Method and device for determining particulate matter in vehicle dpf, computer equipment

Abstract

The application relates to a method and device for determining particulate matter in a vehicle DPF, computer equipment and a storage medium. The method comprises the following steps: acquiring current working condition data of a vehicle and current differential pressure data in the DPF; determining a carbon load prediction model corresponding to the current working condition data based on the current working condition data, wherein the carbon load prediction model comprises a corresponding relationship between carbon load and differential pressure; determining a carbon load differential pressure according to the current differential pressure data and the ash load differential pressure corresponding to the ash load in the DPF; and determining the carbon load in the vehicle DPF according to the carbon load prediction model and the carbon load differential pressure. The method of the application can consider the current working condition and the current differential pressure when determining the carbon load, is more accurate, and can exclude the interference of the ash load when determining the carbon load, thereby further improving the accuracy of the determined carbon load.

Description

Technical Field

[0001] This application relates to the field of engine aftertreatment technology, and in particular to a method, apparatus, computer equipment, and storage medium for determining particulate matter in a vehicle's DPF. Background Technology

[0002] With the development of the automotive industry and the increasing requirements for environmental protection, the country has set higher and higher requirements for vehicle emissions, and the limits on particulate matter in engine emissions have become more and more stringent. In order to effectively reduce the particulate matter content in engine exhaust, it is necessary to install a particulate filter (DPF) on the vehicle. The DPF is the most effective way to purify particulate matter in exhaust emissions. It can capture particulate matter in engine emissions, thereby effectively reducing particulate matter emissions.

[0003] However, as cars operate, more and more particulate matter accumulates in the DPF. Excessive accumulation of particulate matter in the DPF will affect its particulate filtration performance, so it is necessary to remove the particulate matter from the DPF regularly. The best time to remove particulate matter from the DPF depends on the particulate matter load in the DPF. Therefore, how to determine the particulate matter load in the DPF is a problem that needs to be solved. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, and storage medium for determining particulate matter in a vehicle DPF that can determine the amount of particulate matter in the DPF, in order to address the above-mentioned technical problems.

[0005] A method for determining particulate matter in a vehicle's Discrete Pressure Filter (DPF) includes: acquiring current operating condition data of the vehicle and current differential pressure data in the vehicle's DPF; determining a carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data, wherein the carbon loading prediction model includes a correspondence between carbon loading and differential pressure; determining a carbon loading differential pressure based on the current differential pressure data and the ash loading differential pressure corresponding to the ash loading in the DPF; and determining the carbon loading in the vehicle's DPF based on the carbon loading prediction model and the carbon loading differential pressure.

[0006] In one embodiment, the step of determining the carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data includes: acquiring multiple training sample data, wherein each training sample data includes multiple different operating condition data and multiple corresponding relationships between carbon loading and pressure difference under each operating condition data; constructing a neural network model; training the neural network model using the training sample data to obtain a first prediction model, wherein the first prediction model includes the correspondence between the operating condition data and the carbon loading-pressure difference relationship; and determining the correspondence between carbon loading and pressure difference corresponding to the current operating condition data based on the current operating condition data and the first prediction model, as the carbon loading prediction model corresponding to the current operating condition data.

[0007] In one embodiment, the step of constructing a neural network model includes: constructing multiple neural network models with different architectures; the step of training the neural network models using the training sample data to obtain a first prediction model includes: training each neural network model using the training sample data to obtain multiple sub-prediction models; determining a weight model based on each sub-prediction model and the training sample data, wherein the weight model includes the weights corresponding to each sub-prediction model; and determining the first prediction model based on the weight model and the multiple sub-prediction models.

[0008] In one embodiment, the step of determining the weight model based on each sub-prediction model and the training sample data includes: adjusting the weights corresponding to each sub-prediction model until the weights corresponding to each sub-prediction model and the data obtained by each sub-prediction model are consistent with the training sample data, and using the current weights corresponding to each sub-prediction model as the weight model.

[0009] In one embodiment, the step of determining the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference includes: when the carbon load prediction model includes a pressure difference that is the same as the carbon load pressure difference, substituting the carbon load pressure difference into the carbon load prediction model to obtain the carbon load; when the carbon load prediction model does not include a pressure difference that is the same as the carbon load pressure difference, performing linear interpolation based on the correspondence between the carbon load pressure difference and the two sets of carbon loads and pressure differences that are closest to the carbon load pressure difference in the carbon load prediction model to obtain the carbon load.

