Vehicle Fault Diagnosis Method, Device and Medium

Abstract

The present application provides a vehicle fault diagnosis method, device and medium. By means of the event rules corresponding to each preset problem event, in the big data of vehicle usage obtained, historical fault records corresponding to each historical problem event are determined. For each type of problem event in each historical problem event, corresponding frequent item sets are determined according to the corresponding historical fault records, and then rule reference information corresponding to each association rule is determined. A diagnosis rule library corresponding to the current type of problem event is constructed through each rule reference information, realizing the rapid construction of the diagnosis rule library based on the historical usage data of the vehicle, solving the problems of low efficiency, poor integrity and high personnel requirements caused by manual sorting of rules in the prior art. Furthermore, for the obtained fault record to be diagnosed, the root cause event corresponding to the fault record to be diagnosed is determined through the corresponding diagnosis rule library, realizing the rapid determination of the fault cause and improving the determination efficiency of the fault cause.

Description

Technical Field

[0001] This application relates to the technical field of automobiles, and particularly to a vehicle fault diagnosis method, device and medium. Background Art

[0002] Currently, the determination of diagnostic rules in the prior art relies solely on manual sorting. However, this method has the following defects: First, in the trend of electrification and intelligence, the number of signals sent by a vehicle reaches thousands or tens of thousands, and considering the influence of different working conditions, the number of cause events is huge and complex. Relying solely on the method of manually sorting the diagnostic knowledge base rules has low efficiency and poor integrity; Second, for diagnostic rules across ECUs (Electronic Control Unit) or even cross-domain functions, system engineers who are familiar with the vehicle's overall function strategy are required to sort them out, which has high requirements for personnel capabilities and takes a long time; Third, the problem location logic of the manual experience diagnostic knowledge base, also known as the fault tree, is constructed based on known diagnostic experience and is usually updated based on complaint events rather than all fault information, which is prone to omissions of cause events and omission of the correlation relationship between cause events. Summary of the Invention

[0003] In view of the above defects or deficiencies in the prior art, this application aims to provide a vehicle fault diagnosis method, device and medium to solve the problems of low efficiency, poor integrity and high personnel requirements caused by manual sorting of rules in the prior art.

[0004] An embodiment of this application provides a vehicle fault diagnosis method, including:

[0005] Obtain big data on vehicle usage, and determine historical fault records corresponding to each historical problem event in the big data on vehicle usage according to the event rules corresponding to each preset problem event;

[0006] For each type of problem event among the historical problem events, determine the frequent item set corresponding to the current type of problem event according to the historical fault records corresponding to the current type of problem event, where the frequent item set is composed of at least one historical cause event in the historical fault records;

[0007] Determine rule reference information corresponding to each association rule according to the frequent item sets, and construct a diagnostic rule base corresponding to the current type of problem event based on the rule reference information corresponding to each association rule, where the association rule is composed of at least two frequent item sets and the item set association direction;

[0008] In the case of obtaining the fault record to be diagnosed, according to the diagnostic rule base corresponding to the problem event to be diagnosed in the fault record to be diagnosed, determine the root cause event corresponding to the fault record to be diagnosed.

[0009] Optionally, after obtaining the big data of vehicle usage, it further includes:

[0010] Preprocess the big data of vehicle usage, and update the big data of vehicle usage according to the preprocessing result, where the preprocessing includes at least one of data type conversion, empty field filling, field range adjustment, and fixed column processing;

[0011] Perform signal normalization processing on the big data of vehicle usage to obtain the big data of vehicle usage including signal fields corresponding to each signal, where the signal fields include signal time, vehicle identification, signal type, signal name, and signal value.

[0012] Optionally, before determining the historical fault records corresponding to each historical problem event in the big data of vehicle usage according to the event rules corresponding to each preset problem event, it further includes:

[0013] Determine the event fields corresponding to each preset fault event, where each of the preset fault events includes a preset problem event and a preset cause event, and the event fields include event type, event name, and event rule;

[0014] Correspondingly, determining the historical fault records corresponding to each historical problem event in the big data of vehicle usage according to the event rules corresponding to each preset problem event includes:

[0015] Based on the event rules corresponding to each of the preset problem events and each signal field in the big data of vehicle usage, determine the signals that match the event rules, and determine the signals that match the event rules and other signals associated with the signals as the historical fault records corresponding to the historical problem events.

[0016] Optionally, determining the frequent item sets corresponding to the current type of problem event according to each historical fault record corresponding to the current type of problem event includes:

[0017] According to the event rules corresponding to each preset cause event, determine each historical cause event in each historical fault record corresponding to the current type of problem event;

[0018] Construct each candidate item set based on each of the historical cause events, where the candidate item set includes at least one historical cause event;

[0019] Based on each historical fault record corresponding to the current type of problem event, determine the historical support degrees corresponding to each candidate item set, and determine frequent item sets from each candidate item set according to the historical support degrees corresponding to each candidate item set.

[0020] Optionally, the determining the rule reference information corresponding to each association rule according to each frequent item set includes:

[0021] Construct each association rule according to each frequent item set, where the association rule consists of at least two frequent item sets and the item set association direction;

[0022] Based on each historical fault record corresponding to the current type of problem event, determine the confidence degrees corresponding to each association rule, and determine the rule reference information based on the confidence degrees;

[0023] Among them, the determining the confidence degrees corresponding to each association rule based on each historical fault record corresponding to the current type of problem event includes:

[0024] For each association rule, based on the item set association direction of the association rule, determine the antecedent and the consequent in each frequent item set of the association rule;

[0025] According to the historical support degree corresponding to the antecedent and the historical support degree corresponding to the candidate item set composed of the antecedent and the consequent, determine the confidence degree corresponding to the association rule.

[0026] Optionally, the constructing the diagnostic rule base corresponding to the current type of problem event based on the rule reference information corresponding to each association rule includes:

[0027] Based on the confidence degrees corresponding to each association rule and the minimum confidence degree threshold, determine each target rule from each association rule, and construct a diagnostic rule base based on each target rule;

[0028] Obtain an empirical rule base, update the diagnostic rule base based on the empirical rule base, or obtain new fault records, and update the diagnostic rule base based on the new fault records.

[0029] Optionally, after updating the diagnostic rule base, it further includes at least one of the following:

[0030] Perform duplicate removal processing on each target rule in the diagnostic rule base;

[0031] Detect whether a circular rule group appears in each of the target rules. If so, based on the confidence levels and support counts corresponding to the target rules in the circular rule group, perform a deletion process on some of the target rules in the circular rule group, where the item set association directions of the target rules in the circular rule group form a circular direction;

[0032] Generate a diagnostic fault tree based on the diagnostic rule base, where the diagnostic fault tree includes each node and the logic gates connecting the nodes. Perform node trimming processing on the input nodes with the same output node and the same logic gate under different branches of the diagnostic fault tree.

[0033] Optionally, the determining the root cause event corresponding to the to-be-diagnosed fault record according to the diagnostic rule base corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record includes:

[0034] Determine each candidate cause event in the to-be-diagnosed fault record;

[0035] According to each of the candidate cause events, determine a current matching rule in the diagnostic rule base corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record, where each cause event in the current matching rule exists in each of the candidate cause events;

[0036] Determine the antecedent in the current matching rule as the root cause event corresponding to the to-be-diagnosed fault record.

[0037] An embodiment of the present application further provides an electronic device, where the electronic device includes:

[0038] A processor and a memory;

[0039] The processor is configured to execute the steps of the vehicle fault diagnosis method provided in any embodiment of the present application by calling a program or instruction stored in the memory.

[0040] An embodiment of the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores a program or instruction, and the program or instruction causes a computer to execute the steps of the vehicle fault diagnosis method provided in any embodiment of the present application.

[0041] In summary, the present application proposes a vehicle fault diagnosis method. By means of the event rules corresponding to each preset problem event, in the vehicle usage big data obtained, historical fault records corresponding to each historical problem event are determined. For each type of problem event among the historical problem events, according to the historical fault records corresponding to the current type of problem event, corresponding frequent item sets are determined. Furthermore, according to the frequent item sets, rule reference information corresponding to each association rule is determined. By means of each rule reference information, a diagnosis rule library corresponding to the current type of problem event is constructed, achieving the rapid construction of the diagnosis rule library based on the historical usage data of the vehicle, without the need for manual sorting, solving the problems of low efficiency, poor integrity, and high personnel requirements caused by manual sorting of rules in the prior art. Furthermore, for the obtained fault record to be diagnosed, by means of the diagnosis rule library corresponding to the problem event to be diagnosed in the fault record to be diagnosed, the root cause event corresponding to the fault record to be diagnosed is determined, achieving the rapid determination of the fault cause, without the need for manual rule search and screening, and improving the determination efficiency of the fault cause. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0043] Figure 1 is a flowchart of a vehicle fault diagnosis method provided by an embodiment of the present application;

[0044] Figure 2 is a schematic diagram of an empirical rule library corresponding to a problem event provided by an embodiment of the present application;

[0045] Figure 3 is a schematic diagram of a diagnostic fault tree provided by an embodiment of the present application;

[0046] Figure 4 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The following will further elaborate on the present application in conjunction with the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention and are not a limitation to the invention. Additionally, it should be noted that, for the sake of convenience of description, only the parts related to the invention are shown in the drawings.

[0048] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The following will detail the present application with reference to the drawings and embodiments.

[0049] As mentioned in the background art, in view of the problems in the prior art, the present application proposes a vehicle fault diagnosis method, which can be executed by a vehicle fault diagnosis device that can be integrated into electronic devices such as an Electronic Control Unit (ECU), a Vehicle Control Unit (VCU), a computer, or a server. Figure 1 It is a flowchart of a vehicle fault diagnosis method provided by an embodiment of the present application. Refer to Figure 1 The vehicle fault diagnosis method specifically includes:

[0050] S110. Obtain big data on vehicle usage, and determine historical fault records corresponding to each historical problem event in the big data on vehicle usage according to the event rules corresponding to each preset problem event.

[0051] Among them, the big data on vehicle usage can be historical usage data of each vehicle; for example, vehicle DTC (Diagnostic Trouble Code), CAN (Controller Area Network) signals, logs, etc. Exemplarily, the usage data of each vehicle can be obtained through the in-vehicle T-BOX of each vehicle to obtain the big data on vehicle usage.

[0052] In this embodiment, for the obtained big data on vehicle usage, considering that in fact there are some fields with default values or empty values, this situation may affect subsequent rule analysis. Therefore, the obtained big data on vehicle usage can also be subjected to data cleaning to reduce unnecessary data volume, improve the data quality of the big data on vehicle usage, and the subsequent data analysis efficiency, and reduce the data analysis overhead.

[0053] For example, in a specific implementation manner, after obtaining the big data on vehicle usage, it further includes: preprocessing the big data on vehicle usage, and updating the big data on vehicle usage according to the preprocessing result, where the preprocessing includes at least one of data type conversion, empty field filling, field range adjustment, and fixed column processing; performing signal normalization processing on the big data on vehicle usage to obtain the big data on vehicle usage including signal fields corresponding to each signal, where the signal fields include signal time, vehicle identifier, signal type, signal name, and signal value.

