CN115660640A - Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product - Google Patents
Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product Download PDFInfo
- Publication number
- CN115660640A CN115660640A CN202211296026.6A CN202211296026A CN115660640A CN 115660640 A CN115660640 A CN 115660640A CN 202211296026 A CN202211296026 A CN 202211296026A CN 115660640 A CN115660640 A CN 115660640A
- Authority
- CN
- China
- Prior art keywords
- data
- time
- fault
- automobile
- fault diagnosis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention provides an intelligent automobile part fault diagnosis modeling method and an electronic equipment program product, which comprises the steps of obtaining automobile driving history big data and maintenance record data, carrying out data aggregation and cleaning on the history data of the whole life cycle of automobile parts, carrying out data segmentation and marking by taking one-time start-stop in the automobile driving process as a time sliding window, and constructing a fault diagnosis model based on an LSTM time sequence model and a neural network model. According to the invention, the fault detection and prediction of the parts when the automobile runs are realized by using algorithms such as artificial intelligence and the like in a data-driven manner, the faults of the parts which cannot be sensed by using the sensor in the running process of the automobile can be diagnosed and predicted, the early warning of the faults is realized, the accuracy and the timeliness of fault maintenance of a 4S shop are improved, and the fault diagnosis efficiency is improved.
Description
Technical Field
The invention belongs to the field of automobile fault diagnosis and maintenance, and particularly relates to a fault diagnosis and modeling method for automobile parts.
Background
The automobile industry enters the era of intellectualization, electromotion and automatic driving, the integration level and the automation degree of the whole automobile parts are higher and higher, and the requirements on the running state and the health monitoring of the automobile parts are also higher and higher. The traditional automobile part fault diagnosis mainly depends on professional maintenance personnel to use professional equipment to carry out one-by-one troubleshooting and detection on fault sources, the maintenance personnel can solve most of small problems, but the problems of difficulty and need of technical expert guidance are inevitable, and technical experts of a whole automobile factory need to be subjected to on-site diagnosis at the moment, so that the time and the labor are consumed; meanwhile, the device can only perform off-line detection after a fault occurs, and cannot perform early warning on the fault. At present, in the large environment of automobile intellectualization, fault diagnosis of automobile parts becomes possible by using an artificial intelligence technology.
Patent document 202111446107.5 discloses a configurable fault diagnosis method and a fault diagnosis system based on real-time vehicle conditions, which dynamically configure basic fault detection items corresponding to vehicle information and vehicle condition configuration fault detection items corresponding to real-time vehicle condition data according to the vehicle condition information and the real-time vehicle condition data of a vehicle to be detected, and then perform fault diagnosis and detection according to fault diagnosis and detection rules corresponding to the configured fault detection items.
Patent document 201310541642.8 discloses a vehicle-mounted status real-time diagnosis system and a diagnosis method thereof, which perform vehicle fault diagnosis using data of a vehicle OBD fault code, are executed only by existing detection types and detection methods of vehicle OBD fault diagnosis detection, and cannot detect parts not included in the OBD fault diagnosis detection types.
Therefore, the problems that the fault detection, diagnosis time consumption, labor consumption and diagnosis precision of parts are not high, and the fault which cannot be sensed by a sensor in the running process of the automobile still exists in the automobile maintenance process cannot be well solved in the prior art.
Disclosure of Invention
The invention provides an intelligent automobile part fault diagnosis and modeling method aiming at the defects in the prior art, and the method is based on data such as vehicle type data, part data, maintenance records, historical big data in the vehicle running process, user driving behavior data and the like, realizes fault detection and prediction of parts during the vehicle running by utilizing algorithms such as artificial intelligence and the like in a data driving mode, can diagnose and predict the part fault which cannot be sensed by a sensor in the vehicle running process, realizes early fault warning, improves the fault maintenance accuracy and timeliness of a 4S store, improves the fault diagnosis efficiency, avoids misdiagnosis, mischange and mischange of the parts, and reduces fault complaints of users.
Based on the method, the invention also provides intelligent electronic equipment and a computer program product for the fault diagnosis and modeling of the automobile parts.