[0010] In one embodiment, the method for determining particulate matter in a vehicle's DPF further includes: acquiring first operating condition data and first differential pressure data of the vehicle within a set time period after removing carbon from the DPF; determining a first ash load prediction model corresponding to the first operating condition data based on the first operating condition data, wherein the first ash load prediction model includes a correspondence between ash load and differential pressure; and determining the ash load based on the first differential pressure data and the first ash load prediction model.

[0011] In one embodiment, the method for determining particulate matter in a vehicle's DPF further includes: determining a second ash load prediction model corresponding to the current operating condition data based on the current operating condition data; and determining the ash load pressure difference based on the ash load and the second ash load prediction model.

[0012] An apparatus for determining particulate matter in a vehicle's DPF, comprising:

[0013] The data acquisition module is used to acquire the vehicle's current operating condition data and the current differential pressure data in the vehicle's DPF;

[0014] The model determination module is used to determine a carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data, wherein the carbon loading prediction model includes the correspondence between carbon loading and pressure difference;

[0015] The differential pressure determination module is used to determine the carbon loading differential pressure based on the current differential pressure data and the ash loading differential pressure corresponding to the ash loading in the DPF.

[0016] The carbon load determination module is used to determine the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference.

[0017] A computer device includes a memory and a processor, the memory storing a computer program, characterized in that the processor executes the computer program to implement the steps of the aforementioned method.

[0018] A computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the aforementioned method.

[0019] The above-mentioned method, apparatus, computer equipment, and storage medium for determining particulate matter in a vehicle's DPF (Diesel Particulate Filter). The method for determining particulate matter in a vehicle's DPF first acquires the vehicle's current operating condition data and the current pressure difference data in the DPF, thus obtaining the vehicle's current actual operating condition and the current actual pressure difference in the DPF. Then, based on the current operating condition data, a carbon load prediction model corresponding to the current operating condition data is determined, thus obtaining the correspondence between carbon load and pressure difference corresponding to the vehicle's current operating condition. This takes into account the vehicle's current operating condition, facilitating the improvement of the accuracy of the subsequently obtained carbon load. Since carbon and ash are present in the DPF, and carbon and ash are removed using different methods, the interference of ash load needs to be excluded when calculating the carbon load. Therefore, the current pressure difference data is subtracted from the ash load pressure difference corresponding to the ash load in the DPF; the remaining value is the carbon load pressure difference. Then, based on the carbon load pressure difference and the carbon load prediction model, the carbon load in the DPF can be obtained. The method of this application takes into account the current operating conditions and current pressure difference of the vehicle when determining the carbon load, thus making it more accurate. On the other hand, it eliminates the interference of ash load when determining the carbon load, thereby further improving the accuracy of the determined carbon load. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a method for determining particulate matter in a vehicle's DPF in one embodiment;

[0022] Figure 2 A flowchart of a method for determining a carbon loading prediction model in one embodiment;

[0023] Figure 3 This is a flowchart of a method for determining a first prediction model in one embodiment;

[0024] Figure 4 This is a flowchart of a method for determining carbon loading in one embodiment;

[0025] Figure 5 This is a flowchart of a method for determining the ash load in a vehicle's DPF in one embodiment;

[0026] Figure 6 This is a flowchart of a method for determining the ash load pressure difference in a vehicle's DPF in one embodiment;

[0027] Figure 7This is a structural diagram of a device for determining particulate matter in a vehicle's DPF in one embodiment;

[0028] Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0029] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0031] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another.

[0032] It should be noted that when one element is considered to be "connected" to another element, it can be directly connected to the other element or connected to the other element through an intermediary element. Furthermore, in the following embodiments, "connection" should be understood as "electrical connection," "communication connection," etc., if there is transmission of electrical signals or data between the connected objects.

[0033] When used herein, the singular forms of “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” or “having,” etc., specify the presence of the stated feature, whole, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0034] The method for determining particulate matter in a vehicle's DPF provided in this application embodiment can be applied to the vehicle's overall controller, facilitating users in determining the particulate matter content in the vehicle's DPF. Particulate matter includes at least one of carbon and ash. A DPF is a wall-flow filter device installed at the vehicle's exhaust outlet. During vehicle operation, exhaust gas enters the filter chamber within the DPF through the inlet end and reaches the DPF outlet end after passing through the filter chamber wall. During this process, carbon soot and ash particles in the exhaust gas are blocked by the filter chamber wall, achieving the effect of reducing carbon soot emissions. As the vehicle operates, the accumulated carbon soot and ash in the DPF gradually increase. When the DPF contents exceed a certain limit, it can cause blockage, leading to engine and aftertreatment system malfunctions. Therefore, the DPF needs to be cleaned regularly.

