An engine fault diagnosis optimization method based on working condition portrait and related device

Abstract

The present application discloses an engine fault diagnosis optimization method based on working condition portrait and related devices, the method comprising: obtaining the operating parameters of the engine, the operating parameters comprising at least a plurality of working condition parameters; obtaining the current working condition portrait of the engine according to the operating parameters; processing the current working condition portrait and the initial data version using a diagnosis prediction model to obtain a target data version corresponding to the current working condition portrait; the target data version comprising: a fault diagnosis strategy and a diagnosis support parameter corresponding to the fault diagnosis strategy; the diagnosis prediction model is obtained by training using training samples, the training samples are obtained using a typical working condition library, the typical working condition library comprising a plurality of typical working condition portraits; the target data version is sent to the engine controller, and the controller executes the fault diagnosis strategy according to the diagnosis support parameters, so that the use time of the engine before the fault alarm is extended.

Description

Technical Field

[0001] The present application relates to the field of engine technology, and in particular to an engine fault diagnosis optimization method based on operating condition profiling and related devices. Background Art

[0002] At present, engine fault diagnosis usually uses existing road spectrum data and accumulated experience to diagnose and predict engine faults.

[0003] For engine parts, they are replaced or maintained when a possible failure is predicted. However, in actual applications, the life and maintenance cycle of actual parts can be extended due to different operating conditions. For this reason, engineers usually pre-calibrate fault diagnosis strategies and diagnostic support parameters under various operating conditions.

[0004] However, in actual engine applications, operating conditions are ever-changing. Even if engineers have pre-calibrated multiple versions of fault diagnosis strategies and diagnostic support parameters, there are still certain operating conditions under which the life and maintenance cycle of engine components cannot be extended. Summary of the invention

[0005] In view of the above problems, the present application provides an engine fault diagnosis optimization method based on working condition profiling and related devices, so as to extend the engine usage time before the fault alarm. The specific scheme is as follows:

[0006] The first aspect of the present application provides an engine fault diagnosis optimization method based on operating condition profiling, the method comprising:

[0007] Obtaining operating parameters of the engine, wherein the operating parameters include at least a plurality of operating condition parameters;

[0008] According to the operating parameters, a current operating condition profile of the engine is obtained;

[0009] The current operating condition profile and the initial data version are processed by using a diagnostic prediction model to obtain a target data version corresponding to the current operating condition profile; the target data version includes: a fault diagnosis strategy and a diagnostic support parameter corresponding to the fault diagnosis strategy;

[0010] The diagnostic prediction model is obtained by training using training samples, and the training samples are obtained using a typical operating condition library, and the typical operating condition library contains multiple typical operating condition portraits;

[0011] The target data version is sent to a controller of the engine, and the controller executes the fault diagnosis strategy according to the diagnosis support parameters, so that the service life of the engine before a fault alarm is extended.

[0012] In a possible implementation, the diagnosis prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a data optimization model; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter;

[0013] Wherein, using the diagnostic prediction model, processing is performed according to the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including:

[0014] According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library;

[0015] The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait;

[0016] Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine;

[0017] Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure;

[0018] According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree;

[0019] The target fault level, the target occurrence degree, the target risk degree, the current operating condition portrait and the initial data version are input into the data optimization model to obtain the target data version corresponding to the current operating condition portrait output by the data optimization model.

[0020] In a possible implementation, the diagnosis prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a control simulation system; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter;

[0021] Wherein, using the diagnostic prediction model, processing is performed according to the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including:

[0022] According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library;

[0023] The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait;

[0024] Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine;

[0025] Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure;

[0026] According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree;

[0027] Inputting the current operating condition portrait and the initial data version into a control simulation system corresponding to the controller of the engine, the control simulation system can output initial parameters according to the current operating condition portrait using the initial data version as the control data version; the initial parameters include: initial fault level, initial occurrence degree, and initial risk degree;

[0028] Adjusting the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the target parameters;

[0029] The control data version used by the control simulation system when the initial parameters match the target parameters is determined as the target data version corresponding to the current operating condition portrait.

[0030] In a possible implementation, the matching condition is: the portrait similarity between the current working condition portrait and the target working condition portrait is greater than or equal to a first threshold.

[0031] In a possible implementation, according to the current working condition portrait, obtaining a target working condition portrait and an initial data version in a data version library includes:

[0032] Using a clustering model, obtaining a current operating condition type corresponding to the current operating condition portrait;

[0033] Compare the current working condition type with the working condition type of each typical working condition portrait in the data version library;

[0034] If there is a typical operating condition type in the data version library that is consistent with the current operating condition type, determine the typical operating condition portrait in the data version library that is consistent with the current operating condition type as the target operating condition portrait and obtain the optimized data version corresponding to the target operating condition portrait;

[0035] If there is no typical operating condition type in the data version library that is consistent with the current operating condition type, the typical operating condition portrait with the greatest portrait similarity with the current operating condition portrait in the data version library is used as the target operating condition portrait and the optimized data version corresponding to the target operating condition portrait is obtained.

[0036] In a possible implementation, the training sample is obtained in the following manner:

[0037] The following processing is performed on each typical working condition image in the typical working condition library:

[0038] Using the fault level classification model, obtaining the current fault level corresponding to the typical working condition portrait;

[0039] Using the risk classification model, obtaining the current occurrence degree and current risk degree corresponding to the typical working condition portrait;

[0040] According to the current fault level, current occurrence degree and current risk degree corresponding to the typical working condition portrait, in the fault strategy pool, index parameters matching the typical working condition portrait are screened out; the index parameters include: fault level index, occurrence degree index and risk degree index;

[0041] The typical operating condition portrait and the historical data version corresponding to the typical operating condition portrait are input into the control simulation system corresponding to the controller of the engine, and the control simulation system can output initial parameters according to the typical operating condition portrait using the historical data version corresponding to the typical operating condition portrait as the control data version; the initial parameters include: initial fault level, initial occurrence degree, and initial risk degree;

[0042] Adjusting the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the indicator parameters;

[0043] Among them, the control data version used by the control simulation system when the initial parameters match the index parameters is the output sample in the training sample; the index parameters, the typical operating condition portrait and the historical data version corresponding to the typical operating condition portrait are the input samples in the training sample.

[0044] In a possible implementation, the initial parameter matches the indicator parameter, including:

[0045] The initial fault level matches the fault level index, the initial occurrence degree matches the occurrence degree index, and the initial risk degree matches the risk degree index.

[0046] In a possible implementation, after obtaining the target data version corresponding to the current operating condition portrait, the method further includes:

[0047] Generalizing at least one operating condition parameter in the current operating condition profile to obtain a new typical operating condition profile;

[0048] The new typical working condition portrait and its corresponding target data version are added to the data version library.

[0049] A second aspect of the present application provides an engine fault diagnosis and optimization device based on operating condition profiles, the device comprising:

[0050] A parameter obtaining unit, used to obtain operating parameters of the engine, wherein the operating parameters at least include a plurality of operating condition parameters;

[0051] A portrait obtaining unit, used for obtaining a current operating condition portrait of the engine according to the operating parameters;

[0052] A data processing unit, used to process the current operating condition portrait and the initial data version using a diagnostic prediction model to obtain a target data version corresponding to the current operating condition portrait; the target data version includes: a fault diagnosis strategy and a diagnostic support parameter corresponding to the fault diagnosis strategy;

[0053] The diagnostic prediction model is obtained by training using training samples, and the training samples are obtained using a typical operating condition library, and the typical operating condition library contains multiple typical operating condition portraits;

[0054] A data transmission unit is used to send the target data version to the controller of the engine, and the controller executes the fault diagnosis strategy according to the diagnosis support parameters to extend the service life of the engine before the fault alarm.

[0055] The third aspect of the present application provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements an engine fault diagnosis optimization method based on operating condition profiling according to the above-mentioned first aspect or any implementation method of the first aspect.

[0056] A fourth aspect of the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0057] The memory is used to store computer programs;

[0058] The processor is used to execute the computer program so that the electronic device can implement an engine fault diagnosis optimization method based on operating condition profiling according to the above-mentioned first aspect or any implementation method of the first aspect.

[0059] A fifth aspect of the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement an engine fault diagnosis optimization method based on operating condition profiling according to the above-mentioned first aspect or any implementation method of the first aspect.

[0060] By means of the above technical scheme, in an engine fault diagnosis optimization method based on working condition portrait and related devices provided by the present application, after obtaining the operating parameters of the engine including multiple working condition parameters, the current working condition portrait of the engine is first obtained according to the operating parameters, and then the diagnostic prediction model trained by the typical working condition portrait is used to process the current working condition portrait and the initial data version to obtain the target data version corresponding to the current working condition portrait, and the target data version includes: fault diagnosis strategy and diagnostic support parameters corresponding to the fault diagnosis strategy, so that after sending the target data version to the controller of the engine, the controller executes the corresponding fault diagnosis strategy according to the diagnostic support parameters in the target data version, thereby extending the use time of the engine before the fault alarm. It can be seen that in the present application, no matter what working condition state the current working condition portrait of the engine is, the diagnostic prediction model can be used to optimize the fault diagnosis strategy and the corresponding diagnostic support parameters, so that when the engine controller executes the fault diagnosis strategy according to the optimized diagnostic support parameters, the use time of the engine before the fault alarm can be extended, so that the cost of using the engine can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0062] Figure 1 A flowchart of an engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application;

[0063] Figure 2A partial flow chart of an engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application;

[0064] Figure 3 Another partial flow chart of an engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application;

[0065] Figure 4 A flowchart of another part of an engine fault diagnosis optimization method based on operating condition portrait provided in an embodiment of the present application;

[0066] Figure 5 Another flow chart of an engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application;

[0067] Figure 6 Another flow chart of an engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application;

[0068] Figure 7 A schematic diagram of the structure of an engine fault diagnosis and optimization device based on operating condition profiling provided in an embodiment of the present application;

[0069] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0070] Fig. 9 The overall execution logic flow chart of the engine fault automatic diagnosis prediction optimization method based on working condition adaptive learning in this application;

[0071] Fig.10 A flowchart for optimizing the fault diagnosis strategy in this application;

[0072] Fig.11 This is a flow chart for obtaining training samples for the virtual calibration model in this application. DETAILED DESCRIPTION

[0073] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0074] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0075] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0076] The present application can be applied in the field of engines. Taking diesel engines as an example, multiple application scenarios of the products will be introduced below.

[0077] Reference Figure 1 , is a flow chart of an engine fault diagnosis optimization method based on working condition profiling provided in an embodiment of the present application. The method can be applied to an electronic device capable of data processing, which can be a local device connected to the engine or a cloud device connected to the engine. The technical solution in this embodiment is mainly used to extend the use time of the engine before the fault alarm.

[0078] Specifically, the method in this embodiment may include the following steps:

[0079] Step 101: Obtain operating parameters of the engine.

[0080] Among them, the operating parameters include at least multiple operating parameters, such as rotation speed, torque, vehicle speed, altitude, temperature, load, acceleration, idle speed, etc.

[0081] In one implementation, in step 101 , multiple sensors deployed on the engine may be used to collect multiple operating parameters of the engine to obtain the operating parameters of the engine.

[0082] For example, a data connection is established between the electronic device in this embodiment and the sensor deployed on the engine. After the sensor collects the operating parameters, the operating parameters are sent to the electronic device in this embodiment through the data connection, thereby obtaining operating parameters including these operating parameters on the electronic device.

[0083] It should be noted that the operating parameters may also include non-operating parameters, which refer to parameters that will not change after the engine leaves the factory.

[0084] It should be noted that this embodiment can be executed when the operating parameters of the engine change, that is, this embodiment continuously detects whether the operating parameters of the engine have changed. If it is monitored that the operating parameters of the engine have changed, then step 101 and subsequent steps are executed.

[0085] Step 102: Obtain a current operating condition profile of the engine according to the operating parameters.

[0086] Specifically, in step 102, the operating parameters can be first filtered for data standardization by using rules such as data correlation and range limits to remove abnormal data; then, the operating parameters are segmented according to the selected data processing method (such as driving cycle mode, mileage, operating time, etc.), so that corresponding parameter features are extracted from the segmented operating parameters, such as temperature, altitude, humidity, speed, torque, load, etc., and then the current working condition profile of the engine is constructed based on these parameter features. The current working condition profile of the engine can be represented by the extracted parameter features, as shown in Table 1.

[0087] Table 1 Current working condition portrait

[0088]

[0089] Step 103: Use the diagnostic prediction model to process the current operating condition profile and the initial data version to obtain the target data version corresponding to the current operating condition profile.

[0090] The target data version includes the engine fault diagnosis strategy and the diagnostic support parameters corresponding to the fault diagnosis strategy. The diagnostic support parameters refer to the parameters required by the engine controller ECU (electronic control unit) to execute the fault diagnosis strategy. The fault diagnosis strategy refers to the strategy used by the controller to detect and alarm the engine fault. For example, when the oil level in the tank drops to a certain value, the controller outputs an alarm indicating that the oil level is too low.

