A power battery energy consistency risk assessment method and system

Abstract

The present invention relates to the field of electric vehicle battery risk assessment, and specifically to a power battery energy consistency risk assessment method and system. The method includes obtaining operation history data and performing initialization processing to obtain standardized data, extracting initial safety factors based on the standardized data to calculate safety factors, quantifying the safety characteristics of the safety factors to obtain safety characteristic quantization values, summing the safety characteristic quantization values ​​to accumulate risks to obtain risk accumulation curve data pairs, calculating risk accumulation characteristics based on the risk accumulation curve data pairs, and performing vehicle risk warnings based on the risk accumulation characteristics. The system includes a receiving module and a processing module, the receiving module receives the operation history data, and the processing module performs vehicle risk assessment on the operation history data according to the assessment method. The present invention completes risk characteristic calculation and safety quantification during vehicle operation, thereby achieving accurate judgment of the vehicle's comprehensive state and fault identification.

Description

Technical Field

[0001] The present invention relates to the field of electric vehicle battery risk assessment, and in particular to a power battery energy consistency risk assessment method and system. Background Art

[0002] With the adjustment of China's energy structure, the rapid development of technologies such as batteries, motors, and electronic controls, shifting consumer demand, and the introduction and improvement of supporting industrial policies, the production and sales of new energy vehicles in my country have continued to grow in recent years. However, the actual operating environment of new energy vehicles is changeable and complex, and the data representing their safety status is multidimensional, redundant, heterogeneous, and strongly coupled. This poses a significant challenge to exploring and mining vehicle safety features embedded in operational big data. Mechanistic analysis is difficult to assess the status of new energy vehicle power batteries in these strongly coupled, complex systems. Therefore, data-driven safety feature extraction methods combined with mechanistic knowledge can quantitatively describe vehicle risk status from specific dimensions.

[0003] Since vehicles will upload operating data in accordance with the GB-32960 standard during their service, and characterization signals such as voltage and current are a representation of the comprehensive state of the battery, the degree of difference in the state of the power battery during service can be further quantified by analyzing and mining the safety characteristics of the vehicle's historical operation data.

[0004] However, considering the physical properties of the battery system itself, as well as factors such as sensor design and acquisition accuracy, there will inevitably be information coupling, redundancy and errors between different signals, making it difficult to accurately extract and quantify safety features. Summary of the Invention

[0005] The present invention aims to provide a power battery energy consistency risk assessment method to solve the information coupling, redundancy and error problems between different signals when extracting and quantifying safety features of operating data uploaded by vehicles.

[0006] The power battery energy consistency risk assessment method in this solution includes:

[0007] Step 1: Obtain the vehicle's operating history data and initialize the operating history data to obtain standardized data;

[0008] Step 2: extracting initial safety factors from the standardized data according to the calculation model, and performing power operation based on the velocity matrix of the initial safety factors and a preset constant to obtain the safety factors;

[0009] Step 3: Based on the safety factor matrix formed by the safety factors, the ratio of the squared mean of the cell voltage of each parameter at each moment to the squared mean of the cell voltage of each parameter at each moment in the safety factor matrix is ​​used as the quantized safety feature vector;

[0010] Step 4: Sum the values ​​in the safety feature vector to obtain a risk accumulation curve data pair. Use the speed curve vector of the risk accumulation curve data pair as the risk feature. Calculate the relative risk based on the risk feature. Use the maximum value in the risk feature vector as the absolute risk. Use the number of times the absolute risk exceeds the absolute risk threshold as the risk frequency. Use the relative risk, absolute risk, and risk frequency as the risk accumulation feature.

[0011] Step 5: Determine whether the risk accumulation feature meets any warning condition. If so, determine that the vehicle is a high-risk vehicle.