[0035] For carbon soot removal, an active regeneration method is used: by consuming additional fuel to raise the exhaust gas temperature, the DPF is heated to over 480°C, causing the accumulated carbon soot inside to oxidize at high temperatures, achieving a cleaning effect. If active regeneration is performed too early, the carbon load in the DPF is insufficient, thus wasting fuel. Conversely, if active regeneration is performed too late, the carbon load in the DPF is too high, leading to excessive carbon soot oxidation and heat release, potentially burning out the DPF. Therefore, it is necessary to determine the carbon load in the DPF to determine the appropriate regeneration timing, thereby improving vehicle fuel economy and safety.

[0036] In one embodiment, such as Figure 1 As shown, a method for determining particulate matter in a vehicle's DPF is provided, including:

[0037] Step S100: Obtain the vehicle's current operating condition data and the current differential pressure data in the vehicle's DPF.

[0038] Specifically, the vehicle's overall controller can obtain the vehicle's current operating condition data through the Controller Area Network (CAN) bus and obtain the current differential pressure data in the DPF through the pressure sensor.

[0039] The current differential pressure data in the DPF is the pressure difference between the inlet and outlet of the DPF.

[0040] For example, vehicle operating data may include engine intake air mass flow rate, selective catalytic reduction (SCR) inlet temperature, fuel flow rate, the product of engine intake air mass flow rate and SCR inlet temperature, and the ratio of fuel flow rate to intake air mass flow rate. These data will affect the DPF differential pressure data to some extent. The carbon loading and ash loading in the DPF will also affect the current differential pressure data in the DPF.

[0041] Step S110: Based on the current operating condition data, determine the carbon loading prediction model corresponding to the current operating condition data.

[0042] The carbon load prediction model includes the correlation between carbon load and pressure difference. Under different operating conditions, the correlation between carbon load and pressure difference in the DPF will differ. The carbon load prediction model corresponding to the current operating condition data can characterize the correlation between carbon load and pressure difference in the DPF under the current vehicle operating conditions. The carbon load prediction model can be in tabular form, recording multiple sets of data on the correlation between carbon load and pressure difference.

[0043] Step S120: Determine the carbon loading pressure difference based on the current pressure difference data and the ash loading pressure difference corresponding to the ash loading in the DPF.

[0044] Specifically, the current pressure difference data in the DPF is jointly reflected by the pressure difference caused by the carbon loading and the ash loading in the DPF. In other words, part of the pressure difference in the DPF is caused by the carbon accumulated in the DPF, and another part is caused by the ash accumulated in the DPF. Therefore, the current pressure difference in the DPF needs to be subtracted from the ash loading pressure difference corresponding to the ash loading in the DPF. The remaining value is the pressure difference caused by the accumulation of carbon in the DPF, i.e., the carbon loading pressure difference.

[0045] Step S130: Determine the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference.

[0046] Specifically, the carbon loading prediction model includes the correspondence between carbon loading and pressure difference. Therefore, after obtaining the carbon loading pressure difference, the carbon loading in the corresponding DPF can be obtained by combining the carbon loading pressure difference with the carbon loading prediction model.

[0047] In this embodiment, the current operating condition data of the vehicle and the current pressure difference data in the DPF are first obtained to obtain the actual operating condition of the vehicle and the actual pressure difference in the DPF. Then, based on the current operating condition data, a carbon load prediction model corresponding to the current operating condition data is determined, thus obtaining the correspondence between the carbon load and pressure difference corresponding to the current operating condition of the vehicle. This takes into account the current operating condition of the vehicle, which helps to improve the accuracy of the carbon load obtained subsequently. Since carbon and ash are present in the DPF, and carbon and ash are removed in different ways, the interference of ash load needs to be excluded when calculating the carbon load. Therefore, the current pressure difference data is subtracted from the ash load pressure difference corresponding to the ash load in the DPF, and the remaining amount is the carbon load pressure difference. Then, based on the carbon load pressure difference and the carbon load prediction model, the carbon load in the DPF can be obtained. Using the method of this application, on the one hand, the current operating condition and current pressure difference of the vehicle are considered when determining the carbon load, so it is more accurate; on the other hand, the interference of ash load is excluded when determining the carbon load, thus further improving the accuracy of the determined carbon load.

[0048] In one embodiment, such as Figure 2 As shown, step S110 involves determining a carbon loading prediction model corresponding to the current operating condition data, based on the current operating condition data. This includes:

[0049] Step S200: Obtain multiple training sample data.

[0050] The training sample data includes multiple different operating conditions and corresponding carbon loading and pressure differentials for each condition. Since the relationship between carbon loading and pressure differential in the DPF differs under different operating conditions, a large number of these relationships are obtained as training sample data. This training sample data can be downloaded from a cloud-based vehicle network. The training sample data includes multiple different operating conditions, each with multiple different carbon loading and pressure differential relationships; that is, multiple sets of different carbon loadings and corresponding pressure differentials under a given operating condition. The operating condition data can be represented in the form of a feature matrix.

[0051] Step S210: Construct a neural network model.

[0052] Specifically, the neural network model can be a time-series based neural network model, such as a Long Short-Term Memory (LSTM) network model, a Convolutional Neural Network (CNN) model, or a Temporal Convolutional Network (TCN) model. For example, an LSTM model can be designed with 2 layers and 64 nodes; a CNN model can be designed with 10 channels, a kernel size of 3, and an average pooling range of 2; and a TCN model can be designed with 15 hidden nodes, 9 layers, and a kernel size of 2. This allows for the construction of neural network models.

[0053] Step S220: Train the neural network model using training sample data to obtain the first prediction model.

[0054] The first prediction model includes the correspondence between operating condition data and carbon loading-pressure difference.

[0055] Specifically, the training sample data is divided into a training set and a test set. 70% of the data can be used as the training set and 30% as the test set. Then, the training parameters for the neural network model are set, including the number of training iterations, batch size, learning rate, etc. The neural network model is then trained using the training set data. After training, the accuracy of the neural network model is tested using the test set data. A mean absolute error threshold and a coefficient of determination threshold are set. If the model's accuracy is higher than the threshold, the model is usable; otherwise, the model's design and training parameters are adjusted, and training is repeated until the accuracy reaches the threshold.

[0056] Step S230: Based on the current operating condition data and the first prediction model, determine the correspondence between the carbon loading and pressure difference corresponding to the current operating condition data, and use it as the carbon loading prediction model corresponding to the current operating condition data.

[0057] Specifically, after acquiring the current operating condition data, it is necessary to perform data cleaning according to preset conditions to remove abnormal data and ensure the accuracy of the operating condition data. For example, using 100 kg / h and 2000 kg / h as the upper and lower limits of intake airflow, operating condition data with intake airflow outside these ranges are removed; using 100℃ and 400℃ as the upper and lower limits of the selective catalytic reduction system inlet temperature, operating condition data with inlet temperature outside these ranges are removed; using 0 L / h and 300 L / h as the upper and lower limits of fuel flow, operating condition data with fuel flow outside these ranges are removed; and using 0 kPa and 35 kPa as the upper and lower limits of DPF pressure difference, DPF pressure difference outside these ranges is removed. Then, the filtered current operating condition data is combined with the first prediction model to obtain the carbon loading prediction model corresponding to the current operating condition data.

[0058] For example, the current operating condition data can also be operating condition data over a period of time. The time series length is set to 5 seconds. Data segments of 5 seconds in length are successively extracted from the operating condition data collected at one end of the time period. The starting timestamp values ​​of different data segments differ by 1 second. For example, the timestamp of the first time segment is 1s-5s, and the timestamp of the second time segment is 2s-6s. Within each data segment, the timestamps are continuous, and the data is arranged sequentially according to the timestamps. The pressure difference value corresponding to the operating condition data in the last second of each 5-second data segment is taken as the current pressure difference data. This yields several sets of current operating condition data - current pressure difference data, facilitating the subsequent calculation of multiple carbon loadings within this time period.

[0059] In this embodiment, a large amount of training sample data is first used to train the constructed neural network model to obtain a first prediction model that can predict the correspondence between carbon loading and pressure difference based on the operating condition data. Then, the current operating condition data is combined with the first prediction model to obtain the carbon loading prediction model corresponding to the current operating condition data, which facilitates the subsequent determination of carbon loading.

[0060] In one embodiment, such as Figure 3 As shown, step S210 involves constructing a neural network model. This includes:

[0061] Step S300: Construct multiple neural network models with different architectures.

[0062] For example, Long Short-Term Memory (LSTM) models, Convolutional Neural Networks (CNN) models, and Temporal Convolutional Networks (TCN) models can be constructed.

[0063] Step S220: Train the neural network model using training sample data to obtain the first prediction model. This includes:

[0064] Step S310: Train each neural network model using training sample data to obtain multiple sub-prediction models.

[0065] For example, the training loops for an LSTM model can be designed with 300 iterations, a batch size of 32, and a learning rate of 0.0001; the training loops for a CNN model can be designed with 1000 iterations, a batch size of 1000, and a learning rate of 0.0001; and the training loops for a TCN model can be designed with 10 iterations, a batch size of 32, and a learning rate of 0.00001. Each of the three models is trained using the training set data. After training, the accuracy of the three models is tested using the test set data. A mean absolute error threshold of 0.1 and a coefficient of determination threshold of 0.95 are set until the accuracy of all three models is higher than the threshold, resulting in three sub-prediction models.

[0066] Step S320: Determine the weight model based on each sub-prediction model and the training sample data.

[0067] The weight model includes the weights corresponding to each sub-prediction model.

[0068] Specifically, an initial weighted model can be constructed first. Then, operating condition data from multiple training sets are substituted into each sub-prediction model to obtain the pressure drop-carbon loading correspondence sequence output by each sub-prediction model. The weighted model is then used to weight these sequences, with the principle that the sum of the weights of each sub-prediction model is 1, and the resulting pressure drop-carbon loading correspondence sequence matches the actual pressure drop-carbon loading correspondence sequence in the training sample data. The weighted model is then trained. The weights of each sub-prediction model are adjusted until the weights and data obtained by each sub-prediction model match the training sample data. The current weights of each sub-prediction model are then used as the final weighted model.

[0069] For example, a Deep Temporal Clustering (DTC) architecture can be used to construct a weighted model. The DTC model is designed with 50 channels, a kernel size of 3, an average pooling range of 2, and 2 nodes. The working conditions data are used as feature values, and the weight data corresponding to each sub-model is used as the target value. The training and testing sets are divided with a 70% training set and a 30% testing set, yielding the training and testing data for the weighted model. The training parameters of the weighted model are set, including the number of training iterations, batch size, and learning rate. The weighted model is then trained using the training set data. After training, the accuracy of the weighted model is tested. A log loss threshold is set. If the model accuracy is higher than the threshold, the model is usable; otherwise, the design and training parameters are adjusted, and retraining is performed until the accuracy reaches the threshold. A possible setup is 200 training iterations, a batch size of 32, a learning rate of 0.0001, and a log loss threshold of 0.3.

[0070] Step S330: Determine the first prediction model based on the weight model and multiple sub-prediction models.

[0071] Specifically, after obtaining the weighted model, the predicted values ​​of each sub-prediction model are weighted using the weighted model, and the final predicted value is the predicted value of the first prediction model. That is, the first prediction model can be obtained by combining the weighted model and multiple sub-prediction models.

[0072] In this embodiment, multiple neural network models based on different architectural principles are constructed and trained to obtain multiple sub-prediction models. An independent weight model is also constructed. The first prediction model is obtained by combining the weight model and multiple sub-prediction models. Since the first prediction model is composed of multiple sub-prediction models based on different principles, the robustness of the prediction algorithm is improved, the impact of the inaccuracy of a single sub-prediction model at individual working conditions on the final prediction result is reduced, and the accuracy of the result obtained by using the first prediction model is further improved.

[0073] In one embodiment, such as Figure 4 As shown, step S130 involves determining the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference. This includes:

[0074] Step S400: When the carbon loading prediction model includes a pressure difference that is the same as the carbon loading pressure difference, substitute the carbon loading pressure difference into the carbon loading prediction model to obtain the carbon loading.

[0075] Specifically, the carbon loading prediction model includes multiple sets of carbon loading-pressure difference correspondences under the current operating conditions, as shown in Table 1 below. Therefore, if the current measured carbon loading pressure difference data is already recorded in the carbon loading prediction model, for example, if the measured carbon loading pressure difference data is 5 kPa and the carbon loading prediction model also includes data with a carbon loading pressure difference of 5 kPa, then the carbon loading can be directly obtained. For example, using the model in Table 1, the carbon loading is 5 g.

[0076] Table 1. Correspondence between carbon loading and pressure difference.

[0077]

[0078] Step S410: When the carbon loading prediction model does not include a pressure difference that is the same as the carbon loading pressure difference, linear interpolation is performed based on the correspondence between the carbon loading pressure difference and the two sets of carbon loading and pressure difference that are closest to the carbon loading pressure difference in the carbon loading prediction model to obtain the carbon loading.

