worlds are compared, and the difference in the predicted values are then labeled as the "separation'. This predicted separation value forms the basis for the monitoring procedures downstream to be described below.

[0038] For each component or system monitored for the purpose of the on-board diagnostics, a subset of measured vehicle input variables and calibratable parameters defines the enabling conditions in which the monitor will be released for evaluation. For example, the ingested engine air mass as measured by the mass air flow sensor is variable with speed and load. If a downstream system has a high correlation to intake air mass, not only within a specific range, the enabling conditions for release will constrain monitoring via calibratable parameters before evaluation of the downstream system is permitted.

[0039] The context signal list chosen through Bayesian optimization defines the monitoring enabling conditions. Each variable selected is assigned boundary conditions via calibratable parameters, which are then optimized to create a monitoring release range in order to return maximum robustness for both passing margin of a performing system and detection margin of non-performing system. Optimization of the model and the enabling conditions are initially determined on a performance system by statistically minimizing the temperature difference between the protective model temperature provided by the Gibbs scheme described above and the measured temperature diagnosis of a non-performing system is carried out within the same enabling conditions window defined using the performing system. The threshold for detection of a non-performing system is determined by the operating characteristics of a faulty component, the statistical separation and regulated demonstration emission limits.

[0040] Ultimately, the output of the machine learning algorithm 20 provides the fundamental variable list to enable condition creation while simultaneously creating a predictive model for a performing system able to determine a non-performing thermal control during an active gasoline particulate filter regeneration request.

Reference Signs motor vehicle 12 exhaust system 14 gasoline particulate filter 16 temperature sensor 18 electronic computing device machine learning algorithm 22 active regeneration process 24 operating mode 26 combustion engine 28 inlet T temperature

Claims

CLAIMS1. A method for determining an active regeneration process (22) of a gasoline particulate filter (14) of an exhaust system (12) of a motor vehicle (10), comprising the steps of: - capturing a current exhaust gas temperature (T) at the gasoline particulate filter (14) by a temperature sensor (16) of the exhaust system (12); -transmitting the exhaust gas temperature (T) to an electronic computing device (18) of the exhaust system (12); - determining a current operating mode (24) of a combustion engine (26) of the motor vehicle (10); - transmitting the current operating mode (24) to the electronic computing device (18); and - determining the active regeneration process (22) depending on the transmitted exhaust gas temperature (T) and the transmitted current operating mode (24) by using a machine learning algorithm (20) of the electronic computing device (18).
2. The method according to claim 1, characterized in that the current exhaust gas temperature (T) is captured at an inlet (28) of the gasoline particulate filter (14).
3. The method according to claim 1 or 2, characterized in that a beginning of the determining process of the gasoline particulate filter (14) is determined by the machine learning algorithm (20).
4. An exhaust system (12) for a motor vehicle (10), comprising at least one gasoline particulate filter (14), at least one temperature sensor (16) and at least one electronic computing device (18) with a machine learning algorithm (20), wherein the exhaust system (12) is configured to perform a method according to any one of claims 1 to 3.