[0091] It should be noted that the diagnosis prediction model is trained using training samples, and the training samples are obtained using a typical working condition library. The typical working condition library may contain multiple typical working condition portraits.

[0092] In the initial state, the typical operating condition portraits in the typical operating condition library can be manually set by the user. As the engine runs, the engine may generate various different operating condition portraits. Therefore, these actual operating condition portraits can also be added to the typical operating condition library.

[0093] In one implementation, a diagnostic prediction model can be constructed based on a machine learning algorithm and trained using training samples so that the diagnostic prediction model can learn the relationship between input samples and output samples in the training samples. The input samples at least include an operating condition portrait and a data version (such as a fault diagnosis strategy and diagnostic support parameters), and the output samples are optimized data versions corresponding to the operating condition portrait in the input samples. In this way, the diagnostic prediction model can process the current operating condition portrait and the initial data version to obtain the target data version corresponding to the current operating condition portrait.

[0094] Step 104: Send the target data version to the controller of the engine.

[0095] The controller executes the fault diagnosis strategy according to the diagnostic support parameters in the target data version, so that the engine usage time before the fault alarm is extended.

[0096] Specifically, in step 104, the target data version can be sent to the engine controller via OTA (Over The Air), so that when the controller executes the fault diagnosis strategy according to the diagnostic support parameters in the target data version, the use time of the engine before the fault alarm can be extended, specifically: the engine runs from a first time length to a second time length and then outputs a prompt message, so as to facilitate maintenance personnel to maintain, service or replace engine-related parts.

[0097] By means of the above technical scheme, in an engine fault diagnosis optimization method based on working condition portrait provided by the embodiment of the present application, after obtaining the operating parameters of the engine including multiple working condition parameters, the current working condition portrait of the engine is first obtained according to the operating parameters, and then the diagnostic prediction model trained by the typical working condition portrait is used to process the current working condition portrait and the initial data version to obtain the target data version corresponding to the current working condition portrait, and the target data version includes: fault diagnosis strategy and diagnostic support parameters corresponding to the fault diagnosis strategy, so that after sending the target data version to the controller of the engine, the controller executes the corresponding fault diagnosis strategy according to the diagnostic support parameters in the target data version, thereby extending the use time of the engine before the fault alarm. It can be seen that in this embodiment, no matter what working condition state the current working condition portrait of the engine is, the diagnostic prediction model can be used to optimize the fault diagnosis strategy and the corresponding diagnostic support parameters, so that when the controller of the engine executes the fault diagnosis strategy according to the optimized diagnostic support parameters, the use time of the engine before the fault alarm can be extended, so that the cost of using the engine can be reduced.

[0098] In one implementation, the diagnostic prediction model may include: a fault level classification model, a risk classification model, a fault strategy pool, and a data optimization model, and the models here can all be machine learning models. The fault strategy pool here may include at least one strategy parameter group, and each strategy parameter group includes at least: a fault level parameter, an occurrence parameter, and a risk parameter.

[0099] Based on this, step 103 can be implemented in the following ways: Figure 2 As shown in:

[0100] Step 201: According to the current working condition portrait, obtain the target working condition portrait and the initial data version in the data version library.

[0101] The data version library contains at least one typical operating condition portrait, and each typical operating condition portrait corresponds to an optimized data version. The optimized data version means that when the engine controller executes the fault diagnosis strategy according to the diagnostic support parameters in the optimized data version, the engine can be used for a longer time before the fault alarm. For example, the optimized data version corresponding to the typical operating condition portrait is shown in Table 2.

[0102] Table 2 Optimized data version

[0103]

[0104] The target working condition portrait and the current working condition portrait meet the matching conditions, and the initial data version is the optimized data version corresponding to the target working condition portrait.

[0105] Specifically, the matching condition is: the portrait similarity between the current working condition portrait and the target working condition portrait is greater than or equal to a first threshold.

[0106] In one case, the target working condition profile is a typical working condition profile in the data version library that is completely consistent with the current working condition profile;

[0107] In another case, the target operating condition portrait is not completely consistent with the current operating condition portrait, but the target operating condition portrait is a typical operating condition portrait in the data version library that is closest to the current operating condition portrait.

[0108] In one implementation, step 201 may be implemented in the following manner:

[0109] First, the portrait similarity between the current working condition portrait and each typical working condition portrait in the data version library is obtained respectively; then, the typical working condition portrait with the greatest portrait similarity is selected from the data version library as the target working condition portrait.

[0110] The similarity between the working condition portraits can be obtained by a similarity algorithm or a machine learning model that can obtain similarity. Alternatively, in this embodiment, the working condition portraits can be compared within the parameter range of each identical working condition parameter, and the portrait similarity can be obtained according to the difference between the parameter ranges.

[0111] In another implementation, step 202 may be implemented in the following manner:

[0112] First, the clustering model is used to obtain the current working condition type corresponding to the current working condition portrait.

[0113] Then, the current working condition type is compared with the working condition type of each typical working condition portrait in the data version library. If there is a typical working condition type consistent with the current working condition type in the data version library, then the target working condition portrait is determined by the typical working condition portrait consistent with the current working condition type in the data version library, and the optimized data version corresponding to the target working condition portrait is determined as the initial data version; if there is no typical working condition type consistent with the current working condition type in the data version, then the typical working condition portrait with the greatest portrait similarity with the current working condition portrait in the data version library is used as the target working condition portrait, and the optimized data version corresponding to the target working condition portrait is determined as the initial data version.

[0114] The clustering model can be trained through training samples, so that the clustering model can identify the working condition type corresponding to the input working condition portrait. Different working condition types have at least one working condition parameter that is different, specifically, at least one working condition parameter is in a different parameter range. The working condition types that the clustering model has learned can be shown in Table 3:

[0115] Table 3 Working condition types

[0116] Type 1 Type 2 … Type Speed Speed Speed Speed Torque Torque Torque Torque temperature temperature temperature temperature humidity humidity humidity humidity … … … …

[0117] Step 202: using the fault level classification model, obtain the current fault level corresponding to the target operating condition profile.

[0118] The current fault level represents the degree of danger of the engine fault. The fault level classification model may be a machine learning model, which can predict the fault level corresponding to the input working condition profile after model training.

[0119] For example, the fault level classification model can predict the following four fault levels:

[0120] Level of fatal failure: This level affects the safe operation of the engine or protects against situations that do not comply with regulations;

[0121] Major failure level: This level causes the engine or assembly to fail, but does not affect the safe operation of the engine;

[0122] General fault level: This level will cause downtime, but can be repaired with spare parts or low-value parts, or does not cause downtime but has affected normal use and needs adjustment and repair;

[0123] Minor fault level: This level will not cause downtime, will not affect the normal operation of the engine, and does not require replacement of parts. It is a fault that can be easily eliminated.

[0124] Step 203: Using the risk classification model, obtain the current occurrence degree and current risk degree corresponding to the target operating condition profile.

[0125] The current occurrence degree represents the frequency of engine failure, and the current risk degree represents the risk level of engine failure. The risk classification model can be a machine learning model, which can predict the occurrence degree and risk level corresponding to the input working condition profile after model training.

[0126] For example, the risk classification model can predict three types of occurrence: high occurrence, average occurrence, and low occurrence. The risk classification model can predict three types of risk: high occurrence risk, medium occurrence risk, and low occurrence risk.

[0127] It should be noted that the execution order of step 202 and step 203 may not be limited to the order shown in the accompanying drawings. Step 203 may be executed first and then step 202, or step 202 and step 203 may be executed simultaneously. Different technical solutions formed by different execution orders are all within the protection scope of this application.

[0128] Step 204: According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, the target parameters matching the current operating condition profile are screened out in the fault strategy pool.

[0129] Among them, the target parameters include the target fault level, target occurrence rate and target risk level. It should be noted that the strategy parameter group contained in the fault strategy pool is to make customized predictions and reminders for the level, occurrence rate, risk, etc. of engine faults under different working condition portraits according to early warning conditions, diagnostic conditions, maintenance conditions, etc.

[0130] Step 205: Input the target fault level, target occurrence degree, target risk degree, current operating condition profile and initial data version into the data optimization model to obtain the target data version corresponding to the current operating condition profile output by the data optimization model.

[0131] The data optimization model can be trained in advance using training samples. The training samples of the data optimization model can be obtained using typical working condition portraits in a typical working condition library.

[0132] Specifically, in this embodiment, each typical working condition image in the typical working condition library can be processed as follows: Figure 3 The corresponding training samples are obtained by the process shown in:

[0133] Step 301: Using the fault level classification model, obtain the current fault level corresponding to the typical working condition portrait.

[0134] Step 302: Using the risk classification model, obtain the current occurrence degree and current risk degree corresponding to the typical working condition portrait.

[0135] It should be noted that the execution order of step 301 and step 302 may not be limited to the order shown in the accompanying drawings. Step 302 may be executed first and then step 301, or step 301 and step 302 may be executed simultaneously. Different technical solutions formed by different execution orders are all within the protection scope of this application.

[0136] Step 303: According to the current fault level, current occurrence degree and current risk degree corresponding to the typical operating condition portrait, the indicator parameters matching the typical operating condition portrait are screened out in the fault strategy pool.

[0137] Among them, the indicator parameters include: fault level indicator, occurrence indicator, and risk indicator.

[0138] Step 304: Input the typical operating condition portrait and the historical data version corresponding to the typical operating condition portrait into the control simulation system corresponding to the engine controller.

[0139] Among them, the control simulation system can use the historical data version corresponding to the typical working condition portrait as the control data version to output initial parameters according to the typical working condition portrait. The initial parameters include: initial fault level, initial occurrence degree, and initial risk degree.

[0140] Specifically, the control simulation system can also be called a virtual ECU hardware system. The control simulation system can simulate the fault diagnosis function of the engine controller, that is, perform fault judgment according to the control data version, specifically, execute the fault diagnosis strategy in the control data version according to the diagnostic support parameters in the control data version.

[0141] Step 305: Adjust the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the index parameters.

[0142] The initial parameter matching index parameter may be: the initial fault level matches the fault level index, the initial occurrence degree matches the occurrence degree index, and the initial risk degree matches the risk degree index.

[0143] Specifically, in this embodiment, according to the difference between the initial parameters and the indicator parameters output by the control simulation system, at least one diagnostic support parameter and a fault diagnosis strategy in the control data version may be adjusted in sequence until the initial parameters match the indicator parameters.

[0144] Based on this, the control data version used by the control simulation system when the initial parameters match the index parameters is the output sample in the training sample; the index parameters, typical working condition portraits, and the historical data versions corresponding to the typical working condition portraits are the input samples in the training samples. In this way, the input samples are input into the data optimization model, thereby obtaining the predicted data version output by the data optimization model, and then the data optimization model is optimized according to the loss function value between the predicted data version and the data version in the sample. After optimizing multiple training samples, the data optimization model can output the target data version corresponding to the target working condition portrait for the input data.

[0145] In one implementation, the diagnostic prediction model may include: a fault level classification model, a risk classification model, a fault strategy pool, and a control simulation system, and the models here can all be machine learning models. The fault strategy pool here may include at least one strategy parameter group, and each strategy parameter group includes at least: a fault level parameter, an occurrence parameter, and a risk parameter.

[0146] Based on this, step 103 can be implemented in the following ways: Figure 4 As shown in:

[0147] Step 401: According to the current working condition portrait, obtain the target working condition portrait and the initial data version in the data version library.

[0148] Among them, the data version library contains at least one typical working condition portrait, each typical working condition portrait corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching conditions; the initial data version is the optimized data version corresponding to the target working condition portrait.

[0149] Step 402: using the fault level classification model, obtain the current fault level corresponding to the target operating condition profile.

[0150] The current fault level represents the degree of danger of the engine fault.

[0151] Step 403: Using the risk classification model, obtain the current occurrence degree and current risk degree corresponding to the target operating condition profile.

[0152] The current occurrence degree represents the frequency of engine failure, and the current risk degree represents the risk level of engine failure.

[0153] It should be noted that the execution order of step 402 and step 403 may not be limited to the order shown in the accompanying drawings. Step 403 may be executed first and then step 402, or step 402 and step 403 may be executed simultaneously. Different technical solutions formed by different execution orders are all within the protection scope of this application.

[0154] Step 404: According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, the target parameters matching the current operating condition profile are screened out in the fault strategy pool.

[0155] Among them, the target parameters include target failure level, target occurrence degree and target risk degree.

[0156] Specifically, the implementation of step 401 to step 404 can refer to the implementation of step 201 to step 204 in the previous text, and will not be described in detail here.

[0157] Step 405: Input the current operating condition portrait and the initial data version into the control simulation system corresponding to the engine controller. The control simulation system can output initial parameters according to the current operating condition portrait with the initial data version as the control data version.

[0158] The initial parameters include: initial fault level, initial occurrence degree, and initial risk degree.

[0159] Step 406: Adjust the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the target parameters.

[0160] The initial parameters match the target parameters, which may be: the initial fault level matches the target fault level, the initial occurrence degree matches the target occurrence degree, and the initial risk degree matches the target risk degree.