[0012] The beneficial effects of this program are:

[0013] By constructing the energy consistency abnormal risk characteristics and combining them with the vehicle operation history data to complete the extraction of energy safety factors, it is possible to complete the risk characteristic calculation and safety quantification during the vehicle operation process, eliminate information coupling, redundancy and error, and thereby achieve accurate judgment and fault identification of the vehicle's comprehensive status, timely discover the fault risk during the vehicle operation process, and improve the accuracy and timeliness of risk tracing.

[0014] Furthermore, in step 1, the initialization process includes:

[0015] a) Signal boundary value limitation: remove abnormal data of voltage and current signal data exceeding the first specified threshold;

[0016] b) Interference pulse identification and marking: If the difference between the current frame voltage data and the previous frame voltage data exceeds a second specified threshold, the frame data is marked;

[0017] c) Time discontinuity identification and marking: if the difference between the timestamp data of the current frame and the timestamp data of the previous frame exceeds a third specified threshold, the frame data is marked;

[0018] d) Mean filtering.

[0019] The beneficial effects are: through the initialization processing of the operation history data, the signal boundary value limitation can remove individual obviously abnormal data, the identification and marking of interference pulses and time discontinuity points, which is convenient for subsequent processing, and can standardize the operation history data from multiple aspects, effectively preventing extremely individual abnormal data from interfering with the processing and analysis results, so as to facilitate subsequent processing and improve the accuracy of subsequent data processing.

[0020] Furthermore, in step 2, the calculation model is expressed as:

[0021] s=250*(VV m ) 2 +(V+0.35) 2 ;

[0022] Where s is the initial safety factor; V is the cell voltage; V m is the median voltage, i.e., the median voltage; 250 is a constant used to amplify the median voltage difference signal to an amplitude range similar to the voltage; 0.35 is the voltage value offset, which is used to unify the amplitude range of the voltage of different types of battery cells;

[0023] The expression formula of the power operation is:

[0024] S=e (0.2*s′) , where S is the safety factor and s′ is the velocity matrix of the initial safety factor.

[0025] The beneficial effect is that the operation history data of different units can be unified into the same dimension through the calculation in step 2, reducing the interference of different data on the analysis results, so as to accurately analyze the data and obtain accurate safety factors.

[0026] Furthermore, in step 3, the eigenvalue of the security feature vector is expressed as:

[0027]

[0028] Where p is the quantized value of the security feature, and 0≤p≤1; is the one-dimensional vector obtained by squaring the mean value of each cell voltage at each moment in the safety factor matrix. It is a one-dimensional vector obtained by averaging the square of each cell voltage in the safety factor matrix at each moment.

[0029] The beneficial effect is: by quantifying safety features, data analysis becomes more intuitive and historical operation data can be analyzed in a unified manner.

[0030] Furthermore, in step 4, the risk accumulation curve data pair is expressed as:

[0031]

[0032] Among them, Sp is the risk accumulation curve data pair, N is the data length, that is, the number of sampling points; p i is the quantified value of the security feature at the i-th moment;

[0033] The risk characteristics are expressed as:

[0034] r = Sp′, where r is the risk characteristic and Sp′ is the speed curve vector of the risk accumulation curve;

[0035] The relative risk calculation process includes sorting the risk characteristics from large to small, and taking the ratio of the average of the risk characteristics before the first threshold value after sorting to the average of the risk characteristics after the second threshold value after sorting as the relative risk;

[0036] Absolute risk is expressed as:

[0037] Q = max(r);

[0038] Where Q is the absolute risk; max() is the maximum value in the risk feature vector;

[0039] The risk frequency is expressed as:

[0040] R = count(r>r0);

[0041] Among them, R is the risk frequency, r0 is the absolute risk threshold, and the risk frequency is the number of times the absolute risk is greater than the threshold.

[0042] The beneficial effects are: calculating the quantified data to obtain the cumulative risk characteristics, and evaluating the risk situation of vehicle abnormalities from multiple aspects such as relative risk, absolute risk and risk frequency. The evaluation is more complete, so as to accurately extract the characteristics that characterize vehicle risks from a large amount of historical operating data.

[0043] Furthermore, the relative risk is expressed as:

[0044]

[0045] Where η is the relative risk, r max After ranking risk features, the mean of the top 20% risk feature values; r min It is the mean of the risk feature values ​​of the last 60% after risk features are sorted.

[0046] The beneficial effect is that the sorted risk characteristics can show different risk information in different segments, and select different numbers of characteristic values ​​to describe the relative risk, accurately intercept the risk situation of different data, improve the differentiation effect, and thus calculate the accurate relative risk.

[0047] Furthermore, in step 5, the warning conditions include:

[0048] (1) Q>Q0;

[0049] (2), η>η0|Q>Q 01 |R>R0.

[0050] The beneficial effect is that by setting early warning conditions for cumulative risk characteristics, it is possible to warn of abnormal vehicle risks in different situations, so as to accurately and comprehensively locate vehicle risks.

[0051] The power battery energy consistency risk assessment system includes a receiving module and a processing module. The receiving module obtains the operation history data uploaded by the vehicle and sends it to the processing module. The processing module receives the operation history data from the receiving module and stores it.

[0052] The processing module stores computer executable instructions, which, when executed, follow the steps of the above method and evaluate vehicle risks based on operating history data. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flowchart of the power battery energy consistency risk assessment method of the present invention;

[0054] Figure 2 This is a schematic diagram of risk details obtained by simulating the power battery energy consistency risk assessment method of the present invention;

[0055] Figure 3 This is a second schematic diagram of risk details obtained by simulating the power battery energy consistency risk assessment method of the present invention. DETAILED DESCRIPTION

[0056] The following is further explained in detail through specific implementation methods.

[0057] Example

[0058] Energy consistency risk assessment method for power batteries. Energy consistency refers to the description of the difference in equivalent energy released or absorbed by each battery cell within a certain time range. A certain time range refers to a time window, such as a ten-minute time window. Figure 1 As shown, including:

[0059] Step 1: Obtain the vehicle's operating history data and initialize the operating history data to obtain standardized data. The operating history data is obtained from the data uploaded by the vehicle according to actual needs. The initialization process includes:

[0060] a) Signal boundary value limitation: remove abnormal data of voltage and current signal data exceeding the first specified threshold;

[0061] b) Interference pulse identification and marking: If the difference between the current frame voltage data and the previous frame voltage data exceeds the second specified threshold, which is set based on the voltage performance of the actual vehicle when it is abnormal, the frame data will be marked;

[0062] c) Time discontinuity identification and marking: if the difference between the timestamp data of the current frame and the timestamp data of the previous frame exceeds a third specified threshold, the frame data is marked;

[0063] d) Mean filtering: Mean filtering is performed using an existing program.

[0064] When subsequently quantifying the security features, the security feature values ​​at the locations marked in b) and c) are set to 1, effectively reducing the interference of the data anomaly itself on the analysis results. The first specified threshold, the second specified threshold, and the third specified threshold are set according to actual needs.

[0065] Step 2: Extract initial security factors from standardized data according to the calculation model. The calculation model is expressed as:

[0066] s=250*(VV m ) 2 +(V+0.35) 2 ;

[0067] Where s is the initial safety factor; V is the cell voltage; V m is the median voltage, i.e., the median voltage; 250 is a constant used to amplify the median voltage difference signal to an amplitude range similar to the voltage; 0.35 is the voltage value offset, which is used to unify the amplitude range of the voltage of different types of battery cells;

[0068] The safety factor is obtained by performing a power operation based on the velocity matrix of the initial safety factor and the preset constant. The power operation is based on e and the expression formula of the power operation is:

[0069] S=e (0.2*s′) Where S is the safety factor, s′ is the velocity matrix of the initial safety factor, and the velocity matrix is ​​obtained using a velocity filtering method. The velocity filtering method specifies a time window with a fixed step size, calculates the difference between the target matrix's end position and its starting position in the current time window, and then takes the ratio of this difference to the time window step size to obtain the velocity matrix of the target matrix. In this embodiment, the target matrix is ​​the matrix formed by the initial safety factor. The fixed step size is set as required, for example, a fixed step size of ten minutes.

[0070] Step 3: Based on the safety factor matrix formed by the safety factors, the ratio of the squared mean of the single voltage of each parameter at each moment to the squared mean of the single voltage of each parameter at each moment is used as the quantized safety feature vector. Each moment refers to each sampling point, and the eigenvalue of the safety feature vector is expressed as:

[0071]

[0072] Where p is the quantized value of the security feature, and 0≤p≤1; is the one-dimensional vector obtained by squaring the mean value of each cell voltage at each moment in the safety factor matrix. This is a one-dimensional vector obtained by averaging the squared voltages of each cell at each moment in the safety factor matrix. By representing the eigenvalues ​​of the safety feature vector and maintaining them within the range of (0,1), we can uniformly evaluate and assess different data while preserving key features within the data, facilitating further assessment of consistency risks.

[0073] Step 4: Sum the values ​​in the security feature vector to obtain a risk accumulation curve data pair, which is expressed as:

[0074]

[0075] Among them, Sp is the risk accumulation curve data pair, N is the data length, that is, the number of sampling points; p i is the quantized value of the security feature at the i-th moment.

[0076] The speed curve vector of the risk accumulation curve data pair is used as the risk feature, and the risk feature is expressed as:

[0077] r=Sp′, where r is the risk characteristic, Sp′ is the speed curve vector of the risk accumulation curve, and the speed curve vector is also obtained using the speed filtering method.

[0078] The relative risk is calculated based on the risk characteristics. The relative risk calculation process includes sorting the risk characteristics from large to small, taking the ratio of the average risk characteristics before the first threshold value after sorting to the average risk characteristics after sorting, as the relative risk. The first threshold value is 20%, and the second threshold value is 60%. The relative risk is expressed as:

[0079]

[0080] Where η is the relative risk, r max After ranking risk features, the mean of the top 20% risk feature values; r min It is the mean of the risk feature values ​​of the last 60% after risk features are sorted.

[0081] The maximum value in the risk feature vector is taken as the absolute risk, and the absolute risk is expressed as:

[0082] Q = max(r);

[0083] Among them, Q is the absolute risk; max() is the maximum value in the risk feature vector.

[0084] The number of times the absolute risk is greater than the absolute risk threshold is taken as the risk frequency, which is expressed as:

[0085] R = count(r>r0);

[0086] Among them, R is the risk frequency, r0 is the absolute risk threshold, and the risk frequency is the number of times the absolute risk is greater than the threshold.

[0087] Relative risk, absolute risk and risk frequency are used as risk accumulation characteristics.

[0088] Step 5: Determine whether the risk accumulation characteristics meet any warning conditions. Warning conditions include:

[0089] (1)Q>Q0;

[0090] (2)η>η0|Q>Q 01 |R>R0;

[0091] If (1) or (2) is satisfied, the vehicle is judged as a high-risk vehicle. Where Q is the absolute risk, which indicates the absolute size of the local risk; η is the relative risk, which indicates the degree of the overall risk increase; R is the risk frequency, which indicates the number of times high risk occurs. Q0 = 0.005 means that when the absolute risk quantization value of the vehicle at any time exceeds 0.005, the vehicle is judged as a high-risk vehicle; η0 = 5, Q 01 =0.0025, R0=3 means that when a vehicle satisfies all three conditions, it is determined to be a high-risk vehicle. In summary, for these two determination conditions, a vehicle can be determined to be a high-risk vehicle if any one of them is met.

[0092] Taking the actual operation history data of a broken-down vehicle as an example, assuming that the analysis of the existing data is a real-time monitoring and analysis during the operation of the vehicle, then when a high risk occurs, the corresponding early warning information will be given and the simulation will be carried out. Figure 2 and Figure 3 The risk details shown are from Figure 2 It can be seen that the risk detail 1: 2020-07-16 16:52:04, the vehicle's driving intensity was slightly greater than before (SOC < 45). Observing the local details of SOC at this moment, the voltage of cell 33 dropped significantly, the pressure difference increased rapidly, and the SOC showed a slight drop. Figure 3 Risk detail 2: At 4:15:05 PM on September 10, 2020, the vehicle's usage intensity increased significantly compared to before (SOC < 40). Observing the SOC details at this moment, we can see that cell 33 experienced a significant drop in voltage, the pressure differential increased rapidly, and the SOC dropped to 0 very quickly. Simulation results show that the risk accumulation model for this vehicle can provide early warning and identify risks two months in advance of a breakdown.

[0093] Due to the limitations of the progress of battery mechanism research and the low feasibility of actual battery detection operations, the fault diagnosis of power batteries has always been limited by objective factors. For several typical fault modes, such as self-discharge anomaly, abnormal capacity attenuation, connection anomaly, sampling anomaly, etc., the accuracy and feasibility of abnormal detection and fault judgment during vehicle operation are also unsatisfactory. When some faults or anomalies occur in the vehicle, the corresponding risk causes cannot be accurately found. On the basis of effectively shielding data anomalies, this embodiment constructs energy consistency risk features, combines vehicle operation history data to complete the extraction of energy safety factors, and completes the energy consistency feature extraction and quantitative identification during vehicle operation. The difference in stored energy or released energy generated by the battery cell during charging or discharging in the same time period is used. The greater the difference, the worse the battery cell consistency, which will lead to uneven output or energy storage of the battery cell in the battery system, thereby leading to vehicle safety risks. Then, the risk accumulation model is combined to complete the determination of the vehicle safety status and output high-risk point image detail information. The risk accumulation algorithm is combined to realize the identification of vehicle risk features, and the image information of high-risk points is extracted to complete the risk tracing analysis. The overall method is applied to the analysis of operating data reported by actual running vehicles. By extracting and quantifying the safety features of the operating data uploaded by the vehicles, the information coupling, redundancy and error between different signals are eliminated. Based on the various model settings in the method and the corresponding parameter settings in the model, the failure risk of the vehicle can be warned in advance and accurately before the vehicle fails, thereby improving the safety of vehicle operation.

[0094] Example 2

[0095] A power battery energy consistency risk assessment system includes a receiving module and a processing module. The receiving module obtains the operation history data uploaded by the vehicle and sends it to the processing module. The receiving module can use the on-board network module on the existing electric vehicle to transmit the operation history data. The processing module can be a background processor for analyzing and processing the operation history data. The processing module receives the operation history data from the receiving module for storage; the processing module stores computer executable instructions. When the computer executable instructions are executed, they follow the steps of the power battery energy consistency risk assessment method as described in Example 1 and assess vehicle risks based on the operation history data.

[0096] By using the power battery energy consistency risk assessment method of Example 1 equipped with the system of this embodiment, it is possible to process and analyze the operating data reported in real time by the running vehicle, extract and quantify the safety features of the operating data uploaded by the vehicle, eliminate the information coupling, redundancy and error between different signals, and then perform risk analysis to provide early warning of fault risks during the real-time operation of the vehicle, and to provide early warning of vehicle fault risks, thereby improving the safety of vehicle operation.

[0097] The above is only an embodiment of the present invention, and the common knowledge such as the specific structure and characteristics of the scheme is not described in detail here. 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, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A power battery energy consistency risk assessment method, characterized in that: include: Step 1: Obtain the vehicle's operating history data and initialize the operating history data to obtain standardized data; Step 2: Extract the initial safety factors from the standardized data according to the calculation model, and perform power operation based on the speed matrix of the initial safety factors and the preset constant to obtain the safety factors. The power operation is expressed as follows: ,in, For safety factors, is the velocity matrix of the initial safety factors; Step 3: Based on the safety factor matrix formed by the safety factors, the ratio of the squared mean of the cell voltage of each parameter at each moment to the squared mean of the cell voltage of each parameter at each moment in the safety factor matrix is ​​used as the quantized safety feature vector; Step 4: Sum the values ​​in the safety feature vector to obtain a risk accumulation curve data pair. Use the speed curve vector of the risk accumulation curve data pair as the risk feature. Calculate the relative risk based on the risk feature. Use the maximum value in the risk feature vector as the absolute risk. Use the number of times the absolute risk exceeds the absolute risk threshold as the risk frequency. Use the relative risk, absolute risk, and risk frequency as the risk accumulation feature. Step 5: Determine whether the risk accumulation feature meets any warning condition. If so, determine that the vehicle is a high-risk vehicle.

2. The power battery energy consistency risk assessment method according to claim 1, characterized in that: In step 1, the initialization process includes: a) Signal boundary value limitation: remove abnormal data whose voltage exceeds the first specified threshold; b) Interference pulse identification and marking: If the difference between the current frame voltage data and the previous frame voltage data exceeds a second specified threshold, the frame data is marked; c) Time discontinuity identification and marking: if the difference between the timestamp data of the current frame and the timestamp data of the previous frame exceeds a third specified threshold, the frame data is marked; d) Mean filtering.

3. The power battery energy consistency risk assessment method according to claim 2, characterized in that: In step 2, the calculation model is expressed as: ； in, It is the initial safety factor; is the single cell voltage; is the median voltage, i.e., the median voltage; 250 is a constant used to amplify the median voltage difference signal to an amplitude range similar to the voltage; 0.35 is the voltage value offset, which is used to unify the amplitude range of the voltage of different types of battery cells.

4. The power battery energy consistency risk assessment method according to claim 1, characterized in that: In step 4, the risk accumulation curve data pair is expressed as: ； in, is the risk accumulation curve data pair, is the data length, that is, the number of sampling points; For the Quantitative value of security characteristics at a moment; The risk characteristics are expressed as: ,in, is the risk characteristic, is the speed curve vector of the risk accumulation curve; The relative risk calculation process includes sorting the risk characteristics from large to small, and taking the ratio of the average of the risk characteristics before the first threshold value after sorting to the average of the risk characteristics after the second threshold value after sorting as the relative risk; Absolute risk is expressed as: ； in, is the absolute risk; is the maximum value in the risk feature vector; The risk frequency is expressed as: ； in, is the risk frequency, is the absolute risk threshold.

5. The power battery energy consistency risk assessment method according to claim 4, characterized in that: The relative risk is expressed as: ； in, is the relative risk, The average value of the top 20% risk feature values ​​after risk feature ranking; It is the mean of the risk feature values ​​of the last 60% after the risk features are sorted.

6. The power battery energy consistency risk assessment method according to claim 5, characterized in that: In step 5, the warning conditions include: (1) ； (2) ； If (1) or (2) is satisfied, the vehicle is judged as a high-risk vehicle, where Q is the absolute risk, which indicates the absolute size of the local risk; η is the relative risk, which indicates the degree of the overall risk increase; R is the risk frequency, which indicates the number of times high risk occurs; =0.005 means that when the absolute risk quantification value of a vehicle exceeds 0.005 at any time, the vehicle is judged to be a high-risk vehicle; =5, =0.0025, =3 means that when a vehicle meets all three conditions at the same time, it is judged as a high-risk vehicle.

7. A power battery energy consistency risk assessment system, comprising a receiving module and a processing module. The receiving module acquires the operation history data uploaded by the vehicle and sends it to the processing module. The processing module receives the operation history data from the receiving module and stores it. It is characterized by: The processing module stores computer executable instructions, which, when executed, follow the steps of the power battery energy consistency risk assessment method according to any one of claims 1 to 6 and assess vehicle risk based on historical operating data.

Images