Vehicle user big data portrait analysis method and system based on damage model

Abstract

The application discloses a vehicle user big data portrait analysis method and system based on a damage model, which comprises the following steps: pre-building a running characteristic classification matrix of a vehicle, combining the actual running state of the vehicle, performing feature extraction based on the classification matrix, and taking the feature extraction as a user label; performing damage classification definition, damage key factor definition and damage correlation parameter definition according to the failure scene of an engine, and extracting damage correlation parameters according to the damage key factor matched by the user label; performing quantitative calculation and cumulative distribution normalization processing according to the collected key damage correlation parameters, obtaining damage data, and clustering user label portraits and representative users representing different damage classifications of the engine according to the damage data sorting. The application applies the mathematical damage model converted from the failure mechanism of the engine to the vehicle terminal big data, performs statistics and analysis, and uses the data as the input of the bench test and design simulation, so that the development verification and market application are unified.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engines, in particular to a vehicle user big data portrait analysis method and system based on a damage model. BACKGROUND

[0002] User portrait refers to a method of collecting a series of user usage information, such as driving area, usage time, usage frequency, etc. User portrait is an effective tool for associating target users with product development goals. By outlining the user portrait of historical products, accurate knowledge of a certain product application can be obtained, and user characteristics can be quantitatively described to provide targeted services for product design and verification.

[0003] The current mainstream vehicle application user portrait has some shortcomings. For example, the patent with the publication number "CN115221234A" discloses a method and system for user portrait based on powertrain data, which mainly uses power performance and economy as the main analysis carrier for clustering, without falling to the engine level. The reliability concerned by the current customers lacks a quantitative evaluation method, it is difficult to accurately describe from the user end, the utilization rate of engine terminal operation data is insufficient, the recognized user portrait is not accurate, and the engine application scene and working characteristics cannot be subdivided from the reliability angle through the user portrait. How to improve the big data clustering evaluation of user portrait is also an urgent problem in the industry.

[0004] The information disclosed in the above background section is only to increase the understanding of the overall background of the present application, and should not be regarded as acknowledging or implying in any form that the information constitutes prior art known to those skilled in the art. SUMMARY

[0005] In view of the above existing defects, the present application aims to provide a vehicle user big data portrait analysis method and system based on a damage model, which collects running data of the engine under different sub-models and use areas, combines the failure damage mechanism of parts, performs user portrait clustering analysis, determines the typical user scenarios and extreme harsh user scenarios based on the reliability of parts, and performs engine part development verification and reliability improvement work.

[0006] The technical scheme of the present application is as follows:

[0007] A vehicle user big data portrait analysis method based on a damage model, comprising:

[0008] S1: Pre-build the running feature classification matrix of the vehicle, combine the actual running condition of the vehicle, and extract features based on the classification matrix as user tags;

[0009] S2: According to the failure scenario of the engine, the damage classification definition, the damage key factor definition and the damage correlation parameter definition are performed, and the damage correlation parameters are extracted according to the damage key factor matched by the user label;

[0010] S3: Quantitative calculation and cumulative distribution normalization processing are performed according to the collected key damage correlation parameters to obtain damage data, and the user label portrait representing different damage classifications of the engine and the representative user are clustered according to the damage data sorting.

[0011] Specifically, the operation characteristics of the vehicle in S1 include: user operation, external use environment, road characteristics, vehicle driving task profile, load characteristics.

[0012] Specifically, in S2, the damage classification definition and the damage correlation parameter definition include: defining the damage classification and the damage correlation parameter according to the engine component reliability failure mechanism and the engine function failure mechanism, and associating the damage classification with the user label.

[0013] Specifically, in S3, the quantitative calculation and the cumulative distribution normalization processing include the following steps:

[0014] S301: Apply the component damage model formula to calculate the pseudo-damage data value of the component failure according to the collected key damage correlation parameters;

[0015] S302: Calculate the cumulative distribution value by calculating the pseudo-damage data value and accumulating the data of all damage classifications.

[0016] S303: Define the relative weight of different damage cumulative distribution values and calculate the damage weighting factor of different users as the damage data.

[0017] The application also discloses a vehicle user big data portrait analysis system based on a damage model, adopts the method, and comprises an engine ECU, a T-BOX vehicle terminal, a cloud platform and a client computer.

[0018] The application has the following beneficial effects:

[0019] The application associates the failure of the engine with the user portrait, and the innovation is that the mathematical damage model converted from the physical failure is applied to the big data of the vehicle terminal based on the failure mechanism of the engine, the data is statistically analyzed, and the user label portrait and representative user representing different types of engine failure are sorted and clustered. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The application is a flowchart of a vehicle user big data portrait analysis method based on a damage model.