[0161] Specifically, in this embodiment, according to the difference between the initial parameters output by the control simulation system and the target parameters, at least one diagnostic support parameter and a fault diagnosis strategy in the control data version may be adjusted in sequence until the initial parameters match the target parameters.

[0162] Step 407: The control data version used by the control simulation system when the initial parameters match the target parameters is determined to be the target data version corresponding to the current operating condition profile.

[0163] It should be noted that Figure 4In the process shown, multiple adjustments are required in step 406 to make the initial parameters match the target parameters and obtain the target data version. Figure 2 In the process shown, in step 205, the data optimization model only needs to be run once to obtain the target data version that matches the initial parameters with the target parameters, which can improve the efficiency of obtaining the optimized data version.

[0164] Based on the above implementation scheme, after step 103, the method in this embodiment may further include the following processing, such as Figure 5 As shown in:

[0165] Step 105: Generalize at least one operating condition parameter in the current operating condition profile to obtain a new typical operating condition profile.

[0166] The generalization processing in this embodiment refers to generalizing the current working condition portrait into a new working condition portrait that can represent multiple similar working condition states, that is, a typical working condition portrait.

[0167] Specifically, in this embodiment, at least one operating condition parameter in the current operating condition portrait can be adjusted to a parameter range including multiple parameter values ​​to obtain a new typical operating condition portrait.

[0168] Step 106: Add the new typical working condition portrait and its corresponding target data version to the data version library.

[0169] Therefore, each time the target data version is obtained, it is added to the data version library. In subsequent processing, when the operating parameters of the engine change again, after obtaining the new operating parameters and the corresponding new operating condition portrait, the target operating condition portrait that meets the matching conditions with the new operating condition portrait is first obtained in the data version library, and the corresponding target data version is obtained according to the implementation scheme in the previous article. After the new target data version is transmitted to the engine controller, the controller executes the corresponding fault diagnosis strategy according to the diagnostic support parameters in the new target data version, so that the engine's usage time before the fault alarm is extended under the new operating parameters, thereby reducing user maintenance costs.

[0170] In another implementation, the optimized data version in the data version library can also be preset by the user. For example, the fault diagnosis strategy and corresponding diagnostic support parameters manually calibrated by the engineer before the engine leaves the factory are added to the data version library as the optimized data version.

[0171] In one implementation, before step 104, the method in this embodiment may further include the following processing: Figure 6 As shown in:

[0172] Step 107: Determine whether the target data version is consistent with the current data version used by the engine controller. If the target data version is inconsistent with the current data version, execute step 104. At this time, the engine controller uses the diagnostic support parameters in the received target data version to execute the corresponding fault diagnosis strategy. If the target data version is consistent with the current data version, end the current process, that is, do not execute step 104. At this time, the engine controller continues to use the diagnostic support parameters in the original data version to execute the fault diagnosis strategy, so as to extend the engine usage time before the fault alarm and reduce the amount of data transmission.

[0173] The above introduces an engine fault diagnosis optimization method based on operating condition portrait provided by an embodiment of the present application. The following will introduce a device for executing the above-mentioned engine fault diagnosis optimization method based on operating condition portrait.

[0174] See also Figure 7 , is a schematic diagram of the structure of an engine fault diagnosis and optimization device based on working condition profiling provided in an embodiment of the present application. The device can be deployed in an electronic device capable of data processing, and the electronic device can be a local device connected to the engine, or a cloud device connected to the engine. The technical solution in this embodiment is mainly used to extend the use time of the engine before the fault alarm.

[0175] Specifically, the device in this embodiment may include the following units:

[0176] The parameter obtaining unit 701 is used to obtain the operating parameters of the engine, wherein the operating parameters at least include a plurality of operating condition parameters;

[0177] A portrait obtaining unit 702, configured to obtain a current operating condition portrait of the engine according to the operating parameters;

[0178] The data processing unit 703 is used to process the current operating condition profile and the initial data version using the diagnosis prediction model to obtain a target data version corresponding to the current operating condition profile; the target data version includes: a fault diagnosis strategy and a diagnosis support parameter corresponding to the fault diagnosis strategy;

[0179] The diagnostic prediction model is obtained by training using training samples, and the training samples are obtained using a typical operating condition library, and the typical operating condition library contains multiple typical operating condition portraits;

[0180] The data transmission unit 704 is used to send the target data version to the controller of the engine, and the controller executes the fault diagnosis strategy according to the diagnosis support parameters, so that the service life of the engine before the fault alarm is extended.

[0181] By means of the above technical scheme, in an engine fault diagnosis optimization device based on working condition portrait provided by an embodiment of the present application, after obtaining the operating parameters of the engine including multiple working condition parameters, the current working condition portrait of the engine is first obtained according to the operating parameters, and then the diagnostic prediction model trained by the typical working condition portrait is used to process the current working condition portrait and the initial data version to obtain the target data version corresponding to the current working condition portrait, and the target data version includes: fault diagnosis strategy and diagnostic support parameters corresponding to the fault diagnosis strategy, so that after sending the target data version to the controller of the engine, the controller executes the corresponding fault diagnosis strategy according to the diagnostic support parameters in the target data version, thereby extending the use time of the engine before the fault alarm. It can be seen that in the present application, no matter what working condition state the current working condition portrait of the engine is, the diagnostic prediction model can be used to optimize the fault diagnosis strategy and the corresponding diagnostic support parameters, so that when the controller of the engine executes the fault diagnosis strategy according to the optimized diagnostic support parameters, the use time of the engine before the fault alarm can be extended, so that the cost of using the engine can be reduced.

[0182] The present application also provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0183] The memory is used to store computer programs;

[0184] The processor is used to execute the computer program so that the electronic device can implement an engine fault diagnosis optimization method based on operating condition profiling according to the above-mentioned first aspect or any implementation method of the first aspect.

[0185] refer to Figure 8 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present application. The electronic device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 8 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0186] like Figure 8As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 to a random access memory (RAM) 803. When the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 803. The processing device 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0187] Typically, the following devices may be connected to the I / O interface 805: input devices 806 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a memory card, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0188] Also provided in an embodiment of the present application is a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any one of the engine fault diagnosis optimization methods based on operating condition profiling provided in an embodiment of the present application.

[0189] A computer storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When one or more computer programs are executed by an electronic device, the electronic device can implement any engine fault diagnosis optimization method based on operating condition profiling provided in an embodiment of the present application.