In order to solve the technical problem in the invention, the adopted technical scheme is as follows:
the invention provides an intelligent automobile part fault diagnosis and modeling method in a first aspect, which mainly comprises the following steps:
step 1: obtaining automobile driving history big data and maintenance record data
The method comprises the steps of determining the fault type and the corresponding fault maintenance time of each part of the vehicle according to a maintenance recording card of each part of the vehicle, extracting historical data of the whole life cycle of each part in the vehicle driving process from the automobile leaving time to the fault maintenance time from a cloud platform based on the fault type and the fault maintenance time, wherein data fields comprise can signal data of the part, part loads, the external environment of the vehicle in driving, vehicle driving behaviors and the like.
Step 2: data aggregation and cleaning of historical data of full life cycle of parts
The can signal data of each data field of the parts are transmitted through different can buses, the can signal parameters of different data fields have different sampling intervals and sampling periods, the signal acquisition frequency is higher, and the signal data are millisecond-level data.
The method comprises the following steps of data cleaning, wherein the steps of abnormal value elimination and missing value filling of signals of each data field of historical data are included.
And then, data aggregation is carried out, in order to prevent a large amount of data information from being lost and the time sequence from being overlong, data aggregation is carried out by taking minutes as a unit, signals of data fields of historical data are sequenced in time sequence, grouping is carried out by taking minutes as a unit, and the average value of each clock group is taken as the value of the current minute.
And step 3: time-admission sequence segmentation and marking
One-time start-stop (from the start of the automobile to the flameout of the automobile) in the driving process of the automobile is a cyclic process, the invention uses one-time start-stop as a time sliding window to carry out data segmentation and marking, and the operation is as follows:
and (3) sequencing the data aggregated in the step (2) according to a time sequence, and grouping the data according to single start and stop.
Secondly, because different drivers have different vehicle using habits in the driving process of the automobile and each starting and stopping period is different, namely the time series lengths of the single starting and stopping are different, the time series lengths of the single starting and stopping are required to be adjusted to be consistent, according to the vehicle using habits of most users, the time series lengths of the single starting and stopping are fixed to a fixed value L, and when the time lengths of the single starting and stopping exceed L, only data from the starting time to the L time is intercepted and taken as the time series data of the starting and stopping; and when the time length of the single start-stop does not exceed L, keeping the time series data from the starting time to the stopping time unchanged, and filling the time series from the stopping time to the L time with a value of 0.
And finally, marking the corresponding label on the single start-stop data.
And 4, step 4: time series model-based part fault diagnosis modeling
For non-accidental and non-sudden fault types, a slow change process exists in the time dimension, the method utilizes a time series model to perform modeling analysis, and time-allowed segment data generated in the step 3 is sent into the constructed time series model to perform training iteration.
The invention adopts the technical scheme to produce the following beneficial effects:
1. the method utilizes the vehicle type data, the part data, the maintenance record data and the large historical data of the driving of the vehicle to construct the fault diagnosis model of the part of the vehicle, the data has comprehensive and rich sources, and comprises a large amount of data which can not be sensed by the sensor, so that the fault of the part of the vehicle can be more accurately and comprehensively predicted, particularly, the fault which can not be sensed by the sensor can be predicted by the model, the fault can be predicted before the fault of the part of the vehicle occurs, early warning is realized, the fault source is quickly positioned, the timeliness and the accuracy of 4S shop fault maintenance are improved, the misdiagnosis, the mischange and the mischange of the part are avoided, the fault complaint of a user is reduced, and the satisfaction degree of the customer is improved.
2. The initial start time for the failure characterization of the component should be within a period of time before the repair time. However, the initial starting time of fault representation of the parts is difficult to determine in the actual situation, which brings certain difficulty to data marking.
The present invention also provides, in a second aspect, an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor, the memory storing a computer program executable by the at least one processor, the computer program, when executed by the at least one processor, for causing the electronic device to perform the intelligent automotive component failure diagnosis modeling method according to the first aspect.
Further, the present invention also provides in a third aspect a computer program product comprising a computer program, wherein the computer program, when being executed by a processor of a computer, is adapted to make the computer execute the intelligent automobile part failure diagnosis modeling method according to the first aspect.
The above electronic device and computer program product have the same technical effects as the first aspect of the present invention.
Description of the drawings:
FIG. 1 is a fault diagnosis flow diagram of the present invention;
FIG. 2 is a simplified diagram of data sliding window and labeling during data processing according to the present invention;
FIG. 3 is a flow chart of time series length alignment according to the present invention;
FIG. 4 is a flow chart of modeling a time series model according to the present invention.
The specific implementation mode is as follows:
embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present application are shown in the drawings, it should be understood that the present application may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather these embodiments are provided for a more complete and thorough understanding of the present application. It should be understood that the drawings and embodiments of the present application are for illustration purposes only and are not intended to limit the scope of the present application.
It should be understood that the various steps recited in the method embodiments of the present application may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present application is not limited in this respect.
The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description. It should be noted that the terms "first", "second", and the like in the present application are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence of the functions performed by the devices, modules or units.
It is noted that references to "a" or "an" modification in this application are intended to be illustrative rather than limiting, and those skilled in the art will appreciate that references to "one or more" are intended to be exemplary unless the context clearly indicates otherwise.
The names of messages or information exchanged between the units in the embodiments of the present application are for illustrative purposes only, and are not intended to limit the scope of the messages or information.
Referring to fig. 1, in an embodiment, a fault diagnosis modeling process of an automobile part is shown, taking a battery fault diagnosis model as an example, and battery driving history data includes signal data such as battery current, voltage, soc, soh, battery temperature, vehicle speed, external environment temperature, vehicle lamp signal, air conditioner state, and the like. The original data are sampled according to millisecond level, the original data are aggregated according to minute, the average value of each signal in each minute is reserved, and finally the original data are sequenced according to time sequence, the data are input into a fault diagnosis model by taking single start-stop (automobile start-up to automobile flameout) as a time sequence, and then a group of diagnosis results are output.
The following is a detailed description of the process:
step 1: and acquiring big data of the automobile driving history and maintenance record data.
The method comprises the steps of determining the fault type and the corresponding fault maintenance time of each part of the vehicle according to a maintenance recording card of each part of the vehicle, extracting historical data of the whole life cycle of each part in the vehicle driving process from the automobile leaving time to the fault maintenance time from a cloud platform based on the fault type and the fault maintenance time, wherein data fields comprise can signal data of the part, part loads, the external environment of the vehicle in driving, vehicle driving behaviors and the like.
Taking the battery failure diagnosis as an example, the battery running history data includes signal data such as battery current, voltage, soc, soh, battery temperature, vehicle speed, external environment temperature, vehicle light signal, air conditioner state, and the like.
Step 2: and performing data aggregation and cleaning on historical data of the whole life cycle of the automobile parts.
The method comprises the following steps of data cleaning, wherein the steps of abnormal value elimination and missing value filling of signals of each data field of historical data are included.
Then, data aggregation is carried out, in order to prevent a large amount of data information from being lost and a time sequence from being too long, original data is sampled according to millisecond level, aggregation processing is carried out on the original data according to minute unit, signals of data fields of historical data are sequenced according to time sequence and grouped according to minute unit, and an average value of each clock group is taken as a value of a current minute.
And step 3: and performing data segmentation and marking by taking one-time start-stop in the driving process of the automobile as a time sliding window.
Firstly, the data aggregated in the step 2 are sorted according to the time sequence, and then data segmentation is carried out by taking one start-stop in the driving process of the automobile as a time sliding window.
In one embodiment, referring to fig. 2, a schematic diagram of data sliding window and labeling during data processing is shown, which has a slow change process in the time dimension for non-accidental, non-sudden automobile part failure, however, when the automobile fails and goes to a store of 4s for maintenance, the maintenance time recorded by the maintenance record card is not the initial starting time of the part failure characterization, and the initial starting time of the part failure characterization should be a period of time before the maintenance time. However, in real situations, it is difficult to determine the initial starting time of the fault characterization of the component, which brings some difficulty to data marking. In order to solve the accuracy of data labels when marking data, 3 time windows are set in the whole life cycle historical data (from car purchasing time to maintenance time) of the parts, wherein the time windows are T1, T2 and T3 respectively. Wherein the data Label in the time window T1 is set as Label, and the data Label in the time window T3 is set as corresponding Label' according to the specific fault type; the data in the time window T2 is discarded due to the fact that the data of the Label and the Label' are mixed, so that the accuracy of the data Label can be guaranteed, the fault data can be guaranteed to have sufficient sample size, and the problems that positive and negative samples are unbalanced and the like are avoided.
And respectively taking single start and stop of the vehicle (from automobile start to automobile flameout) as sliding time windows in the time period of the time window T1 and the time period of the time window T3, and processing and marking the data.
Secondly, adjust the time sequence length of each single time start-stop time sequence data to unified fixed length L: fixing the time series length of single start-stop to a fixed value L, judging whether the time series data length N of each single start-stop is greater than the set fixed value L, and when N is greater than L, only intercepting the data from the start time to the L time as the time series data of the single start-stop; when N < L, the time series data from the start time to the stop time is kept unchanged, the time series from the stop time to the L time is filled with a value of 0, and the time series length alignment process is shown in fig. 3.
And finally, marking the single start-stop data with a corresponding label.
And 4, step 4: and constructing a fault diagnosis model based on an LSTM time sequence model and a neural network model, and finally inputting the processed data into the constructed algorithm model according to a single start-stop time sequence for training and verification, which is shown in FIG. 4.
The automobile part fault diagnosis modeling method can be used for fault diagnosis and prediction that a sensor cannot sense in the automobile running process, fault diagnosis and early warning of each part are achieved, and accidents are prevented.
Further, an exemplary embodiment of the present application also provides an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor. The memory stores a computer program executable by the at least one processor, the computer program, when executed by the at least one processor, is configured to cause the electronic device to perform the method for modeling fault diagnosis of automobile parts as shown in the above embodiments of the present application.
The exemplary embodiments of the present application also provide a computer program product, which includes a computer program, wherein the computer program is used for causing a computer to execute the automobile part fault diagnosis modeling method demonstrated according to the embodiments of the present application when the computer program is executed by a processor of the computer.
Claims (8)
1. An intelligent modeling method for fault diagnosis of automobile parts is characterized by comprising the following steps:
step 1: acquiring historical big data and maintenance record data of automobile running;
step 2: performing data aggregation and cleaning on historical data of the whole life cycle of the automobile parts;
and 3, step 3: performing data segmentation and marking by taking one-time start-stop in the driving process of the automobile as a time sliding window;
and 4, step 4: and constructing a fault diagnosis model based on the LSTM time sequence model and the neural network model.
2. The intelligent automobile part fault diagnosis modeling method according to claim 1, wherein the step 1 specifically comprises:
the method comprises the steps of determining fault types and corresponding fault maintenance time of all parts of a vehicle according to maintenance recording cards of all parts of the vehicle, extracting historical data of the whole life cycles of all parts in the vehicle driving process from the automobile leaving time to the fault maintenance time from a cloud platform based on the fault types and the fault maintenance time, wherein data fields comprise can signal data of the parts, part loads, the external vehicle driving environment, vehicle driving behaviors and the like.
3. The intelligent automobile part fault diagnosis modeling method according to claim 1, wherein the step 2 comprises:
data cleaning, including removing abnormal values and filling missing values of various data field signals of historical data of the whole life cycle of the part;
and data aggregation, namely performing data aggregation by taking minutes as a unit, sequencing data field signals of the historical data in a time sequence, grouping the data field signals by taking the minutes as a unit, and taking the average value of each clock group as the value of the current minute.
4. The intelligent automobile part fault diagnosis modeling method according to claim 1, wherein the step 3 comprises
Firstly, sequencing according to a time sequence based on aggregated data, and then grouping according to single start-stop;
secondly, fixing the time sequence length of single start-stop to a fixed value L, and when the time length of single start-stop exceeds L, only intercepting data from the starting time to the L time as the time sequence data of the start-stop; when the time length of single start-stop does not exceed L, keeping time series data from the starting time to the stopping time unchanged, and filling the time series from the stopping time to the L time with a 0 value;
and finally, marking the corresponding label on the single start-stop data.
