Reasoning generation method and device for automobile fault diagnosis related phenomena

Abstract

The present invention discloses a method and device for inference generation of automobile fault diagnosis related phenomena, the method is applied to an inference generation system of automobile fault diagnosis related phenomena, the system includes a data middle station, the method for inference generation of automobile fault diagnosis related phenomena includes: the data middle station counts the occurrence frequency of the fault cause of each fault phenomenon in a preset database; the data middle station derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each related phenomenon. The inference generation method of automobile fault diagnosis related phenomena of the present invention, through an inference model combining theory with actual data, systematically combs the correlation logic between various fault phenomena, achieves the purpose of assisting maintenance technicians to synchronously check potential fault phenomena in the actual fault diagnosis process, improves maintenance quality, and improves work efficiency and accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile fault diagnosis, and in particular to a method and device for generating reasoning of automobile fault diagnosis association phenomena. Background Art

[0002] In the actual automobile maintenance process, the common fault phenomenon identification methods are mainly determined by human perception, instrument detection, vehicle self-diagnosis, diagnostic instruments, etc. For some complex fault phenomena, it is necessary to use a variety of detection methods, mainly because when some fault phenomena occur, they are often accompanied by a variety of other related phenomena that have a causal or associated relationship with them. However, there is no further research on the related phenomenon in the prior art, which results in low work efficiency, low accuracy, and low maintenance quality during maintenance.

[0003] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgment or any form of suggestion that the information constitutes the prior art already known to a person skilled in the art. Summary of the invention

[0004] The purpose of the present invention is to provide a method and device for generating reasoning about automobile fault diagnosis related phenomena. Through a reasoning model that combines theory with actual data, the related logic between various fault phenomena is systematically sorted out, thereby achieving the purpose of assisting maintenance technicians to simultaneously check potential fault phenomena during the actual fault diagnosis process, improving maintenance quality, and improving work efficiency and accuracy.

[0005] To achieve the above-mentioned objectives, in a first aspect, the present invention provides an inference generation method for automobile fault diagnosis related phenomena, which is applied to an inference generation system for automobile fault diagnosis related phenomena, the system including a data middle platform, and the inference generation method for automobile fault diagnosis related phenomena includes: the data middle platform counts the occurrence frequency of the fault cause of each fault phenomenon in a preset database; the data middle platform derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each related phenomenon.

[0006] In one embodiment of the present invention, the data center station counts the occurrence frequency of the fault cause of each fault phenomenon in the preset database, including: the data center station extracts the keywords used to describe the fault phenomenon and the fault cause in the preset database; creates corresponding phenomenon common words and cause common words that are common across brands according to the extracted fault phenomenon keywords and fault cause keywords; the data center station performs frequency statistics on the occurrence number of each group of fault phenomena and fault causes in the preset database within the same product subsystem based on each group of phenomenon common words and cause common words created.

[0007] In one embodiment of the present invention, the data middle platform derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each related phenomenon, including: the data middle platform generates a first related phenomenon based on a preset database inference; the data middle platform generates a second related phenomenon based on a preset fault diagnosis and troubleshooting process inference; the data middle platform performs correlation logic judgment based on the first related phenomenon and the second related phenomenon, and generates a correlation ranking of each related phenomenon.

[0008] In one embodiment of the present invention, the data middle platform generates a first associated phenomenon based on inference from a preset database, including: the data middle platform performs quantitative statistics on different fault phenomena with the same fault cause based on information on the fault phenomenon and fault cause extracted from the preset database; establishes a phenomenon correlation algorithm model to count the number of common fault causes between other fault phenomena and the current fault phenomenon; calculates the correlation between other fault phenomena and the current fault phenomenon based on the phenomenon correlation algorithm model to generate a first associated phenomenon.

[0009] In one embodiment of the present invention, the data middle platform generates a second associated phenomenon based on inference of a preset fault diagnosis and troubleshooting process, including: the data middle platform performs quantitative statistics on the general meaning and fault causes of each group of DTC codes contained in historically developed vehicle models; and uses a phenomenon correlation algorithm model to calculate the correlation of DTC-type fault phenomena with the same cause and the same general meaning to generate a second associated phenomenon.

[0010] In one embodiment of the present invention, the data center performs association logic judgment based on the first association phenomenon and the second association phenomenon, and generates an association ranking of each association phenomenon, including: comparing the association degrees of the association phenomena existing in the first association phenomenon and the second association phenomenon, and retaining the association phenomena with high association degrees; retaining the association phenomena that only exist in the first association phenomenon or the second association phenomenon; and generating an association ranking of each association phenomenon.

[0011] In one embodiment of the present invention, the inference generation method of automobile fault diagnosis related phenomena also includes: the data center outputs related phenomenon recommendations corresponding to the user-side request based on the user-side request.

[0012] In the second aspect, the present invention provides an inference generation device for automobile fault diagnosis related phenomena, which is applied to an inference generation system for automobile fault diagnosis related phenomena, wherein the system includes a data middle station, and the inference generation device for automobile fault diagnosis related phenomena includes: a statistical module and a derivation generation module. The statistical module is used by the data middle station to count the occurrence frequency of the fault causes of each fault phenomenon in a preset database; and the derivation generation module is used by the data middle station to derive the correlation between different fault phenomena with the same fault cause, and generate the correlation ranking of each related phenomenon.