[0079] Specifically, if the measured carbon loading pressure difference data is not recorded in the carbon loading prediction model—for example, if the measured carbon loading pressure difference is 8 kPa—then the carbon loading prediction model does not record the corresponding carbon loading value for an 8 kPa pressure difference. In this case, the two sets of carbon loading and pressure difference correspondences that are closest to the carbon loading pressure difference in the carbon loading prediction model are selected. For example, using the model shown in Table 1, two sets of data with pressure differences of 5 kPa and 10 kPa are selected, as they are the values ​​closest to 8 kPa. Then, linear interpolation is performed. The formula for linear interpolation is as follows:

[0080]

[0081] Where a1 represents carbon loading, P1 represents the carbon loading pressure difference, P0 represents the pressure difference data in the carbon loading prediction model that is closest to but smaller than the carbon loading pressure difference, a0 represents the carbon loading value corresponding to P0 in the carbon loading prediction model, P2 represents the pressure difference data in the carbon loading prediction model that is closest to but larger than the carbon loading pressure difference, and a2 represents the carbon loading value corresponding to P2 in the carbon loading prediction model. For example, using the model shown in Table 1, P1 is 8 kPa, P0 is 5 kPa, P2 is 10 kPa, a0 is 5 g, and a2 is 10 g.

[0082] In this embodiment, the carbon loading is calculated by substituting the measured carbon loading pressure difference into the carbon loading prediction model. The carbon loading prediction model can include only the pressure difference-carbon loading data of multiple key points. For the measured carbon loading pressure difference values ​​not included in the carbon loading prediction model, the corresponding carbon loading can be calculated by linear interpolation. Therefore, when training the carbon loading prediction model, only the data of key points needs to be trained, resulting in a smaller amount of training data and saving training time.

[0083] In one embodiment, such as Figure 5 As shown, the method for determining particulate matter in a vehicle's DPF also includes:

[0084] Step S500: Within a set time after the carbon in the DPF is removed, acquire the vehicle's first operating condition data and first differential pressure data.

[0085] Specifically, the accumulation rate of ash in a DPF is very slow, much slower than the accumulation rate of carbon. Therefore, the ash in a DPF does not need to be cleaned very frequently. After the carbon in the DPF is cleaned, it can be approximated that there is no carbon in the DPF, only ash. Therefore, the pressure difference of the DPF during this period is caused by ash. Since the accumulation rate of ash is particularly slow, the calculated ash load during this period can be approximated as a constant value. Until the next time the carbon in the DPF is cleaned, the ash load in the DPF can be considered to be this constant value.

[0086] Step S510: Based on the first working condition data, determine the first ash load prediction model corresponding to the first working condition data.

[0087] The first ash load prediction model includes the correspondence between ash load and pressure difference.

[0088] Specifically, the same method as in the above embodiments can be used, employing a large amount of different operating condition data and the corresponding relationship between ash load and pressure difference under each operating condition data as training samples to train the constructed neural network model, thereby obtaining a second prediction model for the relationship between ash load and pressure difference corresponding to the operating condition data. Substituting the first operating condition data into the second prediction model yields a first ash load prediction model corresponding to the first operating condition data.

[0089] Step S520: Determine the ash load based on the first differential pressure data and the first ash load prediction model.

[0090] Specifically, the ash load is obtained by substituting the measured first pressure difference data within a set time after carbon removal from the DPF into the first ash load prediction model. This ash load can be approximated as a constant value before the vehicle performs the next carbon removal from the DPF.

[0091] The instruction manual states that for ash, the service station provides disassembly and cleaning services for cleaning. Cleaning too early will increase service costs, while cleaning too late will reduce the DPF's carbon soot capacity, leading to frequent regeneration, DPF blockage, and other problems. Therefore, it is necessary to calculate the ash load in the DPF in order to determine the appropriate time for cleaning.

[0092] In this embodiment, a set time is used after the carbon in the DPF is removed, during which time only ash remains in the DPF, so the ash load in the DPF can be calculated. Then, taking advantage of the slow accumulation rate of ash in the DPF, the calculated ash load is taken as a constant value, thus calculating the ash load in the DPF.

[0093] In one embodiment, such as Figure 6 As shown, the method for determining particulate matter in a vehicle's DPF also includes:

[0094] Step S600: Based on the current operating condition data, determine the second ash load prediction model corresponding to the current operating condition data.

[0095] Specifically, based on the vehicle's current operating condition data and combined with the second prediction model in the previous embodiment, a second ash load prediction model corresponding to the vehicle's current operating condition can be obtained, which can characterize the relationship between the pressure difference and ash load in the DPF under the current operating condition.

[0096] Step S610: Determine the ash load pressure difference based on the ash load and the second ash load prediction model.

[0097] Specifically, as shown in the previous embodiment, the ash load can be approximated as a constant value before the vehicle performs the next carbon removal from the DPF, thus the ash load value has already been calculated in the previous embodiment. However, as the vehicle's operating conditions change, even if the ash load value remains constant, the pressure difference in the DPF caused by the ash load will change. Therefore, by substituting the ash load into the second ash load prediction model, the ash load pressure difference corresponding to the ash load value under the current operating conditions can be determined. This determines the ash load pressure difference. This ash load pressure difference can be used in the aforementioned embodiments to determine the carbon load pressure difference based on the current pressure difference data and the ash load pressure difference in the DPF, and can correct the current carbon load pressure difference to obtain a more accurate carbon load.

[0098] In this embodiment, the ash load is approximated as a constant value by taking advantage of the slow accumulation rate of ash in the DPF. However, as the vehicle's operating conditions change, the ash load differential pressure of the DPF will change with the same ash load. Thus, the ash load differential pressure can be obtained based on the current operating conditions and ash load, so as to correct the carbon load differential pressure and obtain a more accurate carbon load.

[0099] It should be understood that, although Figures 1-6 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 1-6 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0100] In one embodiment, such as Figure 7 As shown, a device for determining particulate matter in a vehicle's DPF is provided, comprising: a data acquisition module 701, a model determination module 702, a differential pressure determination module 703, and a carbon load determination module 704, wherein:

[0101] The data acquisition module 701 is used to acquire the current operating condition data of the vehicle and the current differential pressure data in the vehicle's DPF.

[0102] The model determination module 702 is used to determine the carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data. The carbon loading prediction model includes the correspondence between carbon loading and pressure difference.

[0103] The differential pressure determination module 703 is used to determine the carbon loading differential pressure based on the current differential pressure data and the ash loading differential pressure corresponding to the ash loading in the DPF.

[0104] The carbon load determination module 704 is used to determine the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference.

[0105] In one embodiment, the model determination module 702 further includes: a data acquisition unit, a model building unit, a model training unit, and a prediction unit. Wherein:

[0106] The data acquisition unit is used to acquire multiple training sample data, wherein each training sample data includes multiple different operating condition data and the corresponding relationship between multiple carbon loading and pressure difference under each operating condition data.

[0107] Model building unit, used to build neural network models.

[0108] The model training unit is used to train a neural network model using training sample data to obtain a first prediction model, wherein the first prediction model includes the correspondence between operating condition data and carbon loading-pressure difference.

[0109] The prediction unit is used to determine the correspondence between the carbon loading and pressure difference corresponding to the current operating condition data based on the current operating condition data and the first prediction model, and to serve as the carbon loading prediction model corresponding to the current operating condition data.

[0110] In one embodiment, the carbon loading determination module 704 further includes: a first carbon loading determination unit and a second carbon loading determination unit, wherein:

[0111] The first carbon loading determination unit is used to substitute the carbon loading pressure difference into the carbon loading prediction model when the carbon loading prediction model includes the same pressure difference as the carbon loading pressure difference, so as to obtain the carbon loading.

[0112] The second carbon loading determination unit is used to perform linear interpolation based on the correspondence between the carbon loading pressure difference and the two sets of carbon loading and pressure difference that are closest to the carbon loading pressure difference in the carbon loading prediction model when the carbon loading prediction model does not include the pressure difference that is the same as the carbon loading pressure difference, so as to obtain the carbon loading.

[0113] Specific limitations regarding the device for determining particulate matter in a vehicle's DPF can be found in the above-described limitations on the method for determining particulate matter in a vehicle's DPF, and will not be repeated here. Each module in the aforementioned device for determining particulate matter in a vehicle's DPF can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. It should be noted that the module division in this embodiment is illustrative and only represents a logical functional division; other division methods may be used in actual implementation.

[0114] In one embodiment, a computer device is provided, the internal structure of which can be shown in the following diagram. Figure 8 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for determining particulate matter in a vehicle's DPF (Device Power Distribution System).

[0115] Those skilled in the art will understand that Figure 8The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0116] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0117] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0118] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0119] In the description of this specification, references to terms such as "some embodiments," "other embodiments," and "ideal embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative descriptions of the above terms do not necessarily refer to the same embodiments or examples.

[0120] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for determining particulate matter in a vehicle's DPF, characterized in that, include: Obtain the vehicle's current operating condition data and the current differential pressure data in the vehicle's DPF; Based on the current operating condition data, a carbon loading prediction model corresponding to the current operating condition data is determined, wherein the carbon loading prediction model includes the correspondence between carbon loading and pressure difference; The carbon loading pressure difference is determined based on the current pressure difference data and the ash loading pressure difference corresponding to the ash loading in the DPF. The carbon load in the vehicle's DPF is determined based on the carbon load prediction model and the carbon load pressure difference. The step of determining the carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data includes: Acquire multiple training sample data, wherein each training sample data includes multiple different operating condition data and the corresponding relationship between multiple carbon loading and pressure difference under each operating condition data; Construct multiple neural network models with different architectures; Each neural network model is trained using the aforementioned training sample data to obtain multiple sub-prediction models; An initial weight model is constructed, and the training sample data is substituted into each sub-prediction model to obtain the corresponding sequence of pressure difference-carbon loading output by each sub-prediction model. The pressure difference-carbon loading correspondence sequence output by each sub-prediction model is weighted based on the initial weight model. The initial weight model is then trained based on the weighted pressure difference-carbon loading correspondence sequence and the actual pressure difference-carbon loading correspondence sequence in the training sample data to obtain the trained weight model. The sum of the weights of each sub-prediction model in the weight model is 1. Based on the trained weighted model and multiple sub-prediction models, a first prediction model is determined; wherein, the first prediction model includes the correspondence between operating condition data and carbon loading-pressure difference. Based on the current operating condition data and the first prediction model, the correspondence between the carbon loading and pressure difference corresponding to the current operating condition data is determined, which serves as the carbon loading prediction model corresponding to the current operating condition data.

2. The method for determining particulate matter in a vehicle DPF according to claim 1, characterized in that, The step of determining the weight model based on each sub-prediction model and the training sample data includes: Adjust the weights corresponding to each sub-prediction model until the weights corresponding to each sub-prediction model and the data obtained by each sub-prediction model are consistent with the training sample data. Then, use the weights corresponding to each sub-prediction model as the weight model.

3. The method for determining particulate matter in a vehicle DPF according to any one of claims 1-2, characterized in that, The step of determining the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference includes: When the carbon loading prediction model includes a pressure difference that is the same as the carbon loading pressure difference, the carbon loading pressure difference is substituted into the carbon loading prediction model to obtain the carbon loading. When the carbon loading prediction model does not include a pressure difference that is the same as the carbon loading pressure difference, linear interpolation is performed based on the correspondence between the carbon loading pressure difference and the two sets of carbon loading and pressure differences that are closest to the carbon loading pressure difference in the carbon loading prediction model to obtain the carbon loading.

4. The method for determining particulate matter in a vehicle DPF according to any one of claims 1-2, characterized in that, The method further includes: Within a set time after the carbon in the DPF is removed, acquire the vehicle's first operating condition data and first differential pressure data; Based on the first working condition data, a first ash load prediction model corresponding to the first working condition data is determined, wherein the first ash load prediction model includes the correspondence between ash load and pressure difference; The ash load is determined based on the first differential pressure data and the first ash load prediction model.

5. The method for determining particulate matter in a vehicle DPF according to claim 4, characterized in that, The method further includes: Based on the current operating condition data, determine the second ash load prediction model corresponding to the current operating condition data; The ash load pressure difference is determined based on the ash load and the second ash load prediction model.

6. A device for determining particulate matter in a vehicle's DPF, characterized in that, include: The data acquisition module is used to acquire the vehicle's current operating condition data and the current differential pressure data in the vehicle's DPF; The model determination module is used to determine a carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data, wherein the carbon loading prediction model includes the correspondence between carbon loading and pressure difference; The differential pressure determination module is used to determine the carbon loading differential pressure based on the current differential pressure data and the ash loading differential pressure corresponding to the ash loading in the DPF. A carbon load determination module is used to determine the carbon load in the vehicle's DPF based on the carbon load prediction model and the carbon load pressure difference. The step of determining the carbon loading prediction model corresponding to the current operating condition data based on the current operating condition data includes: Acquire multiple training sample data, wherein each training sample data includes multiple different operating condition data and the corresponding relationship between multiple carbon loading and pressure difference under each operating condition data; Construct multiple neural network models with different architectures; Each neural network model is trained using the aforementioned training sample data to obtain multiple sub-prediction models; An initial weight model is constructed, and the training sample data is substituted into each sub-prediction model to obtain the corresponding sequence of pressure difference-carbon loading output by each sub-prediction model. The pressure difference-carbon loading correspondence sequence output by each sub-prediction model is weighted based on the initial weight model. The initial weight model is then trained based on the weighted pressure difference-carbon loading correspondence sequence and the actual pressure difference-carbon loading correspondence sequence in the training sample data to obtain the trained weight model. The sum of the weights of each sub-prediction model in the weight model is 1. Based on the trained weighted model and multiple sub-prediction models, a first prediction model is determined; wherein, the first prediction model includes the correspondence between operating condition data and carbon loading-pressure difference. Based on the current operating condition data and the first prediction model, the correspondence between the carbon loading and pressure difference corresponding to the current operating condition data is determined, which serves as the carbon loading prediction model corresponding to the current operating condition data.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images