An intelligent fault self-diagnosis method for motorcycle electronic control system

Abstract

The present invention relates to the technical field of data processing, and discloses an intelligent fault self-diagnosis method for a motorcycle electronic control system, which performs fault analysis and self-diagnosis processing of a motorcycle based on a big data model; performs risk assessment and monitoring level division of a motorcycle based on a big data model; obtains historical fault data of a motorcycle electronic control system through the cloud, and establishes a big data model of motorcycle faults; obtains real-time data of the current motorcycle through the cloud, including: current fault type, real-time data of sensors, working status of actuators, power supply voltage of the electronic control system, and environmental information; the big data model performs data cleaning, data format conversion, and data missing value processing on the acquired fault data, so that the processed fault data can be used for computer calculation and storage. It can quickly discover potential and occurring problems of a motorcycle, and improve the environmental adaptability of the motorcycle vehicle control system.

Description

Technical Field

[0001] The invention relates to the field of data processing, and in particular to an intelligent fault self-diagnosis method for a motorcycle electronic control system. Background Art

[0002] The motorcycle electronic control system includes a closed-loop electronic carburetor system and an electronic stability system. The electronic stability system ranges from the most basic ABS to power mode selection, as well as traction control, cornering ABS, etc. The closed-loop electronic carburetor system is mainly composed of signal acquisition components such as engine speed, load, oxygen sensor, temperature sensor, data calculation electronic control unit ECU, and solenoid control valve and other execution components. With the development of motorcycle technology, the electronic control system has become more complex and perfect, and higher requirements have been put forward for the maintenance of the electronic control system.

[0003] At present, the fault diagnosis method of the motorcycle electronic control system is mainly to read the fault code through the on-board diagnostic system for diagnosis, and then find out the fault nature and fault location of the electronic control system.

[0004] However, the fault diagnosis scheme of the prior art still has the following shortcomings:

[0005] (1) Testing is only performed when the fault problem occurs, and it is impossible to provide early warning of motorcycle faults;

[0006] (2) External adaptability is poor. When external factors change, the accuracy of fault diagnosis may be reduced. Summary of the invention

[0007] In view of the above technical problems, the present invention proposes an intelligent fault self-diagnosis method for a motorcycle electronic control system.

[0008] The technical solution used in the present invention is: an intelligent fault self-diagnosis method for a motorcycle electronic control system, comprising the following steps: establishing a big data model of motorcycle faults through the cloud; performing fault analysis and self-diagnosis processing on the motorcycle based on the big data model; performing risk assessment and monitoring level division on the motorcycle based on the big data model; obtaining historical fault data of the motorcycle electronic control system through the cloud to establish a big data model of motorcycle faults; obtaining real-time data of the current motorcycle through the cloud, including: current fault type, real-time data of sensors, working status of actuators, power supply voltage of the electronic control system, and environmental information; the big data model performs data cleaning, data format conversion, and data missing value processing on the acquired real-time data, so that the processed real-time data can be used for computer calculation and storage.

[0009] Furthermore, the fault analysis and self-diagnosis processing of the motorcycle based on the big data model also includes: obtaining relevant parameters of the fault type according to the fault type of the real-time data; performing statistical analysis on the relevant parameters to calculate the correlation coefficient between the fault type and the relevant parameters; analyzing the abnormality degree of the relevant parameters to calculate the correlation index between the abnormality degree of the relevant parameters and the actual fault degree; determining the threshold value of different fault types according to the relationship between the abnormality degree of the relevant parameters and the actual abnormality degree, and calculating the self-diagnosis result.

[0010] Furthermore, the method of obtaining statistically relevant parameters of the fault type based on the fault type of the real-time data also includes classifying the processed fault data according to corresponding data types to obtain the relevant parameters of the fault type; converting the fault data into point clusters in a high-dimensional coordinate system based on the relevant parameters, and performing statistical calculations on the relevant parameters corresponding to each point cluster.

[0011] Furthermore, the relevant parameters are statistically analyzed to calculate the correlation coefficient between the fault type and the relevant parameters; the calculation formula of the correlation coefficient is: ;in, is the correlation coefficient between this type of fault and various related parameters, e is a natural constant, Indicates that in any fault cluster class among all fault cluster classes, For the The number of data points with abnormal parameters, for Point The distance between a point and the center of the cluster.

[0012] Furthermore, the abnormality of the relevant parameters is analyzed to calculate a correlation index between the abnormality of the relevant parameters and the actual fault degree; the calculation formula of the correlation index is: ;in, For the The degree of abnormality of the parameters is The correlation index between the actual degree of the fault clusters is Among the relevant data of each fault, The number of abnormal data of each parameter is denoted as k, and c is The jth data in the abnormal data The value of the parameter, is the normal range of the parameter c, The fault severity of this data.

[0013] Furthermore, according to the relationship between the abnormality degree of the relevant parameters and the actual abnormality degree, the fault weight is calculated, the thresholds of different fault types are determined, and the self-diagnosis result is calculated; the fault weight of the parameter is also included; the calculation formula of the fault weight is: ;in, For the Failure The fault weight of the parameter, To consider the first The degree of influence of the parameters on it, for For the Parameter pair The unique representation of a fault, norm is the normalization function of the big data model.

[0014] Furthermore, it also includes the unique representation of the degree of parameter abnormality to the actual degree of the fault; the calculation formula of the unique representation is: ; where e is a natural constant, For the Parameter pair The unique representation of a fault, For the The degree of abnormality of the parameters is The correlation index between the actual degree of failure, For the The degree of abnormality of the parameters is not The first fault The correlation index between the actual degree of failure, is the total number of fault types.

[0015] Further, the first The real-time fault assessment result of the fault is normalized and compared with the Fault preset threshold and threshold Compare and perform self-diagnosis processing to form self-diagnosis data; when the fault evaluation result is greater than the threshold When the fault continues to increase for more than When the fault assessment result is greater than the threshold When the vehicle owner is warned, the self-diagnosis data is recorded in a database.

[0016] Furthermore, the risk assessment and monitoring level division of motorcycles are performed based on the big data model; the self-diagnosis data is stored in the big data model, the big data model monitors the frequency of each abnormal parameter of the motorcycle in real time, and divides the risk level of the motorcycle into a low risk level and a high risk level; the low risk level, in a certain period of time, when the proportion of abnormal data volume of the parameter is greater than a preset threshold When the user reports that there is no abnormality in the motorcycle, and the fluctuation pattern of the value of the abnormal parameter is the same as that under normal circumstances, it is determined that the environment has changed to affect the parameters of the motorcycle, and the abnormal parameter value is set to a normal value; for the high risk level, in a certain period of time, according to the frequency of each abnormal parameter, the parameters that are easy to be judged to be abnormal are set as key monitoring objects, and a shorter period of monitoring is performed. When an abnormality is detected, self-diagnosis processing is immediately performed on the fault related to the parameter.

[0017] Furthermore, the sensors include but are not limited to: throttle position sensor, oxygen sensor, position sensor and speed sensor; the actuators include but are not limited to: fuel injector, ignition coil and idle control valve; the power supply voltage mainly includes: the output of the battery and the generator; the environmental information refers to the environmental conditions when the fault occurs.

[0018] Compared with the prior art, the present invention has the following beneficial effects: by using the cloud big data model and the motorcycle numerical control system fault data, the correlation coefficient between each parameter abnormality and the related fault is obtained according to the classification analysis; the abnormal degree of the detected data is analyzed to confirm the correlation between the abnormal degree of the data and the abnormal degree of the vehicle body performance; the corresponding threshold is set for each fault to record the abnormal data and predict the occurrence of the fault; the frequency of the abnormal parameters is counted and the monitoring intensity of the abnormal parameters is adaptively adjusted. Based on this, the potential and occurring problems of the motorcycle can be quickly discovered, and the environmental adaptability of the motorcycle control system can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A diagram showing the connection relationship between the smart motorcycle of the present invention and the cloud big data model.

[0020] Figure 2 The present invention is a flow chart of an intelligent fault self-diagnosis method of a motorcycle electronic control system. DETAILED DESCRIPTION

[0021] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0022] The present invention will be specifically described below in conjunction with the accompanying drawings:

[0023] like Figure 1 As shown in the figure, the connection relationship between motorcycles and big data models is shown. Each motorcycle has established a real-time connection relationship with the big data model in the cloud.

[0024] The cloud-based big data model obtains historical data of each motorcycle and performs self-diagnosis analysis on the current motorcycle faults, including:

[0025] 1) Using the motorcycle CNC system fault data, the correlation coefficient between each parameter abnormality and related faults is obtained through classification analysis;

[0026] 2) Analyze the degree of abnormality of the detected data and confirm the correlation between the degree of abnormality of the data and the degree of abnormality of the vehicle performance;

[0027] 3) Based on this, set corresponding thresholds for each fault to record abnormal data and predict the occurrence of faults;

[0028] 4) Count the frequency of abnormal parameters, adaptively adjust the monitoring intensity of abnormal parameters, and improve the environmental adaptability of the system.

[0029] An intelligent fault self-diagnosis method for a motorcycle electronic control system, such as Figure 2 As shown, the fault self-diagnosis method comprises the following steps:

[0030] Step S100: Establishing a big data model of motorcycle failures through the cloud.

[0031] The historical fault data of the motorcycle electronic control system is obtained through the cloud, and a big data model of motorcycle faults is established.

[0032] The real-time data of the current motorcycle is obtained through the cloud, including: the current fault type, the real-time data of the sensor, the working status of the actuator, the power supply voltage of the electronic control system, and the environmental information.

[0033] The sensors include but are not limited to: a throttle position sensor, an oxygen sensor, a position sensor and a speed sensor.

[0034] The actuators include but are not limited to: a fuel injector, an ignition coil and an idle speed control valve.

[0035] The power supply voltage mainly includes: the output of the battery and the generator.

[0036] The environmental information specifically refers to the environmental conditions when the fault occurs.

[0037] The big data model performs data cleaning, data format conversion and data missing value processing on the acquired real-time data, so that the processed real-time data can be used for computer calculation and storage.

[0038] Step S200: Performing motorcycle fault analysis and self-diagnosis processing based on the big data model.

[0039] The big data model analyzes the fault data, calculates the correlation coefficient between the fault type and the parameters, and obtains the self-diagnosis results through the degree of abnormality of the system data and makes corresponding records.

[0040] The specific steps include:

[0041] Step S210: According to the fault type of the real-time data, obtain relevant parameters of the fault type through statistics.

[0042] It should be noted that after data cleaning, data format conversion and data missing value processing, the following two steps are also included:

[0043] 1) Classifying the processed fault data according to corresponding data types to obtain the relevant parameters of each type of fault;

[0044] 2) According to the relevant parameters, the fault data is converted into point clusters in a high-dimensional coordinate system, each point cluster corresponds to a fault, and statistical calculations are performed on the relevant parameters involved in the fault data corresponding to each point cluster.

[0045] Step S220: Statistically analyze the relevant parameters to calculate the correlation coefficient between the fault type and the relevant parameters; analyze the abnormality degree of the relevant parameters to calculate the correlation index between the abnormality degree of the relevant parameters and the actual fault degree.

[0046] It should be noted that, when a parameter appears more frequently in a point cluster representing a certain type of fault, it means that the point related to the parameter is closer to the cluster center, and the correlation between the related parameter and this type of fault is stronger.

[0047] The calculation formula of the correlation coefficient between any type of fault and various related parameters is as follows:

[0048]

[0049] In the above formula, For the The first fault and The correlation coefficient of the parameters, e is a natural constant, Indicates that in any fault cluster class among all fault cluster classes, For the The number of data points with abnormal parameters, for Point The distance between a point and the center of the cluster.

[0050] In the above formula, when Among all the points in the fault cluster, The more abnormal data points there are, the The fault clusters and The greater the correlation between the parameters, the The larger the value, the The fault clusters and The larger the correlation coefficient of the parameters, the greater the The distance between the data point with the parameter abnormality and The closer the distance between the center points of the fault clusters, the The fault clusters and The greater the correlation between the parameters, the The smaller, The bigger, the The first fault and The larger the correlation coefficient of the parameters.

[0051] It should be noted that in actual situations, there are cases where the fault data fluctuates slightly but indicates a serious fault, and there are also cases where the fault data fluctuates greatly but indicates a minor fault. For example, the oxygen sensor data provides information about the oxygen content in the engine exhaust, and even a small deviation may warn of combustion efficiency problems or emission control system failures; the battery voltage will fluctuate greatly when the starter motor is working or when a large number of electrical devices are used, but it will not warn of a battery fault.

[0052] Therefore, it is necessary to calculate the relationship between the abnormal degree of all relevant parameters of the fault cluster and the actual fault degree. By quantifying the actual fault degree in all fault data, the more serious the actual situation is, the larger the quantified value will be.

[0053] Step S230: Analyze the abnormality degree of the relevant parameters, and calculate the correlation index between the abnormality degree of the relevant parameters and the actual fault degree.

[0054] Correlation index between the degree of parameter abnormality and the actual degree of fault The calculation formula is as follows:

[0055]

[0056] In the above formula, For the The degree of abnormality of the parameters is The correlation index between the actual degree of the fault clusters is Among the relevant data of each fault, The number of abnormal data of each parameter is denoted as k, and c is The jth data in the abnormal data The value of the parameter, is the normal range of the parameter c, The fault severity of this data.

[0057] In the above formula, as well as Indicates the first The degree of abnormality of each parameter compared with the normal range. Since the units and data sizes of each parameter are different, the abnormal value alone cannot measure the degree of abnormality of the data. The calculation ratio is more accurate.

[0058] In the above formula, when the fault degree is the same, that is, is a fixed value. At this time, the greater the degree of data anomaly, the smaller the correlation between the degree of data anomaly and the actual anomaly. or The larger the value, the The degree of abnormality of the parameters is The correlation index between the actual degree of failure The smaller the data is, the smaller the or is a fixed value. At this time, the greater the actual fault degree, the greater the correlation between the data abnormality degree and the actual abnormality. The larger the value, the The degree of abnormality of the parameters is The correlation index between the actual degree of failure The bigger.

[0059] Step S240: Calculate the fault weight according to the relationship between the abnormality degree of the relevant parameters and the actual abnormality degree, determine the thresholds of different fault types, and calculate the self-diagnosis result.

[0060] It should be noted that when evaluating a fault, it is necessary to consider the correlation between each parameter and the fault, the parameters that can significantly change the actual fault degree with only a slight abnormality, and the unique representativeness of a certain factor for the fault.

[0061] When calculating the failure index Failure Fault weights of parameters The calculation formula is as follows:

[0062]

[0063] In the above formula, For the Failure The fault weight of the parameter, To consider the first The degree of influence of the parameters on it, For the Parameter pair The unique representation of each fault, norm is the normalization function of the big data model; The degree of abnormality of the parameters is The correlation index between the actual degree of failure For The first fault and The correlation coefficient of the parameters Correction is made to make the correlation coefficient more accurate; Reflection time Parameter pair The difference between the ability to represent a fault and the ability to represent other faults is multiplied to obtain the value of the first fault. Parameter pair When a fault has unique representation, it is appropriate to increase the ability of this parameter to influence the evaluation results of this type of fault.

[0064] The fault weight is used to calculate the The real-time fault assessment result of the fault is normalized and compared with the Fault preset threshold and threshold The comparison is performed to perform self-diagnosis processing and form self-diagnosis data.

[0065] When the fault assessment result is greater than the threshold When the fault assessment result keeps increasing for more than 1 second, the information is recorded in the database and the changing trend of the fault assessment result is monitored. When the car owner is alerted,

[0066] When the fault assessment result is greater than the threshold When a vehicle owner is struck by an unexpected error, the vehicle owner is immediately warned and the information is recorded in the database.

[0067] It should be noted that by setting the threshold σ, data that may be a fault can be recorded into the big data model for real-time detection and analysis; the specific value of the threshold σ is set according to the actual scenario needs. The lower the threshold σ is set, the easier it is to record the data into the database, for example: the threshold σ is 0.5.

[0068] It should be noted that by setting the threshold ε, data information that may be a fault can be identified and an alarm can be issued for real-time detection and alarm; the specific value of the threshold ε is set according to the actual alarm judgment requirements. The lower the threshold ε is set, the stricter the fault detection is, for example: the threshold ε is 0.8.

[0069] It can be understood that the threshold ε must be greater than the threshold σ.

[0070] Step S300: Perform risk assessment and monitoring level classification of motorcycles based on the big data model.

[0071] The above self-diagnosis data is stored in a big data model, which monitors the frequency of various abnormal parameters of the motorcycle in real time and divides the risk level of the motorcycle into a low risk level and a high risk level.

[0072] Low risk level: within a certain period of time, when the proportion of abnormal data of the parameter is greater than the set threshold When the user reports that there is no abnormality in the motorcycle, and the fluctuation pattern of the abnormal parameter value is the same as that in the normal situation, it is determined that the environment has changed and thus affected the parameters of the motorcycle, and the abnormal parameter value is set to the normal value.

[0073] High risk level: in a certain period of time, according to the frequency of each abnormal parameter, the parameters that are easy to be judged to be abnormal are set as key monitoring objects, and a shorter period of monitoring is carried out. When an abnormality is detected, the fault related to the parameter is immediately self-diagnosed.

[0074] It is understandable that the computer device structure provided in this embodiment does not constitute a limitation on the computer device, and may include more or fewer components, or a combination of certain components, or different component arrangements.

[0075] It is understandable that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present application.

[0076] The above disclosure only discloses several specific implementation scenarios of the present application. However, the present application is not limited thereto, and any changes that can be conceived by technicians in this field should fall within the scope of protection of the present application.

Claims

1. An intelligent fault self-diagnosis method for a motorcycle electronic control system, characterized in that: The following steps are involved: Build a big data model of motorcycle failures through the cloud; Motorcycle fault analysis and self-diagnosis based on big data models; Risk assessment and monitoring level classification of motorcycles based on big data models; Acquire historical fault data of the motorcycle electronic control system through the cloud and establish a big data model of motorcycle faults; Obtaining the current real-time data of the motorcycle through the cloud, including: the current fault type, the real-time data of the sensor, the working status of the actuator, the power supply voltage of the electronic control system, and environmental information; The big data model performs data cleaning, data format conversion and data missing value processing on the acquired real-time data, so that the processed real-time data can be used for computer calculation and storage; The motorcycle fault analysis and self-diagnosis processing based on the big data model also includes: According to the fault type of the real-time data, statistically obtain relevant parameters of the fault type; Performing statistical analysis on the relevant parameters to calculate the correlation coefficient between the fault type and the relevant parameters; Analyzing the abnormality degree of the relevant parameters and calculating the correlation index between the abnormality degree of the relevant parameters and the actual fault degree; According to the relationship between the abnormality degree of the relevant parameters and the actual abnormality degree, the thresholds of different fault types are determined, and the self-diagnosis results are calculated; Performing statistical analysis on the relevant parameters to calculate the correlation coefficient between the fault type and the relevant parameters; The calculation formula of the correlation coefficient is: ; in, For the The first fault and The correlation coefficient of the parameters, e is a natural constant, Indicates any fault cluster class among all fault cluster classes, For the The number of data points with abnormal parameters, for Point The distance between a point and the center of the cluster; The abnormality degree of the relevant parameters is analyzed to calculate the correlation index between the abnormality degree of the relevant parameters and the actual fault degree; The calculation formula of the correlation index is: ; in, For the The degree of abnormality of the parameters is The correlation index between the actual degree of the fault is Among the relevant data of each fault, The number of abnormal data of each parameter is denoted as k, and c is The jth data in the abnormal data The value of the parameter, is the normal range of the parameter c, The fault severity of this data.

2. The intelligent fault self-diagnosis method of a motorcycle electronic control system according to claim 1, characterized in that: The obtaining, according to the fault type of the real-time data, relevant parameters of the fault type through statistics, further comprises: Classifying the processed fault data according to corresponding data types to obtain the relevant parameters of the fault type; According to the relevant parameters, the fault data is converted into point clusters in a high-dimensional coordinate system, and statistical calculations are performed on the relevant parameters corresponding to each point cluster.

3. The intelligent fault self-diagnosis method of a motorcycle electronic control system according to claim 1, characterized in that: The method further includes calculating the fault weight according to the relationship between the abnormality degree of the relevant parameters and the actual abnormality degree, determining the thresholds of different fault types, and calculating the self-diagnosis result; and also includes the fault weight of the parameter; The calculation formula of the fault weight is: ; in, For the Failure The fault weight of the parameter, To consider the first The degree of influence of the parameters on it, For the Parameter pair The unique representation of a fault, norm is the normalization function of the big data model; It also includes the unique representation of the actual extent of the fault by the degree of parameter abnormality; The calculation formula of the unique representation is: ; in, For the Parameter pair The unique representation of a fault, e is a natural constant, For the The degree of abnormality of the parameters is The correlation index between the actual degree of failure, For the The degree of abnormality of the parameters is not The first fault The correlation index between the actual degree of failure, is the total number of fault types.

4. The intelligent fault self-diagnosis method of a motorcycle electronic control system according to claim 3 is characterized in that: The fault weight is used to calculate the The real-time fault assessment result of the fault is normalized and compared with the Fault preset threshold and threshold Comparison is performed to perform self-diagnosis processing and form self-diagnosis data; When the fault assessment result is greater than the threshold When the fault continues to increase for more than When the car owner is alerted, When the fault assessment result is greater than the threshold When the vehicle owner is warned, the self-diagnosis data is recorded in the database.

5. The intelligent fault self-diagnosis method of a motorcycle electronic control system according to claim 4 is characterized in that: The risk assessment and monitoring level classification of motorcycles based on the big data model; The self-diagnosis data is stored in a big data model, the big data model monitors the frequency of each abnormal parameter of the motorcycle in real time, and divides the risk level of the motorcycle into a low risk level and a high risk level; The low risk level is when the proportion of abnormal data of the parameter is greater than the preset threshold within a certain period of time. When the user reports that the motorcycle has no abnormality, and the fluctuation pattern of the abnormal parameter value is the same as that under normal circumstances, it is determined that the environment has changed and thus affects the parameters of the motorcycle, and the abnormal parameter value is set to the normal value; The high risk level, in a certain period of time, according to the frequency of each abnormal parameter, will easily determine the abnormal parameters as the key monitoring objects, and perform short-term monitoring. When an abnormality is detected, self-diagnosis processing is immediately performed on the fault related to the parameter.

6. The intelligent fault self-diagnosis method of a motorcycle electronic control system according to claim 1, characterized in that: The sensors include but are not limited to: throttle position sensor, oxygen sensor, position sensor and speed sensor; The actuators include but are not limited to: fuel injectors, ignition coils and idle speed control valves; The power supply voltage mainly includes: the output of the battery and the generator; The environmental information refers to the environmental conditions when the fault occurs.

Images