[0190] The following takes a diesel engine as an example to explain the technical solution of the present application in detail:

[0191] At present, engine fault diagnosis and prediction generally use existing road spectrum data and existing accumulated experience to judge and diagnose the current engine operating conditions and monitoring data. In the case of substantial faults, the current fault can be accurately prompted, but this method is a post-judgment. When the relevant fault is prompted, the actual engine has already experienced this fault, and its prediction ability is relatively insufficient.

[0192] For prediction reminders on component life and maintenance, scheduled replacement or maintenance is used. However, in actual applications, due to different operating conditions, the actual component life and maintenance cycle can be extended, thereby reducing customer usage costs.

[0193] Therefore, the current diagnostic strategy is a limited diagnostic strategy based on limited data. How to scientifically diagnose various types of faults and predict the life and maintenance cycle of components is an urgent problem that needs to be solved.

[0194] In view of this, the present application adds a road spectrum collection and extraction strategy on the basis of the existing limited road spectrum data after the engine is installed on the vehicle after leaving the factory, and performs self-learning of engine faults and operating conditions based on the original road spectrum data and the processed new road spectrum data. For the current operating conditions of the vehicle, a reference recommended diagnostic strategy is proposed to ensure that the engine is pre-diagnosed, component life predicted, maintenance cycle predicted, fault warning, etc. are carried out while maintaining the best performance, emissions, economy and other indicators.

[0195] like Fig. 9 As shown in the figure, it is the overall execution logic flow chart of the engine fault automatic diagnosis prediction optimization method based on working condition adaptive learning in this application. Among them, the engine fault automatic diagnosis prediction optimization method based on working condition adaptive learning can be divided into two stages: the first stage is the offline manual optimization stage, Fig. 9 As shown in the left part of the figure; the second stage is the cloud self-learning optimization stage, as shown in Fig. 9 The right part shows:

[0196] In the cloud, the predefined engine diagnostic library is imported into the typical engine diagnostic library (i.e., the typical operating condition library mentioned above), and the engine operating conditions manually updated in the offline manual optimization stage on the left are also imported into the typical engine diagnostic library. Based on this, the initial data version is optimized using the diagnostic prediction model for the typical operating condition portrait in the typical operating condition library to obtain the diagnostic logic and data (i.e., the target data version) based on the local (engine) optimization. Then, the controller fault diagnosis strategy and diagnostic support parameters based on the local optimization are transmitted to the engine controller via OTA to update the data of the fault diagnosis strategy and diagnostic support parameters. As the engine runs after leaving the factory, the engine operating condition parameters can be detected by sensors to realize data collection. The data features can be extracted in the cloud to construct the vehicle engine operating condition portrait (i.e., the current operating condition portrait), and then, the clustering model in the cloud is used for automatic clustering.

[0197] If the operating condition type obtained by clustering is an existing classification, that is, in the portrait-based diagnostic data version library, there is a target operating condition portrait whose operating condition type matches the current operating condition portrait, the diagnostic data version in the clustering result (that is, the corresponding optimized data version) is extracted as the initial data version, and then the target operating condition portrait and the initial data version are processed using the diagnostic prediction model to obtain the target data version (that is, the diagnostic logic and data based on local optimization); and, the typical engine diagnostic library is updated, for example, if the current operating condition portrait is not in the typical engine diagnostic library, the current operating condition portrait is added to the typical engine diagnostic library.

[0198] If the operating condition type obtained by clustering is a new classification, then in addition to manually extracting the new operating condition portrait of the vehicle and then using it to train the clustering model so that the clustering model is added with the new portrait classification, a new operating condition portrait is generated as the data of the typical engine diagnostic library and updated to the typical engine diagnostic library. In addition, after extracting the new vehicle portrait, the diagnostic data version with similar results can be extracted from the portrait-based diagnostic data version library, and then the target operating condition portrait and the initial data version are processed using the diagnostic prediction model to obtain the target data version.

[0199] Furthermore, after obtaining the target data version (i.e., diagnostic logic and data based on local optimization), the new portrait data can be added to the data version library after data generalization processing.

[0200] The detailed scheme is described as follows:

[0201] 1. Offline manual optimization part:

[0202] This part is used when the engine or vehicle of this model is developed for the first time or after a major update. The execution process is as follows:

[0203] 1. Extract fault information and fault classification information from the operating condition information of the typical engine diagnostic library;

[0204] 2. Conduct rule calibration and testing based on the generalized diagnostic rules and component type classification of various components under standard working conditions;

[0205] 3. Cycle adjustment to obtain the optimal diagnostic parameter (the conservative value under the current state, i.e. the indicator parameter mentioned above), and use this parameter as the target value for subsequent calibration;

[0206] 4. Based on the above adjustments, generate a controller diagnostic strategy data version applicable to this type of information;

[0207] 5. Add the above data version and predefined working condition information to the cloud-based portrait-based data version library as the optimized data version;

[0208] 6. Generate the final production version data based on the above data, and perform data flashing (update) on multiple controllers during the production process;

[0209] 7. After the vehicle or engine is delivered to the user, the operating conditions, road spectrum information and various performance parameters of various engines are continuously collected;

[0210] 8. Manually analyze the data to extract the working conditions, and manually update them to the typical engine working condition library in the cloud;

[0211] 2. Cloud self-learning optimization model part:

[0212] After the engine or vehicle is delivered to the user, the continuous optimization work is carried out, combined with the Internet of Vehicles function, to continuously optimize the vehicle controller data. The execution process is as follows:

[0213] 1. Data extraction and vehicle profiling are performed through the data collected by the vehicle end to generate working condition characteristic data based on this specific vehicle;

[0214] 2. Use the clustering model to classify the current vehicle portrait;

[0215] If the classification is normal, the operating condition information of this portrait is used to update the data in the typical engine operating condition library to make it more complete, and the controller data suitable for the portrait is extracted from the portrait-based data version library according to the clustering results;

[0216] If the classification is abnormal:

[0217] (1) Push data to the manual portrait extraction function;

[0218] (2) Through human participation, the clustering model can be retrained to achieve normal classification, that is, the clustering model can classify new working condition types;

[0219] (3) At the same time, human participation can also redefine this portrait as a new vehicle portrait (i.e., working condition portrait), add a new classification identifier to the data version library based on the portrait, and simultaneously extract a controller data closest to this portrait from the data version library (i.e., the initial data version mentioned above);

[0220] 3. The extracted controller diagnostic data is sent to the diagnostic prediction model, and is classified and predicted with the data in the typical engine operating condition library (i.e., the current fault level, current occurrence degree, current risk degree, target fault level, target occurrence degree, and target risk degree are obtained in the previous text), and finally the optimal controller diagnostic data suitable for this machine (i.e., the target data version) is obtained, and then updated to the engine controller through OTA;

[0221] 4. If the new data updated here is a newly added portrait data, the data will be generalized and added to the portrait-based data version library for subsequent use.

