CN109765878A

CN109765878A - The aided analysis method of new-energy automobile CAN bus network failure

Info

Publication number: CN109765878A
Application number: CN201811584342.7A
Authority: CN
Inventors: 周敏; 袁碧珍; 杜朝辉
Original assignee: Shanghai Dajun Technologies Inc
Current assignee: Shanghai Dajun Technologies Inc
Priority date: 2018-12-24
Filing date: 2018-12-24
Publication date: 2019-05-17

Abstract

The invention discloses a kind of aided analysis methods of new-energy automobile CAN bus network failure, this method acquires wave level data in the good situation of vehicle CAN network, wave level data are classified according to the content of CAN ID, respectively the wave level data of every one kind are calculated with the statistical distribution situation of characteristic information, and creates reference characteristic library；Operation vehicle CAN networking waveform level data is obtained in real time, classify to wave level data according to the content of CAN ID, respectively the wave level data of every one kind are calculated with the statistical distribution situation of characteristic information, and it is compared with reference characteristic library, obtains variance data and prompts failure risk feature.This method effectively extracts key feature data at vehicle scene, provides preliminary analytical judgment, provides foundation for the analysis of more detailed technology, promotes Analysis on Fault Diagnosis efficiency, reduce workload.

Description

The aided analysis method of new-energy automobile CAN bus network failure

Technical field

The present invention relates to a kind of aided analysis methods of new-energy automobile CAN bus network failure.

Background technique

With the development of new-energy automobile industry, running new energy vehicle ownership is constantly promoted, therefore runs vehicle The absolute quantity that CAN bus network failure occurs also constantly is increasing.

Traditional CAN bus network fault analyzing method is to be analyzed under lab according to design principle, is searched not Meet the place of design to search failure cause.But in the case where new-energy automobile quantity constantly rises, vehicle net is run Network malfunction is not constantly in stable situation；Time, equipment and technical staff's quantity, which all can not support, simultaneously adopts The method of careful analysis is carried out with complicated laboratory method to search analyzing failure cause.Thus it is badly in need of a kind of quick scene Operable method assists carrying out accident analysis and accurately searches reason.

When CAN bus breaks down, necessarily certain parameters in bus network are changed, but determine failure Point is one, and need will acquire measurement data the analysis of progress technology and the process that is combined with site environment.Therefore tradition Method is time-consuming and laborious, can not cope with the development trend of new-energy automobile.

Summary of the invention

Technical problem to be solved by the invention is to provide a kind of assistant analysis of new-energy automobile CAN bus network failure Method, this method effectively extract key feature data at vehicle scene, provide preliminary analytical judgment, are more detailed technology Analysis provides foundation, promotes Analysis on Fault Diagnosis efficiency, reduces workload.

In order to solve the above technical problems, the aided analysis method of new-energy automobile CAN bus network failure of the present invention includes Following steps:

Step 1: the CAN bus network for operational excellence obtains wave level sampled data；

Step 2: being classified according to CAN ID to wave level data, the wave level data of identical CAN ID are classified as one kind；

Step 3: calculating rising time, the difference waveform of CAN_H, CAN_L of mutually similar wave level data, difference waveform Failing edge time, high level voltage, low level voltage and position time；

Step 4: rising time distributed data T_rise, failing edge time point in the mutually similar wave level data of statistical analysis Cloth data T_drop, high level voltage distributed data L_hl, low level voltage distributed data L_ll and position Annual distribution data T_ bit；

Step 5: each distributed data in step 4 is stored as reference characteristic library, reference characteristic library includes point of distributed data The average value avg and standard deviation s of difference step, distributed data between cloth group number N, group data, use for subsequent referring to calculating；

Step 6: obtaining the wave level sampled data of operation vehicle CAN bus network in real time；

Step 7: the wave level sampled data that will acquire is classified according to CAN ID, the wave level of identical CAN ID is adopted Sample data are classified as one kind；

Step 8: calculating rising time, the difference waveform of CAN_H, CAN_L of mutually similar wave level data, difference waveform Failing edge time, high level voltage, low level voltage and position time；

Step 9: by the distribution situation for the phase homogeneous data being calculated in step 4 statistical analysis step 8；

Step 10: the distribution situation that step 9 obtains is compared with reference characteristic library, with the distribution group in reference characteristic library Difference step between number N, group data carrys out statistical data, calculates the distributed data average value avgx and standard deviation of each phase homogeneous data Sx obtains the variance data of rising time, failing edge time, high level voltage, low level voltage and position time；

Step 11: output variance data, when in avgx or sx and reference characteristic library avg or s differ by more than a step and When above, prompting this feature, there are abnormal risks, and transmit preservation.

Since the aided analysis method of new-energy automobile CAN bus network failure of the present invention uses above-mentioned technical proposal, i.e., originally Method acquires wave level data in the good situation of vehicle CAN network, to wave level data according in CAN ID Appearance is classified, and respectively the wave level data of every one kind are calculated with the statistical distribution situation of characteristic information, and creates benchmark Feature database；Operation vehicle CAN networking waveform level data is obtained in real time, and wave level data are carried out according to the content of CAN ID Classification calculates the wave level data of every one kind the statistical distribution situation of characteristic information respectively, and carries out with reference characteristic library Compare, obtain variance data and prompts failure risk feature.This method effectively extracts key feature data at vehicle scene, Preliminary analytical judgment is provided, foundation is provided for the analysis of more detailed technology, promotes Analysis on Fault Diagnosis efficiency, reduce work Load.

Detailed description of the invention

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments:

Fig. 1 is the Computer Aided Analysis System block diagram for realizing this method；

Fig. 2 is the step schematic diagram of accident analysis in this method；

Fig. 3 is the typical composition schematic diagram of CAN bus message data；

Fig. 4 is CAN bus waveform diagram；

Fig. 5 is typical statistic distribution schematic diagram；

Fig. 6 is the structural block diagram of comparison in difference in this method.

Specific embodiment

The aided analysis method of new-energy automobile CAN bus network failure of the present invention includes the following steps:

Step 5: each distributed data in step 4 is stored as reference characteristic library, reference characteristic library includes point of distributed data The average value avg and standard deviation s of difference step, distributed data between cloth group number N, group data, use for subsequent referring to calculating； General N can take 10, count all sampled datas, remove maximum and the smallest 0.5% data, use the maximum of remaining data Value subtracts minimum value and obtains difference delta, the difference step as organized between data that delta is obtained divided by N；

Step 10: the distribution situation that step 9 obtains is compared with reference characteristic library, with the distribution group in reference characteristic library Difference step between number N, group data carrys out statistical data, calculates the distributed data average value avgx and standard deviation of each phase homogeneous data Sx obtains the variance data of rising time, failing edge time, high level voltage, low level voltage and position time；With step 5 Identical, General N can take 10, count all sampled datas, remove maximum and the smallest 0.5% data, use remaining data Maximum value subtracts minimum value and obtains difference delta, the difference step as organized between data that delta is obtained divided by N；

This method is used for assistant analysis new-energy automobile CAN bus network failure, to make CAN bus Analysis of Network Malfunction It can preferably carry out, find failure source as early as possible.

As shown in Figure 1, this method by Computer Aided Analysis System realize, wherein SE01 be vehicle network, by CAN bus with The SE02 connection of waveform acquisition storage unit, while vehicle power supply is provided to equipment for taking electric unit SE04；Waveform acquisition storage is single First SE02 is used to acquire the wave level data on storage CAN bus network, prepares for wave level analysis；At waveform separation It manages cell S E03 and is used for waveform separation, which is split waveform installation message for unit, while carrying out according to CAN ID Classification；Equipment makes vehicle network SE01 and assistant analysis system for taking electric unit SE04 to provide electric power support for whole system The ground level of system is in identical state；The waveshape feature abstraction cell S E05 that classifies, which is calculated, generates differential level, calculating CAN_H, CAN_L and the rising time of differential level, failing edge time, high level voltage, low level voltage, the value of position time, and point Statistical distribution data are given birth in division；Characteristic storage unit SE06 carries out the storage of data and carries out difference point with reference characteristic library Analysis generates variance data and stores；Graphically display data difference assists point by data transmission and display unit SE07 The problem of analysing CAN bus network, can also carry out data transmission facilitating more advanced analyzing and diagnosing whenever necessary.

It is illustrated in figure 2 the entire step of accident analysis, STP01 first carries out the vehicle CAN bus wave of acquisition operation in real time Shape；STP02 is split waveform as unit of message；STP03 classifies using CAN ID as attribute；After STP04 extracts classification The feature of waveform；STP05 is compared according to the characteristic information of setting；The different information of STP06 acquisition characteristic information；STP07 Store characteristic information and feature difference information；STP08 is shown and transmission feature information data.

The typical composition of a CAN message data is illustrated as shown in Figure 3, by frame starting, arbitration section, control section, data Section, CRC sections and ACK sections of compositions, CAN ID data are existing in arbitration segment body.

Illustrate several characteristics of a CAN bus waveform as shown in Figure 4, rising time be waveform from low to high Time, generally 5% rises to for 95% time, and the failing edge time is the time of waveform from high to low, and generally 95% drops to 5% time, the time that the position time is each, waveform voltage value when high level voltage is in high level, low level voltage Voltage value when in low level.

The typical figure of a statistical distribution is illustrated as shown in Figure 5, and each pillar indicates that data exist in a data area Statistics ratio in all data, the distribution situation for characterize data.

Illustrate the structural block diagram of comparison in difference as shown in Figure 6, obtain current system CAN bus network spy information and CAN bus network feature information when system is good carries out differentiation comparison to each characteristic information, to prompt feature wind Danger.

This method is different from traditional direct DATA REASONING that CAN bus waveform is carried out using instruments such as oscillographs, technology The method that personnel are directly analyzed at the scene.This method can effectively promote the efficiency of analysis CAN bus network failure, simultaneously The workload that technical staff analyzes failure is substantially reduced, the frequency that carrying device instrument goes to vehicle scene is reduced, passes through this method Inspect periodically and can also carry out early warning to the generation of CAN bus network failure, rate of breakdown is effectively reduced.

Claims

1. a kind of aided analysis method of new-energy automobile CAN bus network failure, it is characterised in that this method includes following step It is rapid: