A method for identifying abnormal risks and fault determination of power battery sampling

Abstract

The present invention relates to the technical field of risk identification and discloses a method for identifying the risk of abnormal sampling in power batteries and a method for determining faults, comprising the following steps: Step 1: Collecting basic battery data; Step 2: Cleaning and preprocessing the basic battery data; Step 3: Extracting a sampling abnormality safety factor Cr based on the basic battery data processed in Step 2; the sampling abnormality safety factor Cr is a related projection variance parameter of the voltage difference change rate on the current change rate of a reference battery cell; Step 4: Performing nonlinear feature conversion and amplification processing on the sampling abnormality safety factor Cr to obtain an abnormality factor Sf; Step 5: Quantifying the sampling abnormality safety factor using variance entropy to obtain a quantitative feature. The present invention can achieve early and accurate identification of sampling abnormality risks and accurately identify and determine sampling abnormality faults.

Description

Technical Field

[0001] The present invention relates to the technical field of risk identification, and in particular to a method for identifying abnormal risks in power battery sampling and a method for determining faults. Background Art

[0002] With the increasing number of new energy vehicles, more and more safety issues of new energy vehicles have emerged. Among them, battery safety in new energy vehicles is one of the main safety issues of new energy vehicles today.

[0003] Among battery failure problems, sampling abnormality is a relatively common failure phenomenon. The sampling abnormality failure mode refers to the abnormality of the sampling function module caused by the failure of the battery cell sampling module. The battery cell sampling module involves battery cells, modules, sampling chips, wiring harnesses and their plug-ins, etc. The wiring harness is an important part of the single cell voltage sampling module and is also the part that is more likely to cause sampling abnormalities. There are many causes of sampling abnormalities, including glue dust on the wiring harness terminals, which causes the contact impedance to increase and the voltage sampling to be lost; the spring clips on the wiring harness locking terminals are deformed, resulting in the loss of voltage sampling; the wiring harness connector is withdrawn, resulting in abnormal voltage sampling, etc., which in turn causes changes in the resistance of the sampling line contact points, such as falling off, breakage, damage, and other problems.

[0004] Currently, the identification of sampling anomalies mostly relies on data extracted by the BMS system. However, the data reported by the BMS system may contain many false alarms. This is because data fluctuations are normal, and the BMS system will report the fluctuating data. This can easily lead to many sampling anomalies, resulting in false alarms and inaccurate anomaly determination. Furthermore, due to the changing environments and complex scenarios of the actual operation of new energy vehicles today, their operating data is multidimensional, redundant, heterogeneous, and strongly coupled. Coupled with the physical characteristics of the battery system itself, as well as the design and acquisition accuracy of the sensors, information coupling, redundancy, and errors are inevitable between different signal data, making data analysis and anomaly identification more difficult and increasing errors.

[0005] In addition, the types of faults that power batteries are prone to (i.e., the types of risks that exist) are diverse. In addition to sampling abnormalities, there are also connection abnormalities, self-discharge abnormalities, capacity abnormalities, internal resistance abnormalities, etc. The occurrence of these faults will intuitively cause abnormal fluctuations in the power battery operating data, and the similarity between these fluctuations is very high, which makes it relatively easy to determine the existence of risks based on operating data. However, it is very difficult to accurately determine what kind of fault (risk type) it is from the operating data, because it is impossible to distinguish the risk type from similar numerical representations, and it is very difficult to identify sampling abnormalities in a targeted manner. Summary of the Invention

[0006] The present invention aims to provide a method for identifying abnormal sampling risks and a method for determining faults in power battery sampling, which can achieve early and accurate identification of abnormal sampling risks and can accurately identify and determine abnormal sampling faults.

[0007] To solve the above problems, the present invention provides the following basic solutions:

[0008] Option 1:

[0009] A method for identifying abnormal risk and fault determination of power battery sampling includes the following steps:

[0010] Step 1: Collect basic battery data;

[0011] Step 2: Clean and pre-process the basic battery data;

[0012] Step 3: Based on the basic battery data processed in step 2, extract the sampling abnormality safety factor Cr; the sampling abnormality safety factor Cr is the related projection variance parameter of the voltage difference change rate of the reference battery cell on the current change rate;

[0013] Step 4: Perform nonlinear feature conversion and amplification on the sampled abnormal safety factor Cr, and obtain the abnormal factor Sf; and Sf = e αCr ; Where α is the signal amplification factor;

[0014] Step 5: Use variance entropy to quantify the sampling anomaly safety factor and obtain a quantitative feature; the smaller the quantitative feature value is, the smaller the fluctuation degree of the sampling anomaly feature is on the time scale, and the smaller the anomaly risk is.

[0015] Option 2:

[0016] A method for determining abnormal faults in power battery sampling includes the following steps:

[0017] S1: Risk identification is performed using a power battery sampling abnormality risk identification method as described in Scheme 1;

[0018] S2: Compare the risk probability value with the judgment threshold, and take the time point when the risk probability value is greater than the judgment threshold as the high-risk point;

[0019] S3: Extract abnormal element values ​​and voltage values ​​within a preset time period before and after the high-risk point and draw them into a risk image;

[0020] S4: Determine sampling abnormality faults based on the fluctuation of each numerical curve in the risk image.

[0021] The working principle and advantages of the present invention are:

[0022] This solution combines the knowledge of the mechanism of sampling anomalies with data mining related technologies. Starting from the corresponding mechanism changes of voltage signal data when sampling anomalies occur, by designing the characterization features of sampling anomalies on voltage signal data, that is, the relevant projection variance parameters of the voltage difference change rate corresponding to the reference battery cell on the current change rate, anomalies that are difficult to capture from complex data are extracted in the form of sampling anomaly safety factors, and nonlinear feature conversion and amplification processing, as well as variance entropy, are used to process them into dimensionless quantitative feature data, which can intuitively represent the fluctuation and risk of abnormal characteristics on the time scale, and thus can effectively realize the data expression and automatic identification of fault information. Based on the vehicle operation history data, it can realize the accurate identification of new energy vehicle risks and accurate determination of sampling anomaly faults.

[0023] What is special is that when identifying risks and determining anomalies, this solution does not directly capture abnormal data, but further captures the fluctuation of specific characteristics of voltage signal data during the generation process of sampling anomalies or when sampling anomalies exist, that is, the relevant projection variance parameter of the voltage difference change rate on the current change rate. This parameter actually describes the mechanism state corresponding to the sampling anomaly (that is, the voltages of adjacent battery cells are one too high and one too low, the voltages of adjacent battery cells are offset in opposite directions, etc.). By describing and mining the mechanism state to perform feature mining, sampling anomaly failures can be identified in a targeted and accurate manner.

[0024] Moreover, this parameter reflects pure characteristic fluctuations. While being relatively unaffected by the numerical deviations of the data itself, it can also capture minute anomalies that cannot be reflected by simple voltage values, as well as minute anomalies that are concealed by numerical coupling errors. This parameter helps achieve early and accurate identification of sampling anomaly risks, that is, in the process of sampling anomaly evolution and formation, it can precisely identify the risk of sampling anomalies and precisely mine sampling anomalies in similar numerical appearances. By calculating and extracting this parameter (i.e., the sampling anomaly safety factor), it is possible to effectively eliminate redundant information and noise information generated by external factors in the power battery characterization signal, and more accurately extract the signal characteristics that characterize the power battery's operating status. The resulting safety factor can effectively and accurately describe the characteristic changes in the voltage signal during vehicle operation, i.e., during the operation of the power battery, thereby helping to accurately identify and determine sampling anomaly faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic flow chart of a risk identification method according to an embodiment of a method for identifying abnormal risk of power battery sampling and a method for determining faults of the present invention;

[0026] Figure 2This is a schematic diagram of a fault determination method flow in accordance with an embodiment of a method for identifying abnormal risk of power battery sampling and a method for determining faults according to the present invention;

[0027] Figure 3 This is a risk image schematic diagram of an embodiment of a method for identifying abnormal risk of power battery sampling and a method for determining faults according to the present invention. DETAILED DESCRIPTION

[0028] The following is a further detailed description through specific implementation methods:

[0029] The embodiment is basically as shown in the attached Figure 1 A method for identifying abnormal risk and fault determination of power battery sampling is shown, comprising the following steps:

[0030] Step 1: Collect basic battery data; the basic battery data is parsed from the message log of the electric power battery. Specifically, the basic battery data in this embodiment is parsed from the message log of the electric power battery system that complies with the GB32960 protocol, and the basic data is highly reliable.

[0031] Furthermore, this embodiment analyzes power batteries for new energy vehicles operating in complex and volatile environments. The corresponding operating data is multidimensional, redundant, heterogeneous, and highly coupled, making it difficult to analyze. However, the following steps in this solution enable accurate and targeted early identification of sampling anomaly risks within this complex data.

[0032] Step 2: Perform data cleaning and data preprocessing on the basic battery data. Specifically, the data cleaning is to clean out the abnormal data in the basic battery data. The abnormal data specifically refers to the part of the data in the voltage and current signals of the basic battery data that exceeds the specified threshold. The data preprocessing includes: (1) Interference pulse identification and marking. If the difference between the current frame voltage data and the previous frame exceeds the specified threshold, the frame data is marked to facilitate distinction during subsequent evaluation; (2) Time discontinuity point identification and marking: If the difference between the current frame timestamp data and the previous frame exceeds the specified threshold, the frame data is marked to facilitate distinction during subsequent evaluation; specifically, during subsequent evaluation, the risk of the marked point is set to 0, which helps to improve the accuracy of risk identification. (3) Mean filtering to reduce noise data.

[0033] Step 3: Based on the basic battery data processed in step 2, extract the sampling abnormality safety factor Cr; the sampling abnormality safety factor Cr is the related projection variance parameter of the voltage difference change rate of the reference battery cell on the current change rate.

[0034] The reference cell is selected using the following sub-steps:

[0035] Sub-step 1: Calculate the difference matrix Vd of adjacent cells of the power battery;

[0036] Sub-step 2: Construct a symbolic function; the symbolic function is

[0037]

[0038] Among them, Vaa is the extreme voltage, that is, the difference between the maximum value and the minimum value of the single cell voltage at any time; α∈(0,1) is the sampling abnormality characteristic coefficient.

[0039] Sub-step 2.1: Perform mean filtering on the sign function to smooth the step sign function to achieve the effect of expanding the step position signal.

[0040] Sub-step 3: Calculate the sign function offset of each battery cell and sort them, and select the battery cell corresponding to the largest offset value as the reference battery cell.

[0041] The reference cell selected through the above steps is the cell that best reflects the fluctuation of characteristics (identification factors) in the power battery system. Compared with the general analysis of all cells, this solution effectively limits the analysis scope to a single reference cell, greatly reducing the data processing workload. The selected cell is also very representative and can effectively make up for the analysis accuracy gap after the analysis is streamlined.

[0042] The sampling abnormal safety elements

[0043] Where,

[0044] V v Indicates the vector corresponding to the position of the reference cell in the differential matrix, I v Represents the current change rate vector of the reference cell.

[0045] Step 4: Perform nonlinear feature conversion and amplification on the sampled abnormal safety factor Cr, and obtain the abnormal factor Sf; and Sf = e αCr ; Where α is the signal amplification factor;

[0046] Step 5: Use variance entropy to quantify the sampling anomaly safety factor and obtain a quantitative feature; the smaller the quantitative feature value is, the smaller the fluctuation degree of the sampling anomaly feature is on the time scale, and the smaller the anomaly risk is.

[0047] Specifically, the quantized feature p=1-λ; where λ=E 2 (Sf) / ( 2 ); 0≤λ≤1 and the closer λ is to 1, the smaller p is. On the time scale, the smaller the fluctuation of the sampling abnormal characteristics is, and the smaller the abnormal risk is.

[0048] Step 6: Based on the quantitative characteristics, analyze and obtain the risk probability value.

[0049] Specifically, discrete integration is performed on p on the time scale to obtain a discrete integral function Sp = ∑p, where Sp is a monotonically increasing curve. The amplitude of the slope of the discrete integral function Sp curve is used as the risk probability value. The larger the risk probability value, the greater the fluctuation degree of the sampling abnormality characteristics and the greater the abnormality risk.

[0050] As attached Figure 2 As shown, a method for determining abnormal faults in power battery sampling includes the following steps:

[0051] S1: Risk identification is performed using a power battery sampling abnormality risk identification method as described above;

[0052] S2: Compare the risk probability value with the determination threshold, and take the time point corresponding to when the risk probability value is greater than the determination threshold as a high-risk point; in this embodiment, the determination threshold is set to 0.5. The threshold is set appropriately and can effectively identify high-risk points.

[0053] S3: Extract the abnormal element values ​​and voltage values ​​in the preset time period before and after the high-risk point and draw them into a risk image; as shown in the attached figure. Figure 3 shown.

[0054] S4: Determine sampling abnormality faults based on fluctuations of each numerical curve in the risk image.

[0055] When the fluctuation of the voltage value curve is manifested as the voltage of several adjacent battery cells shifting in two opposite directions, it is determined that there is a sampling abnormality fault. Specifically, when the voltages of two adjacent battery cells shift in two opposite directions, that is, compared with other normal battery cells, the voltages of the two battery cells are obviously higher and lower, and no single battery cell in the two battery cells has the problem of high charging and low discharge (high charging voltage and low discharge voltage), then it is determined that there is a sampling abnormality fault. The judgment method is intuitive and effective.

[0056] This embodiment provides a method for identifying and diagnosing power battery sampling anomaly risks. By combining mechanism knowledge with data mining techniques, this method effectively implements data representation and automated identification of fault information by designing characterization features for sampling anomalies in voltage signal data. Furthermore, by constructing a technical system for identifying and diagnosing the safety status of new energy vehicles and fault mode, encompassing data processing, feature extraction, feature quantification, risk identification, and fault diagnosis, this method accurately identifies risks and diagnoses sampling anomaly faults in new energy vehicles based on historical vehicle operation data.

[0057] The above is only an embodiment of the present invention. Common knowledge such as the specific structure and characteristics of the scheme is not described in detail here. Ordinary technicians in the relevant field are aware of all common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the guidance of this application. Some typical well-known structures or well-known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention. These will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. A method for identifying abnormal risk of power battery sampling, characterized in that: The following steps are involved: Step 1: Collect basic battery data; Step 2: Clean and pre-process the basic battery data; Step 3: Extract sampling abnormal safety factors based on the basic battery data processed in step 2 ; The sampling abnormal safety elements is the related projection variance parameter of the voltage difference change rate of the reference cell on the current change rate; The sampling abnormal safety elements ; Where, ; ; Represents the vector corresponding to the position of the reference cell in the differential matrix, Represents the current change rate vector of the reference cell; Step 4: Sampling abnormal safety factors Perform nonlinear feature conversion and amplification processing to obtain abnormal factors ;and ;in, is the signal amplification factor; Step 5: Use variance entropy to quantify the sampling abnormal safety factors and obtain quantitative features; quantitative features ;in, ; ; The smaller the quantitative characteristic value is, the smaller the fluctuation of the sampling abnormality characteristics is on the time scale, and the smaller the abnormality risk is.

2. The method for identifying abnormal risk of power battery sampling according to claim 1, characterized in that: In step 1, basic battery data is parsed from the message log of the power battery.

3. The method for identifying abnormal risk of power battery sampling according to claim 1, characterized in that: The method further includes step 6: analyzing and obtaining a risk probability value based on the quantitative characteristics.

4. A method for determining abnormal faults in power battery sampling, characterized in that: The following steps are involved: S1: Perform risk identification using the power battery sampling abnormality risk identification method as described in claim 3; S2: Compare the risk probability value with the judgment threshold, and take the time point when the risk probability value is greater than the judgment threshold as the high-risk point; S3: Extract abnormal element values ​​and voltage values ​​within a preset time period before and after the high-risk point and draw them into a risk image; S4: Determine sampling abnormality faults based on the fluctuation of each numerical curve in the risk image.

5. A method for determining abnormal faults in power battery sampling according to claim 4, characterized in that: When the fluctuation of the voltage value curve is manifested as the voltages of several adjacent battery cells shifting in two opposite directions, it is determined that there is a sampling abnormality fault.

Images