[0222] In this application, the diagnostic prediction model is subdivided into a fault classification prediction model, a fault risk prediction model, a diagnostic strategy pool based on working condition profiling, and a machine learning model.

[0223] Among them, the fault classification prediction model: classifies the faults of existing engines to facilitate different business stakeholders to handle the engine faults according to the fault level.

[0224] Fault risk prediction model: predicts the frequency and risk level of this type of fault to remind business-related personnel to focus on key fault categories.

[0225] Fault strategy pool: Under different operating conditions, customized predictions and reminders are made for fault classification, occurrence, risk level, etc., as well as warning conditions, diagnostic conditions, maintenance conditions, etc.

[0226] like Fig.10 As shown in the figure, it is the process of optimizing the fault diagnosis strategy. It includes:

[0227] After collecting the initial data of vehicle operation, the working condition scan is input into the strategy pool. At the same time, the original component diagnosis function model is obtained and the component information is extracted. After passing through the strategy pool, the diagnosis strategies of various components and functions are executed, including:

[0228] The fault classification, risk level and component category of the components are judged. If the fault level is a fatal fault, an alarm is directly issued and the machine is shut down; if the fault level is a major fault and the risk is high, an alarm is also issued and the machine is shut down; if the fault level is a general fault and the risk is medium or low, only an alarm can be issued; if the fault level is a general fault and the risk is low, only a prompt can be issued after the strategy is readjusted through the strategy pool; if the fault level is a minor fault and the risk is low, no processing can be performed after the strategy is readjusted through the strategy pool, thereby extending the engine usage time before the fault alarm.

[0229] The following is an example of optimizing the replacement strategy for a component after a failure:

[0230] Parts are divided into different important categories, such as basic parts, important parts, general parts, electrical parts, wearing parts, etc. This embodiment takes an air filter as an example, as shown in Table 4:

[0231] Table 4 Parts replacement strategy

[0232] Working parameters Standard working conditions User Case 1 User Case 2 Air filter life 1000h Key Parameters Inlet negative pressure Judgment criteria Pressure difference>-20kpa Pressure difference>-20kpa Pressure difference>-20kpa Indirect indicators 0-1000 hours 1000+t1 (500, 250, 125) < 2000 hours 1000+t1+t2 hours Strategy 1 Check the air filter for wear and replace it if it is damaged or broken. Strategy 2 If it is intact, clean the air filter and continue to use it to extend the service life t1; update and record the air filter life at this time {user working condition 1, life time 1000+t1 (500), clean the air filter and continue to use it} Strategy 3 If it is intact, clean the air filter and continue to use it to extend the service life by t2; update and record the air filter life at this time {user working condition 2, life time 1000+t1+t2, clean the air filter and continue to use it}

[0233] Among them, the air filter is calibrated in the laboratory under the following working conditions: speed, internal torque, fuel injection amount, ambient temperature, ambient pressure, vehicle speed, slope, and altitude;

[0234] Direct indicator judgment standard: intake negative pressure difference is greater than -20kpa;

[0235] According to the standard operating environment, the service life of the air filter is 1,000 hours. However, in actual use, the operating conditions of the user's vehicle are different from those of the standard laboratory, and different maintenance strategies are implemented for the life of the air filter.

[0236] Under working condition 1, the user has exceeded the original calibrated service life by 1000 hours and exceeded t1 hours. At this time, the air filter appearance is checked for obvious damage and aging, and the intake negative pressure difference meets the standard. The relationship between the working condition and the life span at this time: According to different working conditions, such as extreme working conditions, urban areas, suburbs and other working conditions, the correlation between the working condition and the life of the parts is established, and the correlation coefficient is obtained; under specific working conditions, the extended life of the parts is equal to the life value under the original calibrated working condition + correlation coefficient × calibrated life value.

[0237] The strategy pool under different working conditions can be used as input for users to personalize the vehicle engine selection. If the user's working conditions change within a certain time range, the user can be reminded that there is a range of parameter selection that can be changed.

[0238] It should be noted that the data output by the above fault classification, fault risk model and strategy pool, as well as the closest supporting data version currently extracted, are sent to the virtual calibration model (i.e., the data optimization model mentioned above) for model training.

[0239] The training samples of the automatic virtual calibration model can be obtained by Fig.11 The process shown obtains:

[0240] First of all, the ECU control logic library is a Simulink-based control logic model of all control logic in the ECU during the electronic control development process. It is a functional module that realizes the entire control logic of the engine.

[0241] The control model is a large logic-corrected model obtained through a large number of engine calibrations, based on the need to adjust the model parameters in various situations, and is generated through data training.

[0242] The working condition scanning input function is a working condition operation logic library generated based on a large amount of calibration experience. This logic library is obtained by generalizing all the data of the classification of the above working condition portraits. The various types of data here are large amounts of automatically generated multidimensional data arrays.

[0243] The virtual ECU hardware system is responsible for performing virtual execution (i.e., simulation) according to the extracted ECU control logic (i.e., data version), and performing individual and multi-dimensional scanning according to each parameter of each of the above-mentioned data arrays, and constantly monitoring the output parameters or intermediate parameters to ensure that they meet the predefined standards (i.e., the indicator parameters mentioned above).

[0244] Stop when the data meets expectations (that is, the initial parameters match the indicator parameters) or converges to a certain degree.

[0245] For parameters with divergent results, manual confirmation is required later to obtain the final result data, that is, the output samples in the training samples.

[0246] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0247] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0248] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0249] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. An engine fault diagnosis optimization method based on operating condition portrait, characterized in that: The method comprises: Obtaining operating parameters of the engine, wherein the operating parameters include at least a plurality of operating condition parameters; According to the operating parameters, a current operating condition profile of the engine is obtained; The current operating condition profile and the initial data version are processed by using a diagnostic prediction model to obtain a target data version corresponding to the current operating condition profile; the target data version includes: a fault diagnosis strategy and a diagnostic support parameter corresponding to the fault diagnosis strategy; The diagnostic prediction model is obtained by training using training samples, and the training samples are obtained using a typical operating condition library, and the typical operating condition library contains multiple typical operating condition portraits; Sending the target data version to a controller of the engine, and having the controller execute the fault diagnosis strategy according to the diagnosis support parameter, so that the service life of the engine before a fault alarm is extended; The diagnostic prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a data optimization model; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter; Wherein, using the diagnostic prediction model, processing is performed according to the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including: According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library; The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait; Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine; Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure; According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree; Inputting the target fault level, the target occurrence degree, the target risk degree, the current operating condition portrait and the initial data version into the data optimization model to obtain the target data version corresponding to the current operating condition portrait output by the data optimization model; or, The diagnostic prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a control simulation system; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter; Wherein, using the diagnostic prediction model, processing is performed according to the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including: According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library; The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait; Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine; Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure; According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree; Inputting the current operating condition portrait and the initial data version into a control simulation system corresponding to the controller of the engine, the control simulation system can output initial parameters according to the current operating condition portrait using the initial data version as the control data version; the initial parameters include: initial fault level, initial occurrence degree, and initial risk degree; Adjusting the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the target parameters; The control data version used by the control simulation system when the initial parameters match the target parameters is determined as the target data version corresponding to the current operating condition portrait.