5. The intelligent automobile part fault diagnosis modeling method according to claim 4, characterized in that 3 time windows, namely T1, T2 and T3, are set in the historical data of the whole life cycle of the part, wherein the data Label in the time window T1 is set as Label, and the data Label in the time window T3 is set as corresponding Label' according to the specific fault type; discarding the data in the time window T2; and respectively taking single start and stop of the vehicle as sliding time windows within the time period of the time window T1 and the time period of the time window T3, and processing and marking the data.
6. The intelligent automobile part fault diagnosis modeling method according to any one of claims 1-5, wherein in step 4, a fault diagnosis model is constructed based on an LSTM time sequence model and a neural network model, modeling analysis is performed, and the time-allowed segment data generated in step 3 is sent to the constructed time sequence model for training iteration.
7. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor, the memory storing a computer program executable by the at least one processor, the computer program, when executed by the at least one processor, for causing the electronic device to perform the intelligent automotive component failure diagnosis modeling method of any one of claims 1 to 6.
8. A computer program product comprising a computer program, wherein the computer program, when executed by a processor of a computer, is adapted to cause the computer to perform the intelligent modeling method for fault diagnosis of automotive parts according to any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211296026.6A CN115660640A (en) | 2022-10-21 | 2022-10-21 | Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211296026.6A CN115660640A (en) | 2022-10-21 | 2022-10-21 | Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115660640A true CN115660640A (en) | 2023-01-31 |
Family
ID=84988746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211296026.6A Pending CN115660640A (en) | 2022-10-21 | 2022-10-21 | Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115660640A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116451122A (en) * | 2023-04-23 | 2023-07-18 | 北京思维实创科技有限公司 | Fault determination method and device, electronic equipment and storage medium |
-
2022
- 2022-10-21 CN CN202211296026.6A patent/CN115660640A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116451122A (en) * | 2023-04-23 | 2023-07-18 | 北京思维实创科技有限公司 | Fault determination method and device, electronic equipment and storage medium |
| CN116451122B (en) * | 2023-04-23 | 2024-03-05 | 北京磁浮有限公司 | Fault determination method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111241154B (en) | Storage battery fault early warning method and system based on big data | |
| CN110647133B (en) | Rail transit equipment state detection maintenance method and system | |
| CN109947080B (en) | Fault diagnosis method and device, computer storage medium and terminal | |
| DE102015214739A1 (en) | Determining a cause of a fault in a vehicle | |
| CN107845159B (en) | Operation monitoring system of automatic driving vehicle evaluation system | |
| CN111997709B (en) | On-line monitoring method and system for vehicle-mounted engine oil | |
| CN116308304B (en) | New energy intelligent operation and maintenance method and system based on meta learning concept drift detection | |
| KR101989962B1 (en) | Integrated management server and building management system using the same | |
| CN115660640A (en) | Intelligent automobile part fault diagnosis modeling method, electronic equipment and program product | |
| CN110618384A (en) | Motor performance test platform | |
| Singh et al. | Data-driven framework for detecting anomalies in field failure data | |
| CN114966422A (en) | Real-time monitoring and early warning system based on power battery parameters | |
| Tichý et al. | Predictive diagnostics usage for telematic systems maintenance | |
| CN116880454A (en) | Intelligent diagnosis system and method for vehicle faults | |
| CN110954335A (en) | Health assessment method and system for industrial vehicle | |
| CN116311593A (en) | Energy consumption calculation method and system | |
| CN114281052A (en) | Mechanical equipment running state monitoring device, system and method | |
| JP2021066232A (en) | Platform door diagnostic device, platform door system, platform door diagnostic method and platform door diagnostic program | |
| CN112036694B (en) | Highway electromechanical equipment life cycle prediction method and system | |
| CN112448967B (en) | Control method and device for uploading data, control equipment and automobile | |
| CN112803587A (en) | Intelligent inspection method for state of automatic equipment based on diagnosis decision library | |
| CN116008701B (en) | Electric mechanism operation diagnosis system and method for intelligent high-voltage switch cabinet | |
| CN115600695B (en) | Fault diagnosis method for metering equipment | |
| CN116775464A (en) | Whole vehicle testing method and system | |
| CN115933619A (en) | Remote diagnosis method, system, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |