Vehicle failure prediction method and device
The vehicle failure prediction method identifies likely failing components by matching detected abnormal signals with past diagnostic data, improving predictive accuracy and enabling proactive maintenance.
Patent Information
- Application Number
- JP2023086571
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-26
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods, such as Patent Document 1, fail to predict irregular vehicle failures and cannot identify which component is likely to fail based on detected abnormality signals.
A vehicle failure prediction method that detects signs of failure, gathers predictive information, refers to diagnostic data for correspondence with past predictive information, and identifies components with a high degree of coincidence as failure candidates.
Enables the identification of likely failing components before actual failure occurs by analyzing past diagnostic data and matching abnormal signal patterns, enhancing predictive accuracy.
Smart Images

Figure 2024169811000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a method and device for identifying a part that is likely to fail when a symptom of a vehicle failure is detected. [Background technology]
[0002] In Patent Document 1, vehicles that belong to the same vehicle classification (model and first year of registration) and have similar driver driving behavior (degree of sway, average speed, distance traveled per ride, degree of sudden braking, etc.) are defined as similar vehicles, and a predicted date of recurrence of a failure of a specific vehicle (e.g. the vehicle itself) is obtained from the average number of occurrences of each type of failure in this group of similar vehicles, and if this predicted date of recurrence has already passed, it is determined that there is a possibility of a failure occurring. In other words, for failures that recur repeatedly, failure prediction is made based on the premise that there is a high probability that the same failure will recur at the same time in similar vehicles with similar driving behavior. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Patent No. 4338652 Summary of the Invention [Problem to be solved by the invention]
[0004] In Patent Document 1, the timing of failure recurrence is merely predicted statistically regardless of whether or not there is actually any sign of failure, and it cannot handle many failures that occur without regularity. Even if some sign of abnormality, such as an abnormal signal, is detected, it is not possible to predict which part is likely to fail based on this sign. [Means for solving the problem]
[0005] The vehicle failure prediction method according to the present invention comprises: Detects signs of failure in vehicle parts, Obtaining predictive information including a plurality of abnormal signals related to the predictive sign, Refer to diagnostic data that summarizes past symptom information from other vehicles and the corresponding repair and replacement parts that were replaced in response to the information, comparing the acquired predictive information with past predictive information in the diagnostic data to determine the degree of agreement; Parts for repair and replacement that have a high degree of match with the predictive information are identified as candidate faulty parts. Effect of the Invention
[0006] According to this invention, when a symptom of a failure is detected, past diagnostic data of other vehicles can be referred to and candidate faulty parts can be narrowed down based on the symptom information, making it possible to respond to the failure before it occurs. [Brief description of the drawings]
[0007] [Figure 1] 1 is a configuration explanatory diagram showing a first embodiment of a failure prediction system according to the present invention; [Diagram 2] FIG. 4 is an explanatory diagram showing data on the results of failure sign detection on the vehicle side in a table format. [Diagram 3] FIG. 4 is an explanatory diagram of a failure sign detection result database on the failed part identifying device side. [Figure 4] 4 is a flowchart showing the flow of processing in the first embodiment. [Diagram 5] FIG. 11 is a diagram illustrating the degree of coincidence of failure symptom information. [Figure 6] FIG. 13 is an explanatory diagram showing an example of a list of parts with high matching degrees. [Figure 7] FIG. 11 is a configuration explanatory diagram showing a second embodiment of the failure prediction system. [Figure 8] 10 is a flowchart showing the flow of processing according to a second embodiment. [Figure 9] FIG. 13 is a configuration explanatory diagram showing a third embodiment of the failure prediction system. [Figure 10] FIG. 13 is an explanatory diagram of a failure sign detection result database including a history of abnormal signals. [Figure 11] FIG. 4 is an explanatory diagram of a failure sign history database. [Figure 12]13 is a flowchart showing the flow of processing according to a third embodiment. [Figure 13] FIG. 13 is an explanatory diagram showing an example of a list of parts with high matching degrees. [Figure 14] FIG. 13 is a configuration explanatory diagram showing a fourth embodiment of the failure prediction system. [Figure 15] FIG. 4 is an explanatory diagram of a signal importance database. [Figure 16] 13 is a flowchart showing the flow of processing according to a fourth embodiment. [Figure 17] FIG. 13 is an explanatory diagram showing an example of a list of ranked parts. [Figure 18] FIG. 13 is a configuration explanatory diagram showing a fifth embodiment of the failure prediction system. [Figure 19] FIG. 4 is an explanatory diagram of a signal feature database. [Figure 20] 13 is a flowchart showing the flow of processing according to the fifth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0008] An embodiment of the present invention will be described in detail below with reference to the drawings. Fig. 1 is a configuration explanatory diagram showing a first embodiment of a failure prediction system according to the present invention. The failure prediction system is composed of a vehicle-side device 001 that is included in each vehicle (e.g., vehicle A), and a faulty part identification device 002 that is composed of a server or cloud computer installed at a dealer or the like.
[0009] The vehicle-side device 001 is a part of a so-called computer system of the vehicle, and includes a diagnostic device 003, a memory 004, a display unit 005, an ECU (Electronic Control Unit) 006, a communication unit 007, and a failure prediction unit 008. The diagnostic device 003 monitors various signals (also called vehicle data) obtained from each part of the vehicle, and detects abnormal signals. The failure prediction unit 008 detects that a part is about to fail based on one or more abnormal signals (or other information), stores the abnormal signals in the memory 004, and transmits the abnormal signals to the failed part identification device 002 via the communication unit 007. The vehicle is configured as a so-called connected car, and is constantly connected to, for example, the Internet via the communication unit 007. Using this communication function, various information is transmitted and received between the vehicle-side device 001 and the failed part identification device 002. The display unit 005 is, for example, a display provided in front of the driver's seat, and displays various information. The display unit 005 may be a mobile terminal (for example, a smartphone) of a user who owns the vehicle.
[0010] When the failure prediction unit 008 detects a failure symptom, the presence of the failure symptom may be displayed on the display unit 005 together with the transmission to the failed part identification device 002. Alternatively, instead of displaying the failure symptom alone, the failure prediction unit 008 may display it together with a candidate parts list, which will be described later.
[0011] In the failure prediction unit 008, when a sign of failure is detected, data of a failure sign detection result 050 is generated in a form as shown in FIG. 2, and this data is transmitted via the communication unit 007. As exemplified in FIG. 2, this data includes the vehicle identification number, the date and time when the sign was detected, and the content of the sign, that is, the content of the abnormal signal (in the illustrated example, three types of abnormal signals a, b, and c are included). Other additional information may be included. That is, when the failure prediction unit 008 detects that there is a sign of failure from some signal or information, it acquires the sign information including a plurality of abnormal signals related to this sign, and generates data as shown in FIG. 2 together with the date and time. Note that the sign may be detected based on the abnormal signals themselves, for example, a, b, and c, included in the sign information, or may be determined to be present when another signal or information indicates a sign of failure, and the abnormal signal at that time may be acquired.
[0012] The faulty part identifying device 002 includes a communication unit 009, a faulty part identifying unit 010, a fault sign consistency diagnosis unit 011, and a fault sign detection result database 012. Note that each of these units does not necessarily have to be configured by one computer, and can be configured by multiple servers or cloud computers.
[0013] The communication unit 009 communicates with the communication unit 007 on the vehicle side via an appropriate communication network, and when a sign of failure is detected on the vehicle side, it receives data of the failure sign detection result 050 shown in Fig. 2 from the vehicle side. When the candidate parts, which will be described later, are to be displayed on the vehicle side as well, the necessary information is transmitted to the vehicle side.
[0014] The failure sign detection result database 012 stores a large amount of diagnostic data that summarizes past symptom information and diagnostic results of other vehicles (which may include past data of the vehicle itself), such as failure sign detection result data 051, a portion of which is illustrated in FIG. 3. As illustrated in FIG. 3, the data 051 of the failure sign detection result database 012 includes information such as the vehicle identification number, the symptom detection content, which indicates what part is abnormal, the date and time when the symptom was detected, the content of the abnormality signal (in other words, the symptom information), the name of the part that was actually repaired and replaced, and the result after the repair and replacement (i.e., whether the defect was resolved). If the result after the repair and replacement is good, it means that the part that was repaired and replaced based on the symptom was appropriate. The data 051 of the failure sign detection result database 012 may include other additional information. For example, information such as the period (time or distance traveled) elapsed from the time when a symptom was detected until the actual repair and replacement, or the period (time or distance traveled) elapsed until the failure occurs if the failure is left unattended, may be added. New data is added to the failure symptom detection result database 012 every time a part is actually repaired or replaced.
[0015] The failure sign coincidence diagnosis unit 011 determines the degree of coincidence between the contents of the sign detected on the vehicle side and the contents of the sign included in the past data 051 in the failure sign detection result database 012. For example, if the contents of the sign obtained by the failure prediction unit 008 are a combination of abnormal signals "a, b, c", a combination pattern including the combination of abnormal signals "a, b, c" is determined to be a combination pattern with a high degree of coincidence in the past data 051 in the failure sign detection result database 012. In the example of FIG. 3, the combinations of abnormal signals "a, b, c" and "a, b, c, f" are determined to be a coincidence pattern with a high degree of coincidence. Note that the combination of "a, b, c" has a higher degree of coincidence than the combination of "a, b, c, f". A combination pattern that does not include "a, b, c" (partially missing) and a combination pattern that includes multiple abnormal signals other than "a, b, c" are determined to be a non-coincidence pattern with a low degree of coincidence. Note that in the present invention, the degree of coincidence required is set to any level.
[0016] As a specific example, the abnormal signals a, b, and c are, for example, the engine speed, the ignition timing, and the air-fuel ratio, respectively, and are determined to be abnormal by processing these constantly monitored signals (detection signals or internal signals) through an abnormality detection process of an appropriate algorithm (for example, statistical abnormality detection for time-series data). The abnormal signal f can be, for example, an air-fuel ratio correction coefficient.
[0017] The faulty part identification unit 010 identifies parts that have a high degree of match with the contents of the symptom based on the degree of match between the contents of the symptom detected on the vehicle side by the fault symptom coincidence diagnosis unit 011 and the contents of the symptom in the past data 051. The parts identified here become candidate faulty parts corresponding to the symptom. These candidate faulty parts may include multiple parts with a high degree of match. Note that it is desirable to exclude from the candidate faulty parts those parts that have been repaired and replaced based on the symptoms in the past data 051 but have not produced good results, even if the degree of match with the contents of the symptom is high.
[0018] Fig. 4 is a flowchart showing the flow of processing in the failure prediction system of the first embodiment. Note that the flowchart in Fig. 4 does not distinguish between the processing performed by the vehicle-side device 001 and the processing performed by the failed part identification device 002, and shows each step in the order of the overall processing flow.
[0019] In step 101, failure prediction (detection of a sign of failure) is performed based on vehicle data by the diagnosis device 003 and the failure prediction unit 008 of the vehicle-side device 001. In step 102, data related to the failure prediction detected here (failure sign detection result 050) is transmitted from the vehicle-side device 001 to the failed part identification device 002.
[0020] In step 103, data 051 stored in the failure sign detection result database 012, that is, information on failure signs of other vehicles, is referred to. In step 104, the degree of agreement between the failure sign detection result 050 sent from the vehicle-side device 001 and the information on failure signs of other vehicles obtained from the failure sign detection result database 012 is diagnosed. As described above, a combination pattern of multiple types of abnormal signals is used to diagnose the degree of agreement. For example, as shown in FIG. 5, if the contents of the detected sign are a combination of abnormal signals "a, b, c", the combination of abnormal signals "a, b, c" or "a, b, c, f" is a matching pattern with a high degree of agreement, and a combination pattern in which the signals are different, such as "g, f" or "s, t", or which includes multiple other abnormal signals, such as "a, b, c, f, g, h", is a non-matching pattern with a low degree of agreement.
[0021] Then, in step 105, the parts with a high degree of match are narrowed down to be faulty part candidates. There may be multiple candidates here. Finally, in step 106, the narrowed down list of faulty part candidates is presented as shown in list 053 in FIG. 6. For example, in the example in FIG. 6, two parts, a plug and an injector, are extracted as faulty part candidates with a high degree of match for the combination of abnormal signals "a, b, c."
[0022] The data extracted as defective part candidates is displayed, for example, on the display unit 005 or the user's mobile terminal, etc. Furthermore, the data is stored in a vehicle management computer in the repair department of the dealer that maintains and manages the vehicle, or in the memory 004 of the vehicle.
[0023] In this manner, in the above embodiment, since the candidate faulty parts are identified based on the degree of coincidence of the combination patterns of multiple types of abnormal signals, it is possible to identify the parts with a high probability. Furthermore, by combining multiple types of abnormal signals, even if the threshold value for determining that each signal is abnormal is set relatively low, the possibility of erroneously determining that a normal part is faulty is reduced.
[0024] Next, a second embodiment of the present invention will be described with reference to Fig. 7 and Fig. 8. The following mainly describes the differences from the first embodiment. The second embodiment differs from the first embodiment in that failure prediction is performed not by the vehicle side device 001 but by a failed part identification device 002 managed by a dealer or the like.
[0025] As shown in FIG. 7, the failure prediction system of the second embodiment is composed of a vehicle-side device 001 installed in each vehicle, and a failed part identification device 002 consisting of a server or cloud computer installed in a dealer or the like.
[0026] As in the first embodiment, the vehicle-side device 001 includes a diagnostic device 003, a memory 004, a display unit 005, an ECU 006, and a communication unit 007. However, the vehicle-side device 001 does not include a failure prediction unit 008.
[0027] The faulty part identification device 002 is basically configured to include a communication unit 009, a faulty part identification unit 010, a fault sign coincidence diagnosis unit 011, and a fault sign detection result database 012, similarly to the first embodiment. In the second embodiment, a fault prediction unit 008 is configured as a part of the faulty part identification device 002. The fault prediction unit 008 detects a sign of a fault in a part based on one or more abnormal signals sent from the diagnosis device 003 of the vehicle-side device 001 via the communication units 007 and 009. When the fault prediction unit 008 detects a fault sign, it acquires predictive information including a plurality of abnormal signals, and thereby generates data of a fault sign detection result 050 as shown in FIG. 2. This data is basically the same as that of the first embodiment, and includes, for example, the vehicle identification number, the date and time when the sign was detected, and the contents of the abnormal signal (including three abnormal signals a, b, and c in the illustrated example), as shown in FIG. 2. It may also contain other additional information.
[0028] 8 is a flowchart showing the process flow of the failure prediction system of the second embodiment. The basic process flow is the same as that of the first embodiment, but first, as shown in step 107, vehicle data is transmitted from the vehicle-side device 001 to the failed part identification device 002. In step 101, failure prediction (detection of a sign) is performed on the failed part identification device 002 side based on the vehicle data transmitted from the vehicle-side device 001.
[0029] In step 103, the data 051 stored in the failure sign detection result database 012, i.e., information on failure signs of other vehicles, is referenced. In step 104, the degree of match between the failure sign detection result 050 output by the failure prediction unit 008 and the information on failure signs of other vehicles obtained from the failure sign detection result database 012 is diagnosed. As described above, a combination pattern of a plurality of abnormal signals is used to diagnose the degree of match.
[0030] Then, in step 105, the components with the highest degree of match are narrowed down to be faulty component candidates. There may be multiple candidates here. Finally, in step 106, the narrowed down faulty component candidates are presented.
[0031] According to the second embodiment, the vehicle does not require the failure prediction unit 008, so the present invention can be applied to a larger number of vehicles, including existing vehicles.
[0032] Next, a third embodiment of the present invention will be described with reference to Figures 9 to 13. The third embodiment differs from the first and second embodiments in that candidate faulty parts are identified in consideration of the history of past signs of faults in the vehicle.
[0033] As shown in FIG. 9, the failure prediction system of the third embodiment is composed of a vehicle-side device 001 installed in each vehicle, and a failed part identification device 002 consisting of a server or cloud computer installed in a dealer or the like.
[0034] The vehicle-side device 001 includes a diagnostic device 003, a memory 004, a display unit 005, an ECU 006, and a communication unit 007, similarly to the second embodiment.
[0035] Similarly to the second embodiment, the failed part identifying device 002 is basically configured to include a failure prediction unit 008, a communication unit 009, a failed part identifying unit 010, a failure sign consistency diagnosis unit 011, and a failure sign detection result database 012. In the third embodiment, the device further includes a detection information acquisition unit 013 that acquires information on the failure signs (information on abnormal signals) detected by the failure prediction unit 008, and a failure sign history database 014 that accumulates information on the failure signs of the vehicle.
[0036] FIG. 11 shows an example of the failure sign history stored in the failure sign history database 014, and stores data 061 including the date and time of failure prediction for the vehicle and the abnormal signal detected at that time.
[0037] Similarly, the failure sign detection result database 012 in the third embodiment stores failure sign detection result data 062 including the history of several abnormal signals in the past for each vehicle, as shown in FIG.
[0038] FIG. 12 is a flowchart showing the process flow of the failure prediction system of the third embodiment. The basic process flow is the same as that of the second embodiment. First, in step 107, vehicle data is transmitted from the vehicle-side device 001 to the failed part identification device 002. In step 101, the failed part identification device 002 performs failure prediction (detection of a sign) based on the vehicle data transmitted from the vehicle-side device 001. Here, data of the failure sign detection result 050 as exemplified in FIG. 2 is generated. In step 110, data of the current failure sign detection result 050 detected in step 101 is added to data 061 of the failure sign history database 014 exemplified in FIG. 11. In other words, the current failure sign detection result 050 is stored. The failure sign history database 014 created in this way is a database related to failure signs of the vehicle itself.
[0039] In step 103, information regarding the signs of failure of other vehicles is referenced from the signs of failure detection result data 062 including past history shown in Fig. 10. As described above, the signs of failure detection result data 062 in this embodiment basically corresponds to the data 051 (Fig. 3) in the signs of failure detection result database 012 in the first embodiment, but also includes the history of several abnormal signals in the past for each vehicle.
[0040] In the next step 109, the degree of agreement between the failure sign information in the failure sign detection result data 062 and the failure sign history in the data 061 of the failure sign history database 014 related to the vehicle is diagnosed. In this agreement diagnosis, a combination pattern of abnormal signals including past history is used. For example, if the combination pattern of abnormal signals in the failure sign history data 061 of the vehicle is "a, b, c" for the first time, "a, b" for the second time, and "a, b, c" for the third time (see FIG. 11), the combination pattern of "a, b, c" and the combination pattern of "a, b" are determined as matching patterns among the combination patterns including the abnormal signal history in the failure sign detection result data 062 of the other vehicle. On the other hand, a combination pattern that includes signals other than a, b, and c, such as "a, b, c, f," or a combination pattern that does not include a and b is determined as a non-matching pattern. In the example of FIG. 10, the part surrounded by a thick solid line is a matching pattern.
[0041] Then, in step 105, the parts with a high degree of match are narrowed down to be faulty part candidates. There may be multiple candidates here. Finally, in step 106, the narrowed down faulty part candidates are presented. For example, in the example of list 053 shown in FIG. 13, a plug is extracted as a faulty part candidate with a high degree of match after considering the past history.
[0042] In this way, in the third embodiment, since candidate faulty parts are identified taking into account the history of the failure symptoms, it is possible to identify the candidate faulty parts with higher accuracy.
[0043] Next, a fourth embodiment of the present invention will be described with reference to Fig. 14 to Fig. 17. The fourth embodiment differs from the first and second embodiments in that the importance of an abnormal signal is determined for each component, and the coincidence of predictive information is determined taking into account this importance.
[0044] As shown in FIG. 14, the failure prediction system of the fourth embodiment is composed of a vehicle-side device 001 installed in each vehicle, and a failed part identification device 002 consisting of a server or cloud computer installed in a dealer or the like.
[0045] The vehicle-side device 001 includes a diagnostic device 003, a memory 004, a display unit 005, an ECU 006, and a communication unit 007, similarly to the second embodiment.
[0046] The faulty part identification device 002 is basically configured to include a fault prediction unit 008, a communication unit 009, a faulty part identification unit 010, a fault sign coincidence diagnosis unit 011, and a fault sign detection result database 012, as in the second embodiment. In the fourth embodiment, the device further includes a signal importance ranking unit 015 that determines the order of importance of abnormal signals for each part, and a signal importance database 016 that accumulates information on the signal importance. The signal importance database 016 stores data of a signal importance list 063 that summarizes the abnormal signal patterns and the importance of each abnormal signal for each part, as illustrated in FIG. 15, for example. The signal importance can be set in advance based on knowledge about each part (design information, test results, etc.). Alternatively, the priority, that is, the importance of the abnormal signal may be obtained from a large amount of fault sign detection result data 051 (see FIG. 3) that is sequentially stored in the fault sign detection result database 012. Even when the failure mechanism of a component is not clear, the importance of the signal can be determined from past data.
[0047] Fig. 16 is a flowchart showing the process flow of the failure prediction system of the fourth embodiment. The basic process flow is the same as that of the second embodiment, and first, in step 107, vehicle data is transmitted from the vehicle-side device 001 to the failed part identification device 002. In step 101, failure prediction (detection of a sign of failure) is performed on the failed part identification device 002 side based on the vehicle data transmitted from the vehicle-side device 001. Here, data of the failure sign detection result 050 as shown in Fig. 2 is generated.
[0048] In step 103, information on signs of failure of other vehicles is referenced from the data 051 (FIG. 3) in the failure sign detection result database 012. In step 104, the degree of agreement between the failure sign detection result 050 of the vehicle itself and the information on signs of failure of other vehicles obtained from the failure sign detection result database 012 is diagnosed. As described above, a combination pattern of a plurality of abnormal signals is used to diagnose the degree of agreement.
[0049] In step 111, the priority of the abnormal signals is assigned using information in the signal importance list 063 (FIG. 15) in the signal importance database 016, and the number of parts with high matching degrees (faulty part candidates) is narrowed down. Finally, in step 106, the narrowed-down list of faulty part candidates is presented. For example, in the example of the list 053 shown in FIG. 17, the plug is presented in first place, and the injector is presented in second place, with the plug ranked first and the injector ranked second.
[0050] In the example of Fig. 15, in the signal importance list 063, for example, when the content of the precursor detection is a plug, the combination patterns of abnormal signals are "a, b, c" and "a, b", and the importance of the abnormal signals is recorded as "a, b>c". In other words, the importance of signal c is low. On the other hand, when the content of the precursor detection is an injector, the combination patterns of abnormal signals are "a, b, c, f" and "a, b, f", and the importance of the abnormal signals is recorded as "a, f>b>c".
[0051] Therefore, in this example, when the combination of abnormal signals that are detected as a sign of failure is "a, b, c," since signal f, which is a signal with high importance for the injector, is not included, when narrowing down the candidate faulty parts, it is determined that the degree of match with the injector is lower than that with the plug (see FIG. 17).
[0052] In this manner, in the fourth embodiment, since the importance of the abnormal signal is taken into consideration to identify a candidate faulty part, it is possible to identify the faulty part with higher accuracy and to easily rank a plurality of candidate faulty parts.
[0053] Next, a fifth embodiment of the present invention will be described with reference to Fig. 18 to Fig. 20. The fifth embodiment differs from the first and second embodiments in that the degree of coincidence of the predictive information is obtained by taking into consideration information on the feature amount of the abnormal signal of each component.
[0054] As shown in FIG. 18, the failure prediction system of the fifth embodiment is composed of a vehicle-side device 001 installed in each vehicle, and a failed part identification device 002 consisting of a server or cloud computer installed at a dealer or the like.
[0055] The vehicle-side device 001 includes a diagnostic device 003, a memory 004, a display unit 005, an ECU 006, and a communication unit 007, similarly to the second embodiment.
[0056] The faulty part identification device 002 is basically configured to include a fault prediction unit 008, a communication unit 009, a faulty part identification unit 010, a fault sign coincidence diagnosis unit 011, and a fault sign detection result database 012, as in the second embodiment. In the fifth embodiment, the device further includes a signal feature diagnosis unit 017 for diagnosing what the feature of an abnormal signal is for each part, and a signal feature database 018 for accumulating information on the signal feature. The signal feature database 018 stores data of a signal feature list 065 that summarizes the abnormal signals of each part and the feature of the abnormal signals, as illustrated in FIG. 19, for example. The signal feature list 065 can be created in advance based on knowledge (design information, test results, etc.) about each part. Alternatively, the abnormal signal and its feature may be obtained from a large amount of fault sign detection result data 051 (see FIG. 3) that is sequentially stored in the fault sign detection result database 012.
[0057] Fig. 20 is a flowchart showing the process flow of the failure prediction system of the fifth embodiment. The basic process flow is the same as that of the second embodiment, and first, in step 107, vehicle data is transmitted from the vehicle-side device 001 to the failed part identification device 002. In step 101, failure prediction (detection of a sign of failure) is performed on the failed part identification device 002 side based on the vehicle data transmitted from the vehicle-side device 001. Here, data of the failure sign detection result 050 as exemplified in Fig. 2 is generated.
[0058] In step 103, information on signs of failure of other vehicles is referenced from the data 051 (FIG. 3) in the failure sign detection result database 012. In step 104, the degree of agreement between the failure sign detection result 050 of the vehicle itself and the information on signs of failure of other vehicles obtained from the failure sign detection result database 012 is diagnosed. As described above, a combination pattern of a plurality of abnormal signals is used to diagnose the degree of agreement.
[0059] In step 112, the abnormal signals are prioritized using the information of the signal feature list 065 in the signal feature database 018, and the number of parts with high matching scores (faulty part candidates) is narrowed down. Finally, in step 106, the narrowed-down list of faulty part candidates is presented.
[0060] In the example of FIG. 19 , the signal feature list 065 records, for example, that when the precursor detection content is a plug, the abnormal signals are a, b, and c, and that the feature of signal a is the kurtosis of 10 minutes of data, the feature of signal b is the variance of 10 minutes of data, and the feature of signal c is the difference value of 1 minute of data.
[0061] On the other hand, when the predictive detection content is an injector, the abnormal signals are a, b, c, and f. It is recorded that the feature of signal a is the kurtosis of 10 minutes of data, the feature of signal b is the variance of 10 minutes of data, the feature of signal c is the average value of 1 minute of data, and the feature of signal f is the variance of 10 minutes of data.
[0062] Therefore, for example, the feature amount of the detected abnormal signal c of a failure sign is the difference value of the data for one minute for plugs and the average value of the data for one minute for injectors, which are different from each other, so by focusing on the abnormal signal c, it is possible to determine the degree of match with higher accuracy. In other words, by diagnosing the degree of match using the type of signal that has a high contribution to the classification of anomalies among the signals in which an abnormality has been detected, the accuracy of identifying candidate faulty parts is increased. [Explanation of symbols]
[0063] 001...Vehicle side device 002…Faulty part identification device 003...Diagnostic equipment 005...Display section 007…Communications Department 008…Failure prediction department 009…Communications Department 010…Faulty part identification section 011…Failure prediction consistency diagnosis section 012…Failure sign detection result database 013…Detection information acquisition unit 014…Fault prediction history database 015…Signal Importance Order Section 016…Signal Importance Database 017…Signal feature diagnostic section 018…Signal feature database
Claims
1. Detects signs of failure in vehicle parts, Obtaining predictive information including a plurality of abnormal signals related to the predictive sign, Refer to diagnostic data that summarizes past symptom information from other vehicles and the corresponding repair and replacement parts that were replaced in response to the information, comparing the acquired predictive information with past predictive information in the diagnostic data to determine the degree of agreement; Identifying parts to be repaired or replaced that have a high degree of coincidence with the predictive information as candidates for faulty parts; A method for predicting vehicle failure.
2. The diagnostic data includes past predictive information and corresponding repair / replacement parts that have been repaired or replaced with good results. The vehicle failure prediction method according to claim 1.
3. determining a degree of agreement between the types of the plurality of abnormal signals included in the acquired abnormality information and the types of abnormality signals included in the past abnormality information of the other vehicle in the diagnostic data; The vehicle failure prediction method according to claim 1.
4. The importance of multiple abnormal signals is determined for each component. Considering the importance of each signal, the degree of agreement between the acquired predictive information and past predictive information is calculated. The vehicle failure prediction method according to claim 1.
5. The importance level is set based on accumulated diagnostic data. The vehicle failure prediction method according to claim 4.
6. When the feature amount of the abnormal signal is different for each of the plurality of components, the abnormal signal is determined as an abnormal signal having a high contribution to the abnormality classification; The abnormal signal with the highest contribution is given priority, and the degree of agreement between the acquired predictive information and past predictive information is calculated. The vehicle failure prediction method according to claim 1.
7. The feature amount is set based on accumulated diagnostic data. The vehicle failure prediction method according to claim 6.
8. The system detects signs of failure in vehicle parts on the vehicle side, The vehicle acquires the warning information and sends it to a server. This server side identifies possible faulty parts. The vehicle failure prediction method according to claim 1.
9. extracting a plurality of candidate faulty parts having a relatively high degree of coincidence based on a degree of coincidence between the obtained predictive information and past predictive information in the diagnostic data; The vehicle failure prediction method according to claim 1.
10. Present multiple faulty part candidates in order of their matching score. The vehicle failure prediction method according to claim 9.
11. a failure prediction unit that detects a sign of a failure in a vehicle component and acquires predictive information including a plurality of abnormality signals related to the sign; A database in which diagnostic data is accumulated that summarizes the correspondence between past symptom information of other vehicles and repair / replacement parts that have been repaired and replaced in response to the past symptom information; a coincidence diagnosis unit for comparing the acquired predictive information with past predictive information in the database to diagnose the coincidence; A part identification unit that refers to the database and identifies repair / replacement parts that have a high degree of coincidence with the predictive information as candidates for faulty parts; A vehicle failure prediction device comprising:
Citation Information
Patent Citations
Vehicle management device
JP4338652B2
Cited By
Frame assemblies for optical fiber distribution elements
US12306451B2
Fiber optic cable management systems and methods
US12339511B2
Optical fiber distribution system
US12422640B2
Fiber optic adapter block
US12461317B2
Telecommunications module arrangements
US12468106B2