2. The method according to claim 1, characterized in that: The matching condition is: the portrait similarity between the current working condition portrait and the target working condition portrait is greater than or equal to a first threshold.

3. The method according to claim 2, characterized in that According to the current working condition portrait, the target working condition portrait and the initial data version are obtained in the data version library, including: Using a clustering model, obtaining a current operating condition type corresponding to the current operating condition portrait; Compare the current working condition type with the working condition type of each typical working condition portrait in the data version library; If there is a typical operating condition type in the data version library that is consistent with the current operating condition type, the typical operating condition portrait in the data version library that is consistent with the current operating condition type is determined as the target operating condition portrait and the optimized data version corresponding to the target operating condition portrait is determined as the initial data version; If there is no typical operating condition type in the data version library that is consistent with the current operating condition type, the typical operating condition portrait with the greatest portrait similarity with the current operating condition portrait in the data version library is used as the target operating condition portrait and the optimized data version corresponding to the target operating condition portrait is determined as the initial data version.

4. The method according to claim 1, characterized in that The training samples are obtained in the following way: The following processing is performed on each typical working condition image in the typical working condition library: Using the fault level classification model, obtaining the current fault level corresponding to the typical working condition portrait; Using the risk classification model, obtaining the current occurrence degree and current risk degree corresponding to the typical working condition portrait; According to the current fault level, current occurrence degree and current risk degree corresponding to the typical working condition portrait, in the fault strategy pool, the indicator parameters matching the typical working condition portrait are screened out; The index parameters include: fault level index, occurrence index, and risk index; The typical operating condition portrait and the historical data version corresponding to the typical operating condition portrait are input into the control simulation system corresponding to the controller of the engine, and the control simulation system can output initial parameters according to the typical operating condition portrait using the historical data version corresponding to the typical operating condition portrait as the control data version; the initial parameters include: initial fault level, initial occurrence degree, and initial risk degree; Adjusting the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the indicator parameters; Among them, the control data version used by the control simulation system when the initial parameters match the index parameters is the output sample in the training sample; the index parameters, the typical operating condition portrait and the historical data version corresponding to the typical operating condition portrait are the input samples in the training sample.

5. The method according to claim 4, characterized in that The initial parameters match the indicator parameters, including: The initial fault level matches the fault level index, the initial occurrence degree matches the occurrence degree index, and the initial risk degree matches the risk degree index.

6. The method according to claim 1, characterized in that After obtaining the target data version corresponding to the current working condition portrait, the method further includes: Generalizing at least one operating condition parameter in the current operating condition profile to obtain a new typical operating condition profile; The new typical working condition portrait and its corresponding target data version are added to the data version library.

7. An engine fault diagnosis and optimization device based on operating condition portrait, characterized in that: The device comprises: A parameter obtaining unit, used to obtain operating parameters of the engine, wherein the operating parameters at least include a plurality of operating condition parameters; A portrait obtaining unit, used for obtaining a current operating condition portrait of the engine according to the operating parameters; A data processing unit, used to process the current operating condition portrait and the initial data version using a diagnostic prediction model to obtain a target data version corresponding to the current operating condition portrait; the target data version includes: a fault diagnosis strategy and a diagnostic support parameter corresponding to the fault diagnosis strategy; The diagnostic prediction model is obtained by training using training samples, and the training samples are obtained using a typical operating condition library, and the typical operating condition library contains multiple typical operating condition portraits; A data transmission unit, used for sending the target data version to the controller of the engine, and the controller executes the fault diagnosis strategy according to the diagnosis support parameter, so that the service life of the engine before the fault alarm is extended; The diagnostic prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a data optimization model; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter; The data processing unit uses the diagnostic prediction model to process the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including: According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library; The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait; Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine; Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure; According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree; Inputting the target fault level, the target occurrence degree, the target risk degree, the current operating condition portrait and the initial data version into the data optimization model to obtain the target data version corresponding to the current operating condition portrait output by the data optimization model; or, The diagnostic prediction model includes: a fault level classification model, a risk classification model, a fault strategy pool and a control simulation system; the fault strategy pool includes at least one strategy parameter group, and the strategy parameter group includes at least a fault level parameter, an occurrence parameter and a risk parameter; The data processing unit uses the diagnostic prediction model to process the current working condition profile and the initial data version to obtain the target data version corresponding to the current working condition profile, including: According to the current working condition portrait, a target working condition portrait and an initial data version are obtained in a data version library; The data version library includes at least one typical working condition portrait, each of which corresponds to an optimized data version; the target working condition portrait and the current working condition portrait meet the matching condition; the initial data version is the optimized data version corresponding to the target working condition portrait; Using the fault level classification model, obtaining a current fault level corresponding to the target operating condition portrait; the current fault level represents the degree of danger of the fault occurring in the engine; Using the risk classification model, the current occurrence degree and the current risk degree corresponding to the target operating condition profile are obtained; the current occurrence degree represents the frequency of the engine failure, and the current risk degree represents the risk degree of the engine failure; According to the current fault level, current occurrence degree and current risk degree corresponding to the target operating condition profile, in the fault strategy pool, target parameters matching the current operating condition profile are screened out, wherein the target parameters include a target fault level, a target occurrence degree and a target risk degree; Inputting the current operating condition portrait and the initial data version into a control simulation system corresponding to the controller of the engine, the control simulation system can output initial parameters according to the current operating condition portrait using the initial data version as the control data version; the initial parameters include: initial fault level, initial occurrence degree, and initial risk degree; Adjusting the control data version used in the control simulation system so that the initial parameters output by the control simulation system match the target parameters; The control data version used by the control simulation system when the initial parameters match the target parameters is determined as the target data version corresponding to the current operating condition portrait.

8. An electronic device comprising at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement an engine fault diagnosis optimization method based on operating condition profiling as described in any one of claims 1 to 6.

Images