[0054] Among them, data type conversion can be converting the current data type of a field to a preset data type. For example, converting data of the double type to the string type. Filling empty fields can be filling fields with empty content according to a preset value. Adjusting the field range can be adjusting the numerical value in a field according to a preset numerical range. Specifically, for a field with a numerical value exceeding the preset numerical range, the numerical value can be adjusted to within the preset numerical range. For example, adjusting the numerical value in a field according to the upper and lower limits in the preset numerical range. Fixed column processing can be deleting fields in one or more columns, or adding fields in one or more columns.

[0055] Specifically, at least one of data type conversion, empty field filling, field range adjustment, and fixed column processing can be performed on the big data of vehicle usage. Further, signal normalization processing is performed on the big data of vehicle usage to uniformly describe the big data of vehicle usage with signals, that is, dividing the big data of vehicle usage into signal fields corresponding to each signal.

[0056] Among them, the signal can be an index, an alarm (including DTC code), a CAN signal, an Ethernet signal, a log parameter, etc. The corresponding signal field can include signal time, vehicle identification, signal type, signal name, and signal value. The signal time can be the time when the signal is generated, the vehicle identification can be the VIN (Vehicle Identification Number), and the signal type can be types such as index, alarm, CAN, Ethernet, log, etc. Exemplarily, Table 1 shows some results after a signal normalization process.

[0057] Table 1 Some Results after a Signal Normalization Process

[0058]

[0059] In the above embodiments, by preprocessing the big data of vehicle usage, the data quality of the big data of vehicle usage can be improved, thereby facilitating the subsequent data analysis efficiency and increasing the speed of rule analysis. By performing signal normalization processing on the big data of vehicle usage, it is convenient to screen each historical fault record in the big data of vehicle usage, and the screening efficiency of the historical fault record is improved.

[0060] After obtaining the big data of vehicle usage, further, based on the event rules corresponding to each preset problem event, the historical fault records corresponding to each historical problem event can be determined in the big data of vehicle usage.

[0061] Among them, the preset problem event can be an event describing a problem with a fault, such as "the air conditioner does not cool". The event rule corresponding to the preset problem event can be a rule for determining whether a preset problem event exists, which can be a precondition or signal abnormality, etc. For example, the event rule corresponding to "the air conditioner does not cool" can be: "The temperature drop in the vehicle (℃) in the previous 5 minutes" < 5 and the indicator "the temperature in the vehicle (℃) at the moment 5 minutes ago" > 25. It should be noted that the event rule can be a rule containing algorithms such as >, ≥, =,!=, <, ≤, etc., and can also support combinations according to logical gates such as AND, OR, and XOR.

[0062] Specifically, through the event rules corresponding to each preset problem event, each historical problem event in the vehicle usage big data can be determined, and then the historical fault records corresponding to each historical problem event can be screened out. Among them, the historical fault records can include signal time, signal value, etc.

[0063] In a specific implementation manner, before determining the historical fault records corresponding to each historical problem event in the vehicle usage big data according to the event rules corresponding to each preset problem event, it further includes:

[0064] Determine the event fields corresponding to each preset fault event, where each preset fault event includes a preset problem event and a preset cause event, and the event fields include event type, event name, and event rule;

[0065] Correspondingly, determining the historical fault records corresponding to each historical problem event in the vehicle usage big data according to the event rules corresponding to each preset problem event includes: based on the event rules corresponding to each preset problem event and each signal field in the vehicle usage big data, determining the signals that match the event rules, and determining the signals that match the event rules and other signals associated with the signals as the historical fault records corresponding to the historical problem events.

[0066] Among them, the preset fault event can be classified into a preset problem event and a preset cause event according to the type; a preset problem event can correspond to one or more preset cause events. Specifically, the corresponding event fields can be set for each preset fault event in advance.

[0067] In this embodiment, the event fields corresponding to the preset fault event can include event type, event name, and event rule. Among them, the event type can be a problem event or a cause event.

[0068] In addition to the event type, event name, and event rule, the event fields can also include event ID, delimited ECU, cumulative times, analysis recent days, single-day peak value, fault description, maintenance suggestion, vehicle series, and vehicle version number, etc.

[0069] Among them, the event ID can be the unique representation corresponding to a preset fault event; the delimiter ECU can be the only main responsible component of the cause event, used to delimit the problem attribution; the cumulative number can be the total number of times required to be determined as a real fault within the recent days time interval; the analysis of the recent days can be the inspection period for judging whether the fault event is a real fault; the single-day peak can be the number of times the fault event occurs in a day. When it reaches or exceeds the set value, it is determined as a real fault; the fault description can be the description of the phenomenon and impact of the fault event; the maintenance suggestion can be the troubleshooting and handling plan for the cause event, including remote handling and offline handling means, used to guide after-sales service personnel for maintenance; the vehicle series can represent the vehicle series applicable to the fault event; the vehicle version number can represent the vehicle version number applicable to the fault event. Exemplarily, as shown in Table 2, the event fields corresponding to the preset fault events are shown.

[0070] Table 2 Event Fields Corresponding to Preset Fault Events

[0071]

[0072] After setting the corresponding event fields for each preset fault event, for each signal value in the signal field corresponding to each signal in the vehicle usage big data, it can be identified according to the event rules in the event fields corresponding to each preset problem event. If the signal value matches the event rule corresponding to a preset problem event, it can be determined that the preset problem event is the historical problem event corresponding to the signal, and the signal and other signals associated with the signal are used as the historical fault records corresponding to the historical problem event. Among them, other signals associated with the signal can be other signals belonging to the same vehicle, or other signals belonging to the same vehicle and generated during one use process, etc.

[0073] It should be noted that after screening the historical fault records according to the event rules corresponding to each preset problem event, in this embodiment, the historical fault records can be further screened according to the single-day peak, cumulative number, or recent analysis days corresponding to each preset problem event, so as to eliminate the historical fault records corresponding to the historical problem events whose actual number does not exceed the cumulative number or the actual peak does not exceed the single-day peak.

[0074] By determining the event fields corresponding to each preset problem event and each preset cause event as described above, and then screening out the historical fault records corresponding to each historical problem event from the vehicle usage big data according to the event fields corresponding to each preset problem event, the rapid acquisition of fault information is realized, without manual screening, and further improves the analysis efficiency of the association rules.

[0075] S120. For each type of problem event in each historical problem event, determine the frequent item set corresponding to the current type of problem event according to each historical fault record corresponding to the current type of problem event, where the frequent item set is composed of at least one historical cause event in the historical fault record.

[0076] Specifically, all historical fault records can be classified according to historical problem events, and then each historical fault record corresponding to each type of problem event can be obtained. In this embodiment, each historical fault record corresponding to a type of problem event can form a transaction database corresponding to a transaction, that is, a set of transaction data, such as historical fault records of a type of historical problem event.

[0077] Furthermore, for each type of problem event, a frequent item set composed of at least one historical cause event can be determined according to each historical fault record corresponding to the current type of problem event. Among them, the frequent item set can be an item set with a historical support degree higher than a preset support degree. An item set can be a set composed of at least one item (i.e., a historical cause event), and an item set containing k items is called a k-item set; for example, {A} is a 1-item set, and {ABC} is a 3-item set.

[0078] In a specific implementation manner, determining the frequent item set corresponding to the current type of problem event according to each historical fault record corresponding to the current type of problem event includes: determining each historical cause event in each historical fault record corresponding to the current type of problem event according to the event rules corresponding to each preset cause event; constructing each candidate item set based on each historical cause event, where the candidate item set includes at least one historical cause event; determining the historical support degree corresponding to each candidate item set based on each historical fault record corresponding to the current type of problem event, and determining the frequent item set from each candidate item set according to the historical support degree corresponding to each candidate item set.

[0079] In this embodiment, each historical fault record can include one or more historical cause events. For example, historical fault record 1: ACDE; historical fault record 2: AC.

[0080] Specifically, each historical fault record corresponding to the current type of problem event can be matched according to the event rules in the event fields corresponding to each preset cause event, and each historical cause event that matches the event rules can be found. Furthermore, all historical cause events can be combined arbitrarily to construct each candidate item set including at least one historical cause event.

[0081] Furthermore, based on each historical fault record corresponding to the current type of problem event, the historical support degrees corresponding to each candidate item set can be counted. Among them, the historical support degree can be the frequency of the candidate item set appearing in each historical fault record corresponding to the current type of problem event, which is used to represent the minimum importance of the candidate item set in terms of statistics. For example, among 100 historical fault records corresponding to the current type of problem event, 3 records contain the candidate item set AC (such as AC, ACDE, ABC), and the support degree P(AC) of the candidate item set AC is 3%.

[0082] Specifically, the candidate item sets with historical support degrees greater than the preset support degree can be determined as frequent item sets to filter out the candidate item sets with historical support degrees less than the preset support degree. Among them, the preset support degree can be a predefined threshold for measuring the historical support degree, which is used to represent the minimum importance requirement of the item set in terms of statistics. For example, 0.1%.

[0083] Exemplarily, as shown in Table 3, the historical support degrees corresponding to each frequent item set are shown. Among them, the historical support count is the number of transactions of the frequent item set. For example, among 100 historical fault records corresponding to the current type of problem event, 3 records contain the historical cause event AC, that is, the historical cause event A & the historical cause event C, then the historical support count corresponding to the frequent item set AC is 3.

[0084] Table 3 Historical support degrees corresponding to each candidate item set

[0085] Frequent itemset k-itemset Historical support Historical support count {A} k-1 50％ 5 {B} k-1 50％ 5 {C} k-1 20％ 2 {D} k-1 40％ 4 {EF} k-2 40％ 4 {BEF} k-3 40％ 4

[0086] In this embodiment, the transaction database (that is, each historical fault record corresponding to the current type of problem event) can be scanned once to find the frequent K-1 item sets, denoted as L, and they are arranged in descending order of the historical support count. Based on L, the transaction database is scanned again to construct an FP (FP-growth) tree representing the association of item sets in the transaction database. The mining process of the FP tree is as follows: starting from the frequent pattern of length 1 (the initial suffix pattern), its conditional pattern base (a "sub-database" composed of the prefix path sets that appear together with the suffix pattern in the FP tree) is constructed; further, its (conditional) FP tree is constructed, and mining is recursively performed on this tree. Pattern growth is achieved by connecting the suffix pattern with the frequent patterns generated by the conditional FP tree.

[0087] The advantages of the FP-growth algorithm are as follows: it can significantly compress the size of the dataset to be searched. First, the database representing frequent item sets is compressed into a frequent pattern tree (FP-tree), which still retains the association information of the item sets. Then, this compressed database is divided into a group of conditional databases, each associated with a frequent item or pattern segment, and each conditional database is mined separately. For each "pattern segment", only the associated dataset needs to be examined. Therefore, as the "growth" of the patterns to be examined, this algorithm can significantly compress the size of the dataset to be searched.

[0088] Through the above implementation manners, the frequent item sets corresponding to all types of problem events can be determined, the rapid determination of the frequent item sets is realized, and then it is convenient to analyze the association rules according to the frequent item sets, and the rapid determination of the association rules is realized.

[0089] S130. Determine the rule reference information corresponding to each association rule according to each frequent item set, and construct a diagnosis rule library corresponding to the current type of problem event based on the rule reference information corresponding to each association rule.

[0090] Among them, the association rule is composed of at least two frequent item sets and the item set association direction; the item set association direction can be used to represent the association direction between each frequent item set. For example, the association rule C→A means that the frequent item set A is associated under the condition of the frequent item set C.

[0091] Specifically, each association rule can be constructed according to each frequent item set, and then the rule reference information corresponding to each association rule can be determined through the transaction database. Among them, the rule reference information can include the confidence level corresponding to the association rule and the historical support count. The confidence level is the probability corresponding to the association rule. For example, P(C|A) can be used to represent the probability that the frequent item set C is associated to obtain the frequent item set A, and P(C|A)=historical support degree P(AC) / historical support degree P(C), that is, it represents the probability that C is associated to obtain A.

[0092] For vehicle fault diagnosis, the probability that the cause event C is associated to obtain the cause event A is equal to the historical support degree of the frequent item set C & the frequent item set A / the historical support degree of the frequent item set C. Exemplarily, as shown in Table 4, a kind of rule reference information corresponding to each association rule is shown, and among them, the rule reference information can also include the confidence level name and the confidence level calculation formula.

[0093] Table 4 A kind of rule reference information corresponding to each association rule

[0094]

[0095] In a specific implementation manner, determining rule reference information corresponding to each association rule according to each frequent item set includes: constructing each association rule according to each frequent item set, where an association rule consists of at least two frequent item sets and an item set association direction; determining the confidence of each association rule based on each historical fault record corresponding to the current type of problem event, and determining the rule reference information based on the confidence; where determining the confidence of each association rule based on each historical fault record corresponding to the current type of problem event includes: for each association rule, determining an antecedent and a consequent in each frequent item set of the association rule based on the item set association direction of the association rule; determining the confidence of the association rule according to the historical support corresponding to the antecedent and the historical support corresponding to the candidate item set composed of the antecedent and the consequent.

[0096] That is, each association rule can be constructed according to each frequent item set, and then the confidence of each association rule can be determined according to the transaction database, and the confidence is used as the rule reference information.

[0097] Specifically, the antecedent and the consequent can be determined according to the item set association direction, where the antecedent is the frequent item set corresponding to the direction pointed out by the association rule, and the consequent is the frequent item set corresponding to the direction pointed to by the association rule. For example, for the association rule {A} -> {C}, {A} is the antecedent and {C} is the consequent. For vehicle fault diagnosis, it can be understood that the cause event is the antecedent and the problem event is the consequent.

[0098] Furthermore, the ratio of the historical support corresponding to the candidate item set composed of the antecedent and the consequent to the historical support corresponding to the antecedent can be used as the confidence of the association rule.

[0099] Through the above method, the rapid determination of the rule reference information corresponding to each association rule is realized, and then it is convenient to construct a diagnostic rule base according to each rule reference information, without manual determination, improving the determination efficiency of the diagnostic rule base.

[0100] After determining the rule reference information corresponding to each association rule, strong association rules can be screened out from the rule reference information, and then a diagnostic rule base containing strong association rules can be constructed. Through this method, a diagnostic rule base corresponding to each type of problem event can be constructed in turn. That is, by repeatedly executing S120 - S130, the construction of the diagnostic rule bases corresponding to all types of problem events can be realized.

[0101] Optionally, based on the rule reference information corresponding to each association rule, a diagnostic rule library corresponding to the current type of problem event is constructed, including: determining each target rule in each association rule based on the confidence level corresponding to each association rule and the minimum confidence threshold, and constructing a diagnostic rule library based on each target rule; obtaining an empirical rule library, and updating the diagnostic rule library based on the empirical rule library, or obtaining new fault records, and updating the diagnostic rule library based on the new fault records.

[0102] Among them, the minimum confidence threshold can be a predefined threshold for measuring confidence, which is used to represent the lowest reliability of the association rule. For example, 95%. Specifically, the association rules not less than the minimum confidence threshold can be determined as target rules, and then all target rules can be written into the constructed diagnostic rule library. It should be noted that when writing into the diagnostic rule library, each target rule can also be encoded in sequence, and then each target rule and the corresponding encoding can be automatically stored in the diagnostic rule library.

[0103] To further ensure the comprehensiveness of the diagnostic rule library, the diagnostic rule library can also be updated according to the empirical rule library. Among them, the empirical rule library is an artificial empirical rule library summarized according to the expert knowledge library or the actual after-sales problem handling, and is composed of the association rules between events. Exemplarily, see Figure 2 , Figure 2 which is a schematic diagram of an empirical rule library corresponding to a problem event provided in an embodiment of the present application.

[0104] For example, for the association rules in the empirical rule library, if they are all different from the target rules in the diagnostic rule library, they can be encoded and then added to the diagnostic rule library. It should be noted that before adding it to the diagnostic rule library, the association rules in the empirical rule library can be converted into the association rule format. For example, for the CD or-gate output A and the EF and-gate output B, the association rules C→A, D→A, and EF→B can be obtained.

[0105] In addition to updating the diagnostic rule library according to the empirical rule library, the diagnostic rule library can also be updated according to the newly obtained fault records in real time. Among them, the new fault records can be the fault records newly obtained within a set time period, such as the fault records within one day.

[0106] Specifically, after constructing the diagnostic rule base, new fault records of each day can be extracted, and the association rules can be extracted by the method provided in this embodiment, so as to obtain new strong association rules, and compare them with the target rules in the diagnostic rule base. If the new strong association rules are different from all the existing target rules in the library, the new strong association rules can be encoded and added to the diagnostic rule base, so as to establish a richer diagnostic rule base; if the new strong association rules are the same as the existing target rules in the library, the confidence level and historical support count of the rules can be updated.

[0107] Through the above method, the real-time update of the constructed diagnostic rule base is realized, and the integrity of the diagnostic rule base is further ensured.

[0108] Considering that there may be duplicate target rules in the diagnostic rule base, or target rules whose item set association directions form a loop, etc., post-processing can also be performed on the diagnostic rule base. Of course, the post-processing of the diagnostic rule base can be executed before updating the diagnostic rule base according to the empirical rule base or new fault records, or can be executed after updating the diagnostic rule base according to the empirical rule base or new fault records.

[0109] For example, after updating the diagnostic rule base, it further includes at least one of the following: removing duplicates from each target rule in the diagnostic rule base; detecting whether there is a loop rule group in each target rule, and if so, based on the confidence levels and support counts corresponding to the target rules in the loop rule group, removing some target rules in the loop rule group, where the item set association directions of the target rules in the loop rule group form a loop direction; generating a diagnostic fault tree based on the diagnostic rule base, where the diagnostic fault tree includes each node and the logic gates connecting the nodes, and performing node trimming processing on the input nodes with the same output node and the same logic gate under different branches of the diagnostic fault tree.

[0110] That is, duplicate target rules can be removed, or some target rules in the loop rule group can be removed, or a diagnostic fault tree can be generated and node correction processing can be performed on the diagnostic fault tree.

[0111] Specifically, the loop rule group can be composed of multiple target rules, and the item set association directions of the target rules therein form a loop direction. For example, the target rules C→A, A→B, B→C form a loop rule group. For the detected loop rule group, the target rules to be removed can be selected according to the confidence level and historical support count in the loop rule group.

[0112] For example, the target rule with the lowest confidence level or historical support count is excluded, or, based on the confidence level, historical support count, a preset first weight corresponding to the confidence level, and a preset second weight corresponding to the historical support count, a reference value is calculated, and then the target rule with the lowest reference value is excluded.

[0113] Specifically, the diagnostic fault tree generated according to the diagnostic rule base can be composed of each node and the logic gates connecting the nodes, for reference, see Figure 2 as shown. In this embodiment, each node can be determined according to the frequent item sets included in the target rule, and the logic gates connecting the nodes can be determined according to the item set association direction.

[0114] Among them, the node can be a historical cause event or a historical problem event, and the logic gate can be determined by the item set association direction corresponding to the target rule. Exemplarily, for C→A and D→A, the or gate exists between node C and node A, and between node D and node A; for EF→B, the and gate exists between node E and node B, and between node F and node B.

[0115] Furthermore, after constructing the diagnostic fault tree, according to the principle of the minimal cut set, each rule can be regularized and redundant rules can be excluded. That is, for different branches with the same output node and the same logic gate, they can be merged into one branch to implement node trimming processing. Exemplarily, as Figure 3 shown, Figure 3 is a schematic diagram of a diagnostic fault tree provided by an embodiment of the present application. Among them, for branch 1 and branch 2 with the same output node and the same logic gate, branch 1 and branch 2 can be merged into one branch, that is, C, D, G, and H are respectively connected to an A, and the logic gate is an or gate.

[0116] Through the above deduplication processing, excluding some target rules in the circular rule group, or performing node trimming processing on the diagnostic fault tree, the occurrence of duplicate rules, unreasonable rules, or redundant rules in the diagnostic rule base is avoided, thereby reducing the storage overhead of the diagnostic rule base, and improving the efficiency of locating the cause event in the diagnostic rule base.

[0117] S140. When the to-be-diagnosed fault record is obtained, according to the diagnostic rule base corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record, determine the root cause event corresponding to the to-be-diagnosed fault record.

[0118] Among them, the to-be-diagnosed fault record can be a fault record for which the fault cause needs to be determined. Specifically, according to the event rules of each preset problem event, the to-be-diagnosed problem event in the to-be-diagnosed fault record can be determined, and then in all diagnostic rule bases, the diagnostic rule base corresponding to the to-be-diagnosed problem event is queried, and the root cause event is located from this diagnostic rule base.

[0119] In a specific embodiment, according to the diagnostic rule base corresponding to the problem event to be diagnosed in the fault record to be diagnosed, determining the root cause event corresponding to the fault record to be diagnosed includes: determining each candidate cause event in the fault record to be diagnosed; according to each candidate cause event, determining the current matching rule in the diagnostic rule base corresponding to the problem event to be diagnosed in the fault record to be diagnosed, wherein each cause event in the current matching rule exists in each candidate cause event; and determining the antecedent in the current matching rule as the root cause event corresponding to the fault record to be diagnosed.

[0120] Specifically, all cause events in the fault record to be diagnosed, that is, each candidate cause event, can be determined according to the event rules corresponding to each preset cause event. Further, in the diagnostic rule base corresponding to the problem event to be diagnosed, the current matching rule corresponding to each candidate cause event is determined, that is, for all target rules in the diagnostic rule base, the target rule whose included all frequent itemsets exist in the candidate cause events can be used as the current matching rule.

[0121] Exemplarily, each candidate cause event in the fault record to be diagnosed is {ACE}, and the diagnostic rule base can include the following rules: C→A→problem event to be diagnosed, D→A→problem event to be diagnosed, E→B→problem event to be diagnosed, F→B→problem event to be diagnosed. Among them, for the target rule "C→A→problem event to be diagnosed", the cause events C and A therein both exist in {ACE}. Therefore, the target rule "C→A→problem event to be diagnosed" can be used as the current matching rule and output this current matching rule.

[0122] Further, after determining the current matching rule, the antecedent in the current matching rule can be used as the root cause event corresponding to the fault record to be diagnosed. Among them, the antecedent in the current matching rule can be the cause event at the bottom layer. Continuing with the above example, the root cause event is the cause event C; or, if the current matching rule is "EF→B→problem event to be diagnosed", then the root cause events are the cause event E and the cause event F.

[0123] In the above embodiment, by locating the current matching rule and then using the antecedent in the current matching rule as the root cause event, the rapid determination of the root cause of the fault is realized, and thus the rapid location of the vehicle fault point is realized. There is no need for manual screening, which greatly improves the fault location efficiency.

[0124] In this embodiment, the determined root cause event can facilitate the user to perform fault repair on the vehicle. Optionally, when outputting the root cause event, the delimited ECU in the event field corresponding to the root cause event can also be output to help the user perform vehicle repair.

[0125] It should be noted that, in this embodiment, big data on vehicle usage can be obtained through the cloud, and a diagnostic rule library corresponding to each type of problem event can be further constructed. Further, a fault record to be diagnosed can be received through the cloud, and then the corresponding root cause event can be fed back to achieve remote vehicle fault diagnosis.

[0126] The vehicle fault diagnosis method provided by the embodiment of the present application, through the event rules corresponding to each preset problem event, determines the historical fault records corresponding to each historical problem event in the obtained big data on vehicle usage. For each type of problem event among the historical problem events, according to the historical fault records corresponding to the current type of problem event, the corresponding frequent item sets are determined, and then the rule reference information corresponding to each association rule is determined according to the frequent item sets. The diagnostic rule library corresponding to the current type of problem event is constructed through each rule reference information, realizing the rapid construction of the diagnostic rule library based on the historical usage data of the vehicle, without manual sorting, solving the problems of low efficiency, poor integrity, and high personnel requirements caused by manual sorting of rules in the prior art. Furthermore, for the obtained fault record to be diagnosed, through the diagnostic rule library corresponding to the problem event to be diagnosed in the fault record to be diagnosed, the root cause event corresponding to the fault record to be diagnosed is determined, realizing the rapid determination of the fault cause, without manual rule search and screening, and improving the determination efficiency of the fault cause.

[0127] Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 4 shown, the electronic device 500 includes one or more processors 501 and a memory 502.

[0128] The processor 501 can be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and can control other components in the electronic device 500 to perform desired functions.

[0129] The memory 502 can include one or more computer program products, and the computer program products can include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory can include, for example, random access memory (RAM) and / or cache memory, etc. The non-volatile memory can include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions can be stored on the computer-readable storage medium, and the processor 501 can run the program instructions to implement the vehicle fault diagnosis method of any embodiment of the present application described above and / or other desired functions. Various contents such as initial external parameters and thresholds can also be stored in the computer-readable storage medium.

[0130] In one example, the electronic device 500 may further include: an input device 503 and an output device 504, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown). The input device 503 may include, for example, a keyboard, a mouse, and the like. The output device 504 may output various information to the outside, including warning prompt information, braking force, and the like. The output device 504 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, and the like.

[0131] Of course, for simplicity, Figure 4 only some of the components related to the present application in the electronic device 500 are shown, and components such as a bus, an input / output interface, and the like are omitted. In addition, according to specific application scenarios, the electronic device 500 may further include any other appropriate components.

[0132] In addition to the above methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, and when the computer program instructions are run by a processor, the processor is caused to execute the steps of the vehicle fault diagnosis method provided in any embodiment of the present application.

[0133] The computer program product may be written in any combination of one or more programming languages for programming code to perform the operations of the embodiments of the present application. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on a user computing device, partially on the user device, executed as a stand-alone software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0134] In addition, an embodiment of the present application may also be a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are run by a processor, the processor is caused to execute the steps of the vehicle fault diagnosis method provided in any embodiment of the present application.

[0135] The computer-readable storage medium may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0136] It should be noted that the terms used in this application are only for describing specific embodiments and do not limit the scope of this application. As shown in the specification and claims of this application, unless the context clearly indicates otherwise, words such as "a", "an", "one", and / or "the" do not specifically refer to the singular and may also include the plural. The term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, or device comprising the element.

[0137] It should also be noted that the orientation or positional relationship indicated by terms such as "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing this application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation of this application. Unless otherwise clearly specified and limited, terms such as "installed", "connected", "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0138] In this article, specific examples are used to illustrate the principles and implementation modes of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application. The above is only the preferred implementation mode of the present application. It should be noted that due to the limited nature of literal expression and the objectively infinite specific structures, for those of ordinary skill in the art, without departing from the principles of the present application, several improvements, refinements or changes can be made, or the above technical features can be combined in an appropriate manner; these improvements, refinements, changes or combinations, or directly applying the inventive concept and technical solution to other occasions without improvement, shall all be regarded as the protection scope of the present application.

Claims

1. A vehicle fault diagnosis method, characterized in that, Including: Obtain big data on vehicle usage, and according to the event rules corresponding to each preset problem event, determine the historical fault records corresponding to each historical problem event in the big data on vehicle usage. The historical fault records corresponding to the historical problem event include the signals corresponding to the historical problem event and other signals associated with the signals; For each type of problem event among the historical problem events, according to the historical fault records corresponding to the current type of problem event, determine the frequent item sets corresponding to the current type of problem event. The frequent item sets are composed of at least one historical cause event in the historical fault records, and the frequent item sets are item sets with a historical support degree higher than a preset support degree; Determine the rule reference information corresponding to each association rule according to the frequent item sets, and based on the rule reference information corresponding to each association rule, construct a diagnostic rule library corresponding to the current type of problem event. The association rule is composed of at least two of the frequent item sets and the item set association direction; When a to-be-diagnosed fault record is obtained, according to the diagnostic rule library corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record, determine the root cause event corresponding to the to-be-diagnosed fault record. The root cause event is the cause event at the bottom layer in the currently matched rule in the diagnostic rule library, and the currently matched rule is the target rule in which all frequent item sets exist in the to-be-diagnosed fault record.

2. The method according to claim 1, characterized in that After obtaining the big data on vehicle usage, it further includes: Perform preprocessing on the big data on vehicle usage, and update the big data on vehicle usage according to the preprocessing result. The preprocessing includes at least one of data type conversion, empty field filling, field range adjustment, and fixed column processing; Perform signal normalization processing on the big data on vehicle usage to obtain big data on vehicle usage including signal fields corresponding to each signal. The signal fields include signal time, vehicle identifier, signal type, signal name, and signal value.

3. The method according to claim 2, wherein Before determining the historical fault records corresponding to each historical problem event in the big data on vehicle usage according to the event rules corresponding to each preset problem event, it further includes: Determine the event fields corresponding to each preset fault event. Each preset fault event includes a preset problem event and a preset cause event, and the event fields include event type, event name, and event rule; Correspondingly, determining the historical fault records corresponding to each historical problem event in the big data on vehicle usage according to the event rules corresponding to each preset problem event includes: Based on the event rules corresponding to each preset problem event and the signal fields in the big data on vehicle usage, determine the signals that match the event rules, and determine the signals that match the event rules and other signals associated with the signals as the historical fault records corresponding to the historical problem event.

4. The method according to claim 3, wherein Determining the frequent item sets corresponding to the current type of problem event according to the historical fault records corresponding to the current type of problem event includes: Determine each historical cause event in each historical fault record corresponding to the current type of problem event according to the event rules corresponding to each preset cause event; Construct each candidate item set based on each of the historical cause events, where the candidate item set includes at least one historical cause event; Determine the historical support corresponding to each candidate item set based on each historical fault record corresponding to the current type of problem event, and determine the frequent item sets in each candidate item set according to the historical support corresponding to each candidate item set.

5. The method according to claim 1, characterized in that The determining the rule reference information corresponding to each association rule according to each of the frequent item sets includes: Construct each association rule according to each of the frequent item sets, where the association rule is composed of at least two of the frequent item sets and the item set association direction; Determine the confidence corresponding to each association rule based on each historical fault record corresponding to the current type of problem event, and determine the rule reference information based on the confidence; where the determining the confidence corresponding to each association rule based on each historical fault record corresponding to the current type of problem event includes: For each of the association rules, determine the antecedent and the consequent in each of the frequent item sets of the association rule based on the item set association direction of the association rule; Determine the confidence corresponding to the association rule according to the historical support corresponding to the antecedent and the historical support corresponding to the candidate item set composed of the antecedent and the consequent.

6. The method according to claim 5, characterized in that, The constructing the diagnostic rule base corresponding to the current type of problem event based on the rule reference information corresponding to each of the association rules includes: Determine each target rule in each of the association rules based on the confidence corresponding to each association rule and the minimum confidence threshold, and construct a diagnostic rule base based on each target rule; Obtain an empirical rule base, update the diagnostic rule base based on the empirical rule base, or obtain new fault records and update the diagnostic rule base based on the new fault records.

7. The method according to claim 6, characterized in that, After updating the diagnostic rule base, it further includes at least one of the following: Perform duplicate removal processing on each target rule in the diagnostic rule base; Detect whether there is a circular rule group in each of the target rules. If so, based on the confidence and support count corresponding to each target rule in the circular rule group, perform removal processing on some target rules in the circular rule group, where the item set association directions of each target rule in the circular rule group form a circular direction; Generate a diagnostic fault tree based on the diagnostic rule base, where the diagnostic fault tree includes each node and the logic gates connecting the nodes, and perform node trimming processing on each input node with the same output node and the same logic gate under different branches of the diagnostic fault tree.

8. The method according to claim 1, characterized in that, The determining the root cause event corresponding to the to-be-diagnosed fault record according to the diagnostic rule base corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record includes: Determine each candidate cause event in the to-be-diagnosed fault record; According to each of the to-be-selected cause events, determine a current matching rule in a diagnostic rule library corresponding to the to-be-diagnosed problem event in the to-be-diagnosed fault record, wherein each cause event in the current matching rule exists in each of the to-be-selected cause events; Determine the antecedent in the current matching rule as the root cause event corresponding to the to-be-diagnosed fault record.

9. An electronic device, characterized in that, The electronic device includes: A processor and a memory; The processor is configured to execute the steps of the vehicle fault diagnosis method according to any one of claims 1 to 8 by calling a program or instruction stored in the memory.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instruction, and the program or instruction causes a computer to execute the steps of the vehicle fault diagnosis method according to any one of claims 1 to 8.

Images