[0013] In one embodiment of the present invention, the data center station counts the occurrence frequency of the fault cause of each fault phenomenon in the preset database, including: the data center station extracts the keywords used to describe the fault phenomenon and the fault cause in the preset database; creates corresponding phenomenon common words and cause common words that are common across brands according to the extracted fault phenomenon keywords and fault cause keywords; establishes the association logic between the fault phenomenon keywords and the fault cause keywords and the phenomenon common words and cause common words; determines the technical logic between the fault phenomenon and the fault cause; the data center station performs frequency statistics on the occurrence number of each group of fault phenomena and fault causes in the preset database within the same product subsystem according to each group of phenomenon common words and cause common words created.

[0014] In one embodiment of the present invention, the data middle platform derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each related phenomenon, including: the data middle platform generates a first related phenomenon based on a preset database inference; the data middle platform generates a second related phenomenon based on a preset fault diagnosis and troubleshooting process inference; the data middle platform performs correlation logic judgment based on the first related phenomenon and the second related phenomenon, and generates a correlation ranking of each related phenomenon.

[0015] Compared with the prior art, the inference generation method and device for automobile fault diagnosis related phenomena according to the present invention, through an inference model that combines theory with actual data, systematically sorts out the correlation logic between various fault phenomena, thereby achieving the purpose of assisting maintenance technicians in the actual fault diagnosis process to simultaneously check potential fault phenomena and improve maintenance quality, and also improves work efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a method for inferring and generating automobile fault diagnosis related phenomena in the first embodiment of the present invention;

[0017] Figure 2 is a flow chart of counting occurrence frequency in the first embodiment of the present invention;

[0018] Figure 3 is a flow chart of generating a correlation ranking of each correlation phenomenon in the first embodiment of the present invention;

[0019] Figure 4 is a flow chart of generating a first correlation phenomenon in the first embodiment of the present invention;

[0020] Figure 5 is a flow chart of generating a second correlation phenomenon in the first embodiment of the present invention;

[0021] Figure 6 It is a structural schematic diagram of an inference generation device for automobile fault diagnosis associated phenomena in the first embodiment of the present invention. DETAILED DESCRIPTION

[0022] The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention, rather than to limit the embodiments of the present invention. It is also necessary to explain that, for ease of description, only parts related to the embodiments of the present invention are shown in the accompanying drawings, rather than all structures.

[0023] For ease of understanding, the main implementation concepts of the embodiments of the present invention are first briefly described.

[0024] With the electrification, new energy, and intelligent networking of automobiles, automobile failures are no longer simple mechanical problems. There are more electrical failures, system communications, and network bus failures. Automobile failures have become more complicated, involving comprehensive problems in multiple fields, which has brought unprecedented challenges to maintenance technicians. The number of intelligent and networked cars in society is constantly increasing, but it takes a long time to train a qualified auto repair expert, so that the supply and demand relationship is seriously unbalanced. In particular, verifying and solving faults based on experience or consulting relevant personnel for technical guidance is time-consuming and prone to diagnostic errors. In addition, it is impossible to simultaneously check potential fault phenomena, and the maintenance quality is low.

[0025] The inventor discovered the technical defects as described in the previous background technology, and then hoped to find a way to systematically sort out the correlation logic between various fault phenomena based on the benchmark database through a reasoning model that combines theory with actual data, so as to achieve the purpose of assisting maintenance technicians to simultaneously check potential fault phenomena in the actual fault diagnosis process, improve the maintenance quality, and improve work efficiency and accuracy.

[0026] In order to solve the technical problems existing in the prior art that it takes a long time and is prone to diagnostic errors to verify and solve faults based on experience or consulting relevant personnel for technical guidance, and that potential fault phenomena cannot be checked simultaneously, resulting in low maintenance quality, the inventor of the present invention has obtained an inference generation method and device for automobile fault diagnosis related phenomena through creative labor.

[0027] The inference generation method and device of automobile fault diagnosis related phenomena of the present invention are based on a reference database, and through an inference model combining theory with actual data, the related logic between various fault phenomena is systematically sorted out, thereby achieving the purpose of assisting maintenance technicians to simultaneously check potential fault phenomena in the actual fault diagnosis process, improving the maintenance quality, and improving work efficiency and accuracy.

[0028] Specifically, the above-mentioned benchmark database is constructed by a method for extracting information from automobile fault diagnosis big data, which specifically includes:

[0029] Based on the historical diagnosis database, the historical fault diagnosis data in the historical diagnosis database is classified and cleaned to generate data to be sampled. As a carrier of repair and maintenance information that has occurred during the actual driving of the vehicle, the historical diagnosis database contains a large number of different fault phenomena and the corresponding fault cause information; by classifying and cleaning the historical fault diagnosis data in the historical diagnosis database, the data to be sampled is generated for subsequent sampling and extraction. The automobile fault diagnosis big data information extraction method is based on the historical fault diagnosis data in the historical diagnosis database. The historical fault diagnosis data includes a large number of different fault phenomena and the corresponding fault cause information of a large number of different vehicles. By classifying and cleaning the information in the historical diagnosis database in a targeted manner, it is convenient for subsequent sampling and use, which greatly avoids the waste of manpower due to processing invalid data.

[0030] Through the data processing tool, the sampled data is sampled and information is extracted to generate benchmark diagnostic data. Two operation permissions are preset in the data processing tool: the phenomenon processing end (phenomenon processing module) and the cause processing end (cause processing module). The fault phenomenon and fault cause information in the sampled data are extracted and reviewed (through the sampling module) to generate benchmark diagnostic data. Through the data processing tool, the sampled data is sampled and information is extracted to generate benchmark diagnostic data. The benchmark diagnostic data is the data after the effective information in the historical diagnostic data is extracted, which makes the technical logic of the generated benchmark diagnostic data clearer, and is convenient for maintenance personnel to directly learn from experience, thereby achieving the purpose of improving their skill level.

[0031] The benchmark diagnostic data is submitted and stored in the benchmark database. The benchmark database stores the final benchmark diagnostic data, which is convenient for maintenance personnel to retrieve, view and learn at any time, so as to achieve the purpose of improving their skills.

[0032] The data processing tool includes a sampling module, a phenomenon processing module, and a cause processing module. The data processing tool can be a virtual software or a physical organization, with multiple functions (modules) built in. By setting parameters, specific functions can be realized. The data processing tool can expand different functions according to different user needs, which is convenient for users to use, thereby effectively improving the efficiency of data extraction, classification, cleaning and generation of benchmark diagnostic data.

[0033] Based on the historical diagnostic database, the historical fault diagnosis data in the historical diagnostic database is classified and cleaned, and the data to be sampled is generated, including: extracting the information of the historical fault diagnosis data stored in the historical diagnostic database through a first preset strategy to generate diagnostic data; classifying and cleaning the diagnostic data through a second preset strategy to generate the data to be sampled. The first preset strategy can specifically be the maintenance type, brand, vehicle model, fault phenomenon description, fault cause part number and fault cause name, etc.; through the first preset strategy, effective information can be extracted from the historical fault diagnosis data, thereby generating diagnostic data. Through the first preset strategy, the required information can be targetedly screened and extracted to generate diagnostic data, and it is beneficial to reduce the amount of data calculation of the fault diagnosis big data information extraction device in subsequent data processing. The second preset strategy is specifically to classify the data in the historical diagnostic database, clear the useless data in the historical diagnostic database that has extremely low reference value for analyzing fault diagnosis, and then retain the data to be sampled in the historical diagnostic database that is beneficial to analyzing fault diagnosis information. For example, the diagnostic data in the historical diagnostic database is classified into "invalid data" for the maintenance type of "normal maintenance" and "quality rectification" and the data containing "passive upgrade", "commercial vehicle inspection", "body corrosion" and other words in the "fault phenomenon description". Through the two preset strategies, the data to be sampled that is conducive to the subsequent extraction of effective fault diagnosis information can be accurately screened, which greatly reduces the workload of processing complex data.

[0034] The data to be sampled is sampled and information is extracted by a data processing tool to generate the reference diagnostic data, including: presetting a DTC diagnostic data identification rule in the sampling module; presetting a sample sampling rule in the sampling module, sampling the data to be sampled, and generating sample sampling data; extracting the fault diagnosis information in the DTC and non-DTC data in the sample sampling data, and generating the reference diagnostic data. The data processing tool is presetting a DTC diagnostic data keyword identification rule, and when the preset keyword appears in the "fault phenomenon description" of the fault diagnosis big data, the data processing tool automatically identifies the diagnostic data containing the DTC fault phenomenon, so as to facilitate the subsequent extraction of information in the diagnostic data, and extract the DTC diagnostic data separately; the identification rule is specifically: presetting words, letters, and numbers such as "current, read, fault code, report code, instrument display, computer check, OBD, FC, SPN, FMI, DTC" in the data processing tool, matching the preset keywords with the diagnostic big data, and identifying the DTC diagnostic data containing the keywords; and performing rolling maintenance on the keywords preset in the data processing tool for identifying the DTC diagnostic data. Preset DTC diagnostic data keyword recognition rules in the data processing tool can quickly obtain and identify DTC diagnostic data, which is convenient for subsequent extraction of DTC fault information. Set sampling rules and sample DTC and non-DTC fault diagnostic data in the data processing tool, and ensure that the sample covers all the diagnostic data of the fault cause parts that have appeared as much as possible; the data processing tool samples and extracts information from the sampled data by setting sampling rules, and covers all the diagnostic data of the fault cause parts that have appeared during the sampling process, which is conducive to cutting the workload of data processing step by step and summarizing and improving the information extraction rules for different types of diagnostic data in the future.

[0035] Submitting the reference diagnostic data to the reference database includes: the data processing tool submits the reference diagnostic data to the reference database. The final reference diagnostic data is stored in the reference database, which is convenient for maintenance personnel to retrieve and view at any time, thereby improving the work efficiency of maintenance personnel and effectively ensuring the accuracy of maintenance diagnosis.

[0036] Extracting fault diagnosis information from the DTC and non-DTC data in the sampled data and generating the reference diagnostic data includes: the phenomenon processing module extracting phenomenon keywords from the data of the marked meaning DTC class, comparing the phenomenon keywords extracted from the entry with the DTC code general meaning library, determining the DTC code corresponding to the DTC phenomenon keyword and the general meaning of the corresponding code, and using the corresponding DTC code as the first phenomenon general word corresponding to the DTC phenomenon keyword. The phenomenon processing module extracts phenomenon keywords from the data of the non-marked meaning DTC class, creates corresponding second phenomenon general words according to the extracted phenomenon keywords, and determines the technical logic between the phenomenon keywords and the second phenomenon general words. The cause processing module extracts fault cause keywords from the data of the non-marked meaning DTC class, creates corresponding cause general words according to the extracted fault cause keywords, and determines the technical logic between the second phenomenon general words, the fault cause keywords and the cause general words. The reference diagnostic data is generated based on the first phenomenon general words, the second phenomenon general words and the cause general words. The extraction of DTC diagnostic data information is to match the extracted keywords with the DTC codes in the DTC code general meaning library; if there is only one DTC that matches the keyword, the corresponding DTC code and the general meaning of the corresponding code are called; if there are multiple DTCs in the DTC code general meaning library that match the DTC phenomenon keywords in the diagnostic data, the VIN corresponding to the vehicle in the diagnostic data is used to retrieve the loading list of the vehicle model corresponding to the VIN, and the key parts number associated with the product subsystem to which each DTC belongs (i.e., the different functional systems or parts assemblies divided in the vehicle) is matched one by one with the part number in the loading list of the target vehicle model, and the DTC phenomenon keyword corresponding to the DTC code and the corresponding code general meaning are determined by the matching results. Since there is a corresponding standard fault cause description in the meaning of the DTC fault code, there is no need to repeatedly create the corresponding fault cause keyword and cause general word. Among them, the DTC code general meaning library contains: DTC code, product subsystem, fault code general meaning, key parts number and other elements.The information extraction of non-DTC diagnostic data includes the extraction of fault phenomenon information. The phenomenon processing module extracts the phenomenon keywords in the fault diagnosis data, creates corresponding common phenomena based on the extracted phenomenon keywords, reviews the technical logic between the phenomenon keywords and the phenomenon common words, completes the extraction of fault phenomenon information, and submits it to the benchmark database; for example: the phenomenon description in the fault diagnosis data is: the right side window lifter cannot be raised or lowered; the extractable phenomenon keywords are: right side glass, cannot be raised or lowered; the corresponding phenomenon common words can be created as: the right door glass cannot be raised or lowered; among them, in order to show that there are causal or associated relationships between different fault phenomena, when extracting keywords for multiple fault phenomena contained in the same diagnostic data, all the keywords of the fault phenomena with causal or associated relationships need to be extracted and shared, and the fault phenomena without causal or associated relationships have fault keywords extracted separately, and cross-brand common phenomenon common words are created according to different fault phenomenon keywords. The extraction of non-DTC diagnostic data information also includes the extraction of fault cause information. The cause processing module extracts the corresponding fault cause keywords in the diagnostic data, and creates corresponding cause general words based on the extracted fault cause keywords. It performs technical and logical judgment on the phenomenon general words, fault cause keywords, and cause general words to complete the extraction and submission of fault cause information. When the cause base word is created, it is created in the form of fault cause component + failure mode. For example, the fault diagnosis data is described as: the right side window lifter cannot be raised or lowered. After inspection, it is found that the glass lifter motor is not working. The extractable cause keywords are: right side glass, glass lifter motor, not working; the corresponding cause general words can be created as: right door glass lifter motor is damaged; the fault cause component is: right door glass lifter motor, and the failure mode is: damaged; when a single phenomenon involves multiple fault causes, it is necessary to extract the corresponding fault cause keywords for different fault causes, and create corresponding fault cause general words according to different fault cause keywords. Since different brands and models may have different names for the same parts, or different keyword descriptions for the same fault phenomenon, when creating common words, the common words corresponding to different fault phenomena and fault cause keywords can be the same; the common words corresponding to the same fault phenomenon and fault cause keywords must be the same. The benchmark diagnostic data generated here refers to the standard data stored in the benchmark database after the common words and logical relationships extracted above are determined. The extracted information is submitted to the benchmark database. For the fault phenomena and fault cause information that need further review and confirmation, the phenomenon processing end and the cause processing end can color-code them and label them as "pending", and organize experts to conduct technical reviews; the diagnostic data after the review and confirmation is stored in the benchmark database and the benchmark diagnostic data after information extraction is used as historical maintenance data.

[0037] The above content elaborates on the origin of the benchmark database, and the present invention is based on the benchmark database to generate the inference of automobile fault diagnosis related phenomena.

[0038] Embodiment 1

[0039] Figure 1 FIG. 1 is a flow chart of a method for generating an inference of an automobile fault diagnosis association phenomenon in the first embodiment of the present invention. Figure 1 As shown, the first embodiment provides a method for generating an inference of an automobile fault diagnosis association phenomenon, including:

[0040] Step S100: The data center station counts the occurrence frequency of the fault cause of each fault phenomenon in the preset database.

[0041] Specifically, the above-mentioned preset database is the benchmark database mentioned above, and the data middle platform is a data processing center that processes the benchmark data in the benchmark database, mainly to conduct frequency statistics on the causes of various fault phenomena in the benchmark data.

[0042] Based on the above analysis, the data center can lay the foundation for the subsequent reasoning of the correlation logic between various fault phenomena by statistically analyzing the occurrence frequency of the fault causes of various fault phenomena in the benchmark database.

[0043] In step S200, the data center derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each related phenomenon.

[0044] Specifically, the data center derives the correlation between different fault phenomena with the same fault cause based on the benchmark data extracted from the benchmark database, and generates a correlation ranking for each related phenomenon.

[0045] Based on the above analysis, it can be seen that through the reasoning model that combines theory with actual data, the correlation logic between various fault phenomena is systematically sorted out, which helps maintenance technicians to simultaneously check potential fault phenomena in the actual fault diagnosis process, improve maintenance quality, and improve work efficiency and accuracy.

[0046] In this embodiment, if Figure 2 As shown, step S100 includes:

[0047] Step S101, the data center extracts keywords from a preset database for describing fault phenomena and fault causes.

[0048] Step S102 , creating corresponding phenomenon common words and cause common words that are common across brands respectively according to the extracted keywords of the fault phenomenon and the extracted keywords of the fault cause.

[0049] In step S103, the data center platform performs frequency statistics on the number of occurrences of each group of fault phenomena and fault causes in the preset database within the same product subsystem based on each group of phenomenon common words and cause common words created.

[0050] Specifically, keywords used to describe fault phenomena and fault cause information in historical fault diagnosis big data (benchmark data) are extracted. Since different brands and models have different names for the same parts and different keyword descriptions for the same fault phenomenon, corresponding cross-brand common phenomenon common words and cause common words are created based on the extracted keywords, and the association logic between keywords and common words is established, and technical logic judgments are made between fault phenomena and fault causes. Based on each group of cause common words and phenomenon common words created, the data center performs frequency statistics on the number of occurrences of each group of phenomena + causes in the benchmark data within the same product subsystem. For example: Fault phenomenon A has 1, 2, 3, 4... and other fault causes. When counting the frequency, the data center performs statistics in the benchmark database according to the fault cause name and part number of the fault cause part in the fault cause according to the phenomenon + cause (A-1, A-2...).

[0051] Based on the above analysis, the data center can lay the foundation for the subsequent reasoning of the correlation logic between various fault phenomena by statistically analyzing the occurrence frequency of the fault causes of various fault phenomena in the benchmark database.

[0052] In this embodiment, if Figure 3 As shown, step S200 includes:

[0053] Step S201: The data center generates a first correlation phenomenon based on inference from a preset database.

[0054] Specifically, the data middle platform infers related phenomena based on the benchmark database. According to the fault phenomena and fault cause information extracted from the big data of various fault diagnosis that actually occurred in history, the correlation between different fault phenomena with the same fault cause is inferred to generate the first related phenomenon.

[0055] Based on the above analysis, the data center conducts quantitative statistics on other fault phenomena in the benchmark database that have the same cause as the current fault phenomenon according to the automobile sub-functional product system (i.e., different functional systems or parts assemblies divided in the vehicle), establishes a phenomenon correlation algorithm model, counts the number of common fault causes between other fault phenomena and the current fault phenomenon, and calculates the correlation between other fault phenomena and the current fault phenomenon based on the phenomenon correlation algorithm model.

[0056] In step S202, the data center generates a second associated phenomenon by reasoning based on a preset fault diagnosis and troubleshooting process.

[0057] Specifically, the data center generates the second related phenomenon based on the fault diagnosis and troubleshooting process reasoning. According to the control modules on the vehicle (such as the engine electronic control module, ABS electronic control module, etc.), the DTC-type fault phenomena of the phenomenon troubleshooting process in the historical fault diagnosis manual are extracted, and the extracted DTC-type fault phenomena are matched with the DTC code general meaning library to determine the general meaning of the DTC code corresponding to the DTC-type phenomenon code and the fault cause, which are stored in the benchmark database.

[0058] Based on the above analysis, the data center performs quantitative statistics (i.e., frequency) on the general meaning and fault causes of each group of DTC codes contained in the historically developed vehicle models, and calculates the correlation of DTC-type fault phenomena with the same cause and the same general meaning with reference to the correlation algorithm model in the above step S201, generates the correlation of DTC-type fault phenomena in the same control module, and completes the correlation phenomenon reasoning of DTC-type fault phenomena in the fault phenomenon troubleshooting process.

[0059] In step S203, the data center performs association logic judgment based on the first association phenomenon and the second association phenomenon, and generates an association ranking of each association phenomenon.

[0060] Specifically, the data center performs association logic judgment based on the first association phenomenon and the second association phenomenon, performs association logic judgment on the association phenomena that are inferred to have causal or associated relationships in two ways in step S201 and step S202, and generates a correlation ranking of each association phenomenon. When confirming the recommended association phenomena, for phenomena that have been technically judged and confirmed to have association logic, there is no need to perform the association calculation in the above steps S201 and S202; for unconfirmed inference phenomena, the data center sets an association threshold screening condition (the association is greater than or equal to 80%, and this threshold can be adjusted according to the actual calculation results), so as to recommend association phenomena that meet the association threshold condition on the PC software. The data center collects the recommended association phenomenon results checked by the maintenance technician in the PC software, and technically judges that the association logic between the recommended association phenomenon and the known phenomenon is reasonable, thereby completing the confirmation of the association logic between each association phenomenon (generating a correlation ranking of each association phenomenon).

[0061] In this embodiment, if Figure 4 As shown, step S201 includes:

[0062] In step S2011, the data center performs quantitative statistics on different fault phenomena with the same fault cause based on the information on the fault phenomenon and fault cause extracted from the preset database.

[0063] Specifically, the data center counts the number of other fault phenomena in the benchmark database that have the same cause as the current fault phenomenon according to the automobile sub-functional product system (i.e., the different functional systems or parts assemblies divided in the vehicle).

[0064] Step S2012: Establish a phenomenon correlation algorithm model to count the number of common fault causes between other fault phenomena and the current fault phenomenon.

[0065] Specifically, a phenomenon correlation algorithm model is established to count the number of common fault causes between other fault phenomena and the current fault phenomenon. Phenomenon correlation = (number of occurrences of the lower cause-phenomenon group) / (number of occurrences of the higher cause-phenomenon group)*100%.

[0066] Step S2013: Calculate the correlation between other fault phenomena and the current fault phenomenon according to the phenomenon correlation algorithm model to generate a first correlation phenomenon.

[0067] Specifically, the correlation between other fault phenomena and the current fault phenomenon is calculated according to the phenomenon correlation algorithm model; when the number of common causes between the phenomenon and the associated phenomenon is n, the correlation algorithm model calculates n rounds, and the result of the nth round of calculation is the correlation between the two fault phenomena.

[0068] For example: A and B are fault phenomena; the fault causes of A are a, c, and e; the fault causes of B are a and c; the number of identical causes is 2;

[0069] When A occurs, the probability of B occurring is: 1. The frequency of A~a is 50, the frequency of A~c is 30, and the frequency of A~e is 40;

[0070] 2. The frequency of B~a is 20, and the frequency of B~c is 60;

[0071] 3. 20 / 50 = 40%, 30 / 60 = 50% (first round);

[0072] Then, the correlation between phenomenon A and phenomenon B = 40% / 50% = 80% (round 2).

[0073] In this embodiment, if Figure 5 As shown, step S202 includes:

[0074] In step S2021, the data center performs quantitative statistics on the common meanings and fault causes of each group of DTC codes contained in the historically developed vehicle models.

[0075] Specifically, the data center performs quantitative statistics (i.e., frequency) on the common meanings and fault causes of each set of DTC codes contained in historically developed vehicle models.

[0076] Step S2022: A phenomenon correlation algorithm model is used to calculate the correlation of DTC-type fault phenomena with the same cause and the same general meaning, so as to generate a second correlation phenomenon.

[0077] Specifically, the correlation degree of DTC-type fault phenomena with the same cause and the same general meaning is calculated by referring to the correlation degree algorithm model in the above step S201, and the correlation degree of DTC-type fault phenomena in the same control module is generated, so as to complete the correlation phenomenon reasoning of DTC-type fault phenomena in the fault phenomenon troubleshooting process. For example: when A occurs, the probability of B occurring (A and B represent the general meaning of DTC corresponding to different DTC-type fault phenomena, and a and c are the corresponding fault causes):

[0078] 1. The frequency of A~a is 10, and the frequency of A~c is 8;

[0079] 2. The frequency of B~a is 6, and the frequency of B~c is 7;

[0080] 3. Calculation of correlation: 6 / 10=60%, 7 / 8=87.5% (first round);

[0081] 4. The correlation between A and B is: 60% / 87.5%=68.57% (second round).

[0082] In this embodiment, step S203 includes:

[0083] Step S2031 , comparing the correlation degrees of the correlation phenomena in the first correlation phenomenon and the second correlation phenomenon, and retaining the correlation phenomena with the high correlation degree.

[0084] Specifically, for the association phenomena existing in both the reasoning methods in step S201 and step S202, the association degree is compared, and the one with the higher association degree is retained. For example, in both the above steps S201 and S202, A and B are both inferred as association phenomena. In the step S201 method, the association degree between A and B is 70%; in the step S202 method, the association degree between A and B is 80%. When the association logic is confirmed, the data of the inference of the association degree between A and B of 80% in the step S202 method is retained as the confirmation basis.

[0085] Step S2032: retain the associated phenomena that only exist in the first associated phenomenon or the second associated phenomenon.

[0086] Specifically, for the associated phenomena that only exist in one of the above-mentioned methods of reasoning, they are all retained. For example: in the above-mentioned step S201, A and B are inferred as associated phenomena; in the above-mentioned step S202, A and C are inferred as associated phenomena. When determining the associated phenomena, the associated phenomena B and C inferred by the two methods need to be retained and the associated logic determination is performed with A respectively.

[0087] Step S2033, generate the correlation ranking of each related phenomenon, which is used to assist maintenance technicians to sort each related phenomenon in order according to the correlation. For example: the related phenomenon of phenomenon A is inferred, and B, C, and D are inferred through the method in step S201 above; through the method B, D, and E in step S202 above, B, C, D, and E are sorted in descending order according to the size of their respective correlations. Among them, B and D are inferred in both step S201 and step S202; when sorting, the method with the higher correlation in the two methods is taken as the sorting basis.

[0088] Specifically, when the data center confirms the recommended related phenomena, for phenomena that have been technically determined and confirmed to have related logic, there is no need to perform the correlation calculation in the above steps S201 and S202; for unconfirmed reasoning phenomena, the data center sets a correlation threshold screening condition (the correlation is greater than or equal to 80%, and this threshold can be adjusted according to the actual calculation results), so as to recommend related phenomena on the PC software for related phenomena that meet the correlation threshold conditions. The data center collects the recommended related phenomenon results checked by the maintenance technician in the PC software, and technically determines that the correlation logic between the recommended related phenomena and the known phenomena is reasonable, thereby completing the confirmation of the correlation logic between each related phenomenon.

[0089] In this embodiment, the inference generation method of automobile fault diagnosis related phenomena also includes: step S300, the data center outputs related phenomenon recommendations corresponding to the user-side request based on the user-side request.

[0090] Specifically, the user end may be a user PC, and the user sends a request to the data center through the software on the PC.

[0091] Embodiment 2

[0092] Figure 6 1 is a schematic diagram of the structure of a device for generating reasoning of automobile fault diagnosis association phenomena in the first embodiment of the present invention. Figure 6 As shown, a vehicle fault diagnosis big data information extraction device provided in the second embodiment includes: a sampling data generation module 401, a diagnostic data generation module 402 and a data sending module 403. The sampling data generation module 401, based on the historical diagnostic database, cleans and classifies the historical fault diagnosis data in the historical diagnostic database to generate data to be sampled; the diagnostic data generation module 402 is used to sample the data to be sampled through a data processing tool to generate diagnostic data; and the data sending module 403 is used to submit the diagnostic data to the reference database for storage; wherein the data processing tool includes a sampling module, a phenomenon processing module and a cause processing module.

[0093] In this embodiment, the data center station counts the occurrence frequency of the fault cause of each fault phenomenon in the preset database, including: the data center station extracts the keywords used to describe the fault phenomenon and the fault cause in the preset database; creates corresponding phenomenon common words and cause common words that are common across brands according to the extracted fault phenomenon keywords and fault cause keywords; establishes the association logic between the fault phenomenon keywords and the fault cause keywords and the phenomenon common words and cause common words; determines the technical logic between the fault phenomenon and the fault cause; the data center station counts the occurrence frequency of each group of fault phenomena and fault causes in the preset database within the same product subsystem according to each group of phenomenon common words and cause common words created.

[0094] In this embodiment, the data center derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each correlation phenomenon, including: the data center generates a first correlation phenomenon based on a preset database inference; the data center generates a second correlation phenomenon based on a preset fault diagnosis and troubleshooting process inference; the data center performs correlation logic judgment based on the first correlation phenomenon and the second correlation phenomenon, and generates a correlation ranking of each correlation phenomenon.

[0095] In this embodiment, the data center generates the first associated phenomenon based on the preset database reasoning, including: the data center performs quantitative statistics on different fault phenomena with the same fault cause based on the information of the fault phenomenon and the fault cause extracted from the preset database; establishes a phenomenon correlation algorithm model to count the number of common fault causes between other fault phenomena and the current fault phenomenon; calculates the correlation between other fault phenomena and the current fault phenomenon according to the phenomenon correlation algorithm model to generate the first associated phenomenon.

[0096] In this embodiment, the data middle platform generates the second associated phenomenon based on the preset fault diagnosis and troubleshooting process reasoning, including: the data middle platform counts the general meanings of the DTC codes and the fault cause groups contained in the historically developed vehicle models; uses the phenomenon correlation algorithm model to calculate the correlation of DTC-type fault phenomena with the same cause and the same general meaning to generate the second associated phenomenon.

[0097] In this embodiment, the data center performs association logic judgment based on the first association phenomenon and the second association phenomenon, and generates an association ranking of each association phenomenon, including: comparing the association degrees of the association phenomena existing in the first association phenomenon and the second association phenomenon, and retaining association phenomena with high association degrees; retaining association phenomena that only exist in the first association phenomenon or the second association phenomenon; and generating an association ranking of each association phenomenon.

[0098] In this embodiment, the data center outputs a recommendation of related phenomena corresponding to the user-side request based on the user-side request.

[0099] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0100] In summary, the inference generation method and device for automobile fault diagnosis related phenomena of the present invention, through an inference model that combines theory with actual data, systematically sorts out the correlation logic between various fault phenomena, thereby achieving the purpose of assisting maintenance technicians in the actual fault diagnosis process to simultaneously check potential fault phenomena and improve maintenance quality, and also improves work efficiency and accuracy.

[0101] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for generating reasoning of automobile fault diagnosis association phenomena, characterized in that: The method is applied to a reasoning generation system for automobile fault diagnosis related phenomena, the system includes a data middle platform, and the reasoning generation method for automobile fault diagnosis related phenomena includes: The data center counts the occurrence frequency of the fault cause of each fault phenomenon in the preset database; The data center derives the correlation between different fault phenomena with the same fault cause, and generates a correlation ranking of each correlation phenomenon, specifically including: The data middle platform generates a first correlation phenomenon based on the preset database by inference, specifically including: The data center station counts the number of different fault phenomena with the same fault cause based on the information of the fault phenomenon and the fault cause extracted from the preset database; Establish a phenomenon correlation algorithm model to count the common fault causes between other fault phenomena and the current fault phenomenon; According to the phenomenon correlation algorithm model, calculating the correlation between other fault phenomena and the current fault phenomenon to generate a first correlation phenomenon; Among them, phenomenon correlation = (lower cause ~ phenomenon group occurrence number) / (higher cause ~ phenomenon group occurrence number) * 100%, when the number of common causes of the current fault phenomenon and other associated fault phenomena is n, the correlation algorithm model calculates n rounds, and the calculation result of the nth round is the correlation of the two fault phenomena; The data center generates a second associated phenomenon based on the preset fault diagnosis and troubleshooting process by inference, specifically including: The data center performs quantitative statistics on the general meaning and fault causes of each set of DTC codes included in historically developed vehicle models; The phenomenon correlation algorithm model is used to calculate the correlation of DTC-type fault phenomena with the same cause and the same general meaning, and generate the second correlation phenomenon; The data center performs association logic judgment based on the first association phenomenon and the second association phenomenon to generate an association ranking of the association phenomena, specifically including: Comparing the correlation degree between the correlation phenomena existing in the first correlation phenomenon and the second correlation phenomenon, and retaining the correlation phenomenon with a high correlation degree; Retaining the associated phenomenon that only exists in the first associated phenomenon or the second associated phenomenon; Generate a correlation ranking of each correlation phenomenon.

2. The inference generation method of automobile fault diagnosis related phenomena according to claim 1, characterized in that: The occurrence frequency of the fault causes of each fault phenomenon in the data center statistics preset database includes: The data center extracts keywords from the preset database for describing fault phenomena and fault causes; Creating corresponding phenomenon common words and cause common words that are common across brands respectively according to the extracted keywords of the fault phenomenon and the extracted keywords of the fault cause; The data center performs frequency statistics on the number of occurrences of each group of fault phenomena and fault causes in the preset database within the same product subsystem based on each group of common words for the phenomenon and common words for the cause created.

3. The inference generation method of automobile fault diagnosis related phenomena according to claim 1, characterized in that: Also includes: The data center outputs recommendations of related phenomena corresponding to the user-side request based on the user-side request.

4. A device for generating reasoning of automobile fault diagnosis related phenomena, characterized in that: The device is applied to a reasoning generation system for automobile fault diagnosis related phenomena, the system includes a data middle platform, and the reasoning generation device for automobile fault diagnosis related phenomena includes: A statistical module, used for the data center to count the occurrence frequency of the fault cause of each fault phenomenon in the preset database; and The derivation generation module is used for the data center to derive the correlation between different fault phenomena with the same fault cause and generate the correlation ranking of each related phenomenon, specifically including: The data middle platform generates a first correlation phenomenon based on the preset database by inference, specifically including: The data center station counts the number of different fault phenomena with the same fault cause based on the information of the fault phenomenon and the fault cause extracted from the preset database; Establish a phenomenon correlation algorithm model to count the common fault causes between other fault phenomena and the current fault phenomenon; According to the phenomenon correlation algorithm model, calculating the correlation between other fault phenomena and the current fault phenomenon to generate a first correlation phenomenon; Among them, phenomenon correlation = (lower cause ~ phenomenon group occurrence number) / (higher cause ~ phenomenon group occurrence number) * 100%, when the number of common causes of the current fault phenomenon and other associated fault phenomena is n, the correlation algorithm model calculates n rounds, and the calculation result of the nth round is the correlation of the two fault phenomena; The data center generates a second associated phenomenon based on the preset fault diagnosis and troubleshooting process by inference, specifically including: The data center performs quantitative statistics on the general meaning and fault causes of each set of DTC codes included in historically developed vehicle models; The phenomenon correlation algorithm model is used to calculate the correlation of DTC-type fault phenomena with the same cause and the same general meaning, and generate the second correlation phenomenon; The data center performs association logic judgment based on the first association phenomenon and the second association phenomenon to generate an association ranking of the association phenomena, specifically including: Comparing the correlation degree between the correlation phenomena existing in the first correlation phenomenon and the second correlation phenomenon, and retaining the correlation phenomenon with a high correlation degree; Retaining the associated phenomenon that only exists in the first associated phenomenon or the second associated phenomenon; Generate a correlation ranking of each correlation phenomenon.

5. The inference generation device for automobile fault diagnosis related phenomena according to claim 4, characterized in that: The occurrence frequency of the fault causes of each fault phenomenon in the data center statistics preset database includes: The data center extracts keywords from the preset database for describing fault phenomena and fault causes; Creating corresponding phenomenon common words and cause common words that are common across brands respectively according to the extracted keywords of the fault phenomenon and the extracted keywords of the fault cause; Establishing association logic between the keywords of the fault phenomenon and the keywords of the fault cause and the common words of the phenomenon and the common words of the cause; Determine the technical logic between the fault phenomenon and the fault cause; The data center performs frequency statistics on the number of occurrences of each group of fault phenomena and fault causes in the preset database within the same product subsystem based on each group of common words for the phenomenon and common words for the cause created.

Images