[0021] Figure 2 The application is a schematic diagram of a vehicle user big data portrait analysis system based on a damage model. DETAILED DESCRIPTION

[0022] In order to describe the technical content, purposes and effects of the application in detail, the following describes the embodiments with reference to the drawings. In the description of the embodiments, it should be understood that the terms indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only used for the convenience of describing the embodiments and simplifying the description, and cannot be understood as indicating or implying that the devices or elements must have a specific orientation, structure and operation, and therefore cannot be understood as limiting the application. EMBODIMENT

[0023] As Figure 1As shown, according to the embodiment of the present application, a vehicle user big data portrait analysis method based on damage model is realized through the following steps:

[0024] S1: Build a running feature classification matrix of the vehicle, combine the actual running condition of the vehicle, and perform feature extraction based on the classification matrix as the user label. Specifically, the following steps are included:

[0025] User matrix classification: based on user operation, external use environment, road feature, whole vehicle driving task profile, and load feature, etc. to classify users; as shown in Table 1.

[0026]

[0027] User label definition: extract the user matrix classification combined with the actual running feature as the basis of the user portrait; as shown in Table 2, abc1d4e3 represents the user portrait of mountain road overload and overspeed driving, which can be classified as high load reliability damage in the future.

[0028]

[0029] S2: According to the failure scene of the engine, perform damage classification definition, damage key factor definition, and damage correlation parameter definition, and extract the damage correlation parameter according to the damage key factor matched by the user label; specifically including the following steps:

[0030] Reliability failure division: according to the reliability failure mechanism of the main components of the engine, the load change with normal use causes the decline of the function or performance of the components, and the damage classification definition and damage correlation parameter definition are performed; as shown in Table 3.

[0031]

[0032] Functional failure division: according to the failure mechanism that the main functions of the engine do not meet, the damage classification definition and damage correlation parameter definition are performed; as shown in Table 4.

[0033]

[0034] Damage data extraction: according to the extracted damage correlation parameters, according to the user defined by the user label definition, perform T-BOX data collection and key parameter extraction analysis, T-BOX is a vehicle-mounted remote data collection terminal, which is popular in vehicles, and can directly read GPS and engine ECU data stream on the cloud platform. Data analysis is to extract the proportion distribution, amplitude distribution, etc. of damage correlation parameters according to damage key factors.

[0035] S3: Quantitative calculation and cumulative distribution normalization processing are performed according to the collected key damage correlation parameters to obtain damage data. According to the damage data, user label portraits and representative users representing different damage classifications of the engine are sorted and clustered. Specifically, the quantitative calculation and cumulative distribution normalization processing include the following processes:

[0036] User damage calculation: The part failure pseudo-damage calculation analysis is performed according to the collected key damage correlation parameters by applying the part damage model formula. For example, the low-cycle fatigue calculation part load amplitude can be calculated by using the rainflow counting method to calculate the amplitude change size and frequency within the data time history. The torque amplitude damage is calculated as the 5th power of the torque amplitude * frequency / collection time, which is used to represent the low-cycle fatigue pseudo-damage. The calculation results are shown in Table 5, and Table 6 shows part of the key part damage model formula.

[0037]

[0038]

[0039] wherein, is the part temperature change amplitude; is the failure factor; is the burst pressure; is the engine speed; is the frequency under the corresponding load; is the frequency of engine shutdown ≥ 2s.

[0040] Damage distribution sorting: According to the pseudo-damage data of the N users calculated in step six, the cumulative distribution of all similar failure data is calculated. For example, the mean and variance can be calculated using EXCEL, and then the cumulative distribution value can be calculated using the NORM.DIST function. The cumulative distribution value can represent the damage degree. For example, Table 6 shows the calculation distribution using low-cycle fatigue torque fluctuation. The cumulative distribution function is to normalize the data to a value between 0 and 1, which is used for data sorting and value selection. In engineering, it is generally believed that a value between 0.9 and 0.95 represents a typical user, which can be used as a representative user of this type of label.

[0041]

[0042] Damage weighted calculation: It can be calculated in three categories. Reliability failure can be sorted and classified according to single failure mode damage. Different failure damages can be defined with different failure damage relative weights, such as low-cycle weight Z1 + high-cycle weight Z2 + friction and wear weight Z3 + aging weight Z4 = 1. The relative reliability failure weighting factor of different users is calculated. The functional failure parameters are calculated and sorted according to each failure. For example, as shown in Table 7.

[0043]

[0044] User clustering: according to the reliability damage weighting calculation factor or according to the cumulative distribution classification of a certain reliability failure mode, 0.8-1 is divided into high reliability failure risk, 0.6-0.8 is medium reliability failure risk, and <0.6 is low failure risk. User clustering is carried out according to the principle, combined with user label division, and the user portrait distribution characteristics based on reliability failure are determined.

[0045] In summary, the embodiment proposes a user big data portrait analysis method based on damage model, which is suitable for all engines. By collecting T-BOX data of user operation, combining common reliability damage and functional damage parameters of engine, through quantitative calculation and cumulative distribution normalization processing of damage parameters, the characteristic user portrait of common reliability and functional failure mode of engine is refined, the user use characteristic parameter is quantitatively described, based on the determined reliability damage and functional damage user distribution, the user label feature is clustered, the target input is provided for design simulation and test verification, and the precise development and verification are realized. Embodiment

[0046] Based on the basis of embodiment 1, the embodiment discloses a vehicle user big data portrait analysis system based on damage model, referring to Figure 2 As shown in the figure, including engine ECU, T-BOX vehicle terminal, cloud platform, client computer, engine ECU and T-BOX vehicle terminal are connected by wire communication, cloud platform is connected with T-BOX vehicle terminal and client computer by wireless communication, wherein:

[0047] The engine ECU is used for recording the actual running condition data of the vehicle, and transmitting the data to the cloud platform for storage through the T-BOX vehicle terminal;

[0048] The client computer is used for obtaining the vehicle running condition data stored in the cloud platform, and extracting the feature to form the user label according to the classification matrix;

[0049] The client computer is also used for extracting the damage correlation parameters of the engine according to the damage key factors matched by the user label;

[0050] The client computer is also used for quantitatively calculating and cumulatively distributing the damage correlation parameters, and obtaining the damage data;

[0051] The client computer is also used for sorting and clustering the user label portrait and representative user representing different types of engine failure according to the damage data.

[0052] The parameters required by the user label are all delivered to the cloud platform by the data stream recorded by the engine ECU through the T-BOX vehicle terminal, the corresponding data is downloaded in the enterprise, the pseudo-damage calculation is carried out according to various reliability / function failure models, the damage data is dimensionless, then the weighted calculation or direct sorting is carried out, the user represented by 0.9~0.95 data is selected as the representative of this kind of user label, and the failure characteristics represented thereby are the subsequent engine bench test target and design target input.

[0053] Although the present application has been described in detail above with specific implementation, some modifications or improvements can be made on the basis of the present application, which is obvious to those skilled in the art. Therefore, these modifications or improvements made on the basis of not deviating from the spirit of the present application, all belong to the scope of the present application claimed.

Claims

1. A vehicle user big data profiling method based on a damage model, characterized by, Comprise: S1: pre-built vehicle operation feature classification matrix, combined with the actual operation of the vehicle, based on the classification matrix feature extraction as user label; S2: according to the failure scenario of the engine, damage classification definition, damage key factor definition and damage correlation parameter definition, and according to the damage key factor matched by the user label, extract the damage correlation parameter; S3: according to the collected key damage correlation parameter, quantitative calculation and cumulative distribution normalization processing are carried out, and the damage data is obtained, and the user label portrait and representative user representing different damage classification of engine are sorted and clustered according to the damage data; Wherein, S3 carries out quantitative calculation and cumulative distribution normalization processing, including the following steps: S301: according to the collected key damage correlation parameter, the pseudo damage data value of the component failure is calculated by applying the component damage model formula; S302: according to the calculated pseudo damage data value, the cumulative distribution of all data belonging to the same damage classification is calculated, and the cumulative distribution value is calculated; S303: define the relative weight of different damage cumulative distribution value, and calculate the damage weighting factor of different users, which is used as damage data.

2. The vehicle user big data profiling method based on damage model according to claim 1, characterized in that, The operation characteristics of the vehicle in S1 include: user operation, external use environment, road feature, whole vehicle driving task profile, load characteristic. 3.The vehicle user big data profiling method based on damage model according to claim 1, wherein, In S2, the damage classification definition and damage correlation parameter definition include: according to the reliability failure mechanism of engine components and the engine function failure mechanism, the damage classification and damage correlation parameter are defined, and the damage classification is associated with the user label.

4. A vehicle user big data profiling system based on damage model, adopting the method of claims 1-3, characterized in that, Including engine ECU, T-BOX vehicle terminal, cloud platform, client computer, the engine ECU and T-BOX vehicle terminal have wired communication connection, the cloud platform is respectively connected with T-BOX vehicle terminal, client computer wireless communication connection, wherein: The engine ECU is used to record the actual operation data of the vehicle, and the T-BOX vehicle terminal is transmitted to the cloud platform storage; The client computer is used to obtain the vehicle operation data stored in the cloud platform, and the feature extraction is carried out according to the classification matrix to form the user label; The client computer is also used to extract the damage correlation parameter of the engine according to the damage key factor of the engine matched by the user label; The client computer is also used for quantitative calculation and cumulative distribution normalization processing of damage correlation parameter to obtain damage data; The client computer is also used for sorting and clustering the user label portrait and representative user representing different types of engine failure according to the damage data.

Images