Battery capacity correction method in serious data loss scene

Abstract

The invention discloses a battery capacity correction method in a serious data loss scene, and belongs to the technical field of battery application. The method comprises the following steps: firstly, marking a position where the data loss time length is less than a, filling by adopting a proximity method, obtaining an SOC denominator theoretical value based on an SOC denominator repair method after filling, and using the SOC denominator theoretical value for marking repair of a data loss fragment capacity value where the data loss time length is greater than a. According to the method, the mode of combining the proximity method and the SOC denominator correction method is utilized, the advantages of each method are fully utilized, accurate correction of the battery capacity under the serious data loss scene of large-segment data loss is finally achieved, and the method is simple, easy to implement and high in practicability. It can be guaranteed that a vehicle with serious data loss can still achieve accurate battery health and safety diagnosis, and potential safety hazards caused by the fact that the vehicle is in a non-diagnosis state are avoided.

Description

Technical Field

[0001] The present invention relates to the field of battery application technology, and more specifically to a battery capacity correction method in a severe data loss scenario. Background Art

[0002] Cloud diagnosis of new energy vehicle batteries is one of the important means to ensure vehicle safety. The accuracy of cloud-based battery diagnosis is inseparable from its reliance on data quality. When the data quality is poor, it is easy to cause unstable diagnostic results or increased errors. One manifestation of poor data quality is data loss, that is, the data is not saved at a fixed frequency, and the relevant parameters at individual time points are missing. Among them, the most serious impact on diagnosis is the loss of large segments of data, that is, the complete loss of data for a long period of time such as more than ten minutes to half an hour. When this phenomenon occurs, it is usually because the lost data is difficult to make up or the compensation error is extremely large, so it can only be abandoned, resulting in the inability to perform health and safety diagnosis of the vehicle in the relevant time period, and unable to provide safety protection for the vehicle.

[0003] At present, the existing methods for compensating for data loss scenarios include window average method, nearest filling method, interpolation fitting method, etc., such as a method for filling missing values ​​of power data disclosed in publication number CN114519379A, and a method for repairing missing wind speed data of wind turbines disclosed in publication number CN115330076A. However, the above methods are not suitable for compensating for current data missing scenarios during battery discharge, because the current in this scenario changes rapidly, which does not conform to the characteristics of the same window average, the same nearest, and linear interpolation changes, and is prone to introduce a large number of errors. To this end, we provide a battery capacity correction method for severe data loss scenarios. Summary of the invention

[0004] The present invention provides a battery capacity correction method in a severe data loss scenario, aiming to solve the shortcomings of existing data loss scenario compensation methods, such as large correction errors and unsuitability for compensation in scenarios where current data is missing during battery discharge.

[0005] The present invention adopts the following technical solution: The battery capacity correction method in the severe data loss scenario includes the following steps: Step 1: Obtain the original ampere-hour integrated capacity array; Step 2: Mark the position where the data loss time length is less than a, and mark the position where the data loss time length is greater than a; Step 3: For the position where the length of the lost data time is less than a, the neighbor method is used to repair. The specific repair method is: intercept the data of the same length of time before or after the lost data position to fill the data of the lost data time period, and use the filled data to calculate the lost capacity data; Step 4: Use the capacity data corrected by the proximity method as input to carry out further capacity correction using the SOC denominator correction method. The specific method of the SOC denominator correction method is: Calculate the actual SOC denominator VD 0 ', carry out clustering algorithm to obtain the theoretical denominator VD 1 , the theoretical denominator VD 1 With the actual denominator VD 0 The difference is the lost capacity value, which is compensated at the corresponding capacity value; Step 5: Apply the SOC denominator repair method to the position where the length of the marked data loss time is greater than a, so as to achieve capacity correction for the position where the large segment of data is lost.

[0006] Preferably, the capacity is obtained by the following calculation formula in step 1: ; Where I is the current, Δt is the sampling time interval, i is the sampling number, Q original It is the capacity obtained by integrating ampere-hours.

[0007] In the above step 1, the initial value of the capacity is set to the rated capacity, and may also be set to 0 or other fixed values.

[0008] The data loss time length threshold a in the above step 2 is 2-5 minutes; it can also be adjusted according to the current characteristics of the actual scene. The adjustment method is: use the method of artificially creating data loss to analyze the compensation accuracy of the proximity method compensation data under different data loss lengths, and use the data loss time length when the proximity method error reaches an unsatisfactory level as the threshold a.

[0009] The clustering algorithm in step 4 above is specifically for the actual denominator VD of SOC 0 'Implement 3-class clustering, take the group with the largest average value in the clustering result as the target class array, and find the maximum value of the probability density distribution of the target class data, or find the average value. The clustering algorithm can use various clustering algorithms such as Kmean or Dbscan.

[0010] The correction method of the present invention can be applied to application fields such as new energy vehicle battery systems or energy storage battery systems.

[0011] It can be seen from the above description of the present invention that, compared with the prior art, the present invention has the following advantages: The present invention takes into account the problem that the denominator value obtained by direct clustering of the initial denominator when large segments of data are lost is usually too small, resulting in a large capacity correction error. The advantages of each method are fully utilized by combining the proximity method with the SOC denominator correction method, and ultimately achieves accurate correction of the battery capacity even in a severe data loss scenario where large segments of data are lost. The method is simple and easy to implement, and can ensure that vehicles with severe data loss can still achieve accurate battery health and safety diagnosis, avoiding the safety hazards caused by being in a no-diagnosis state. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0013] Refer to the following Figure 1 The specific implementation of the present invention is described. In order to fully understand the present invention, many details are described below, but for those skilled in the art, the present invention can be implemented without these details. For well-known components, methods and processes, they are not described in detail below.

[0014] Reference Figure 1 This embodiment provides a method for correcting battery capacity in a severe data loss scenario, including the following steps: Step 1: Get the original ampere-hour integrated capacity array and use the following calculation formula to get the capacity: ; Where I is the current, Δt is the sampling time interval, i is the sampling number, Q original It is the capacity obtained by integrating ampere-hours.

[0015] Step 2: Mark the positions where the data loss time length is less than a, and mark the positions where the data loss time length is greater than a.

[0016] In this step, the data loss time length threshold a is usually set to 2-5 minutes; it can also be adjusted according to the current characteristics of the actual scene. The adjustment method is: use the method of artificially creating data loss to analyze the compensation accuracy of the proximity method compensation data under different data loss lengths, and use the data loss time length when the proximity method error reaches an undesirable level as the threshold a.

[0017] Step 3: For the position where the length of the lost data time is less than a, the proximity method is used to repair. The specific repair method is: intercept the data of the same time length before or after the lost data position to fill the data of the lost data time period, and use the filled data to calculate the lost capacity data.

[0018] Step 4: Use the capacity data corrected by the proximity method as input to carry out further capacity correction using the SOC denominator correction method. The specific method of the SOC denominator correction method is: Calculate the actual SOC denominator VD 0 ', carry out clustering algorithm to obtain the theoretical denominator VD 1 , the theoretical denominator VD 1 With the actual denominator VD 0 The difference is the lost capacity value, which is compensated at the corresponding capacity value.

[0019] The clustering algorithm in step 4 above is specifically for the actual denominator VD of SOC 0'Implement 3-class clustering, take the group with the largest average value in the clustering result as the target class array, and find the maximum value of the probability density distribution of the target class data, or find the average value. The clustering algorithm can use various clustering algorithms such as Kmean or Dbscan.

[0020] Step 5: Apply the SOC denominator repair method to the position where the length of the marked data loss time is greater than a, so as to achieve capacity correction for the position where large segments of data are lost.

[0021] The battery capacity correction method in a severe data loss scenario of the present invention can be applied to application fields such as new energy vehicle battery systems or energy storage battery systems.

[0022] The above is only a specific implementation of the present invention, but the design concept of the present invention is not limited thereto. Any non-substantial changes to the present invention using this concept shall be deemed as an infringement of the protection scope of the present invention.

Claims

1. A battery capacity correction method in a severe data loss scenario, characterized in that: The following steps are involved: Step 1: Obtain the original ampere-hour integrated capacity array; Step 2: Mark the position where the data loss time length is less than a, and mark the position where the data loss time length is greater than a; Step 3: For the position where the length of the lost data time is less than a, the neighbor method is used to repair. The specific repair method is: intercept the data of the same length of time before or after the lost data position to fill the data of the lost data time period, and use the filled data to calculate the lost capacity data; Step 4: Use the capacity data corrected by the proximity method as input to carry out further capacity correction using the SOC denominator correction method. The specific method of the SOC denominator correction method is: Calculate the actual SOC denominator VD 0 ', carry out clustering algorithm to obtain the theoretical denominator VD 1 , the theoretical denominator VD 1 With the actual denominator VD 0 The difference is the lost capacity value, which is compensated at the corresponding capacity value; Step 5: Apply the SOC denominator repair method to the position where the length of the marked data loss time is greater than a, so as to achieve capacity correction for the position where the large segment of data is lost.

2. The battery capacity correction method in a severe data loss scenario as claimed in claim 1, characterized in that: The step 1 uses the following calculation formula to obtain the capacity: ; Where I is the current, Δt is the sampling time interval, i is the sampling number, Q original It is the capacity obtained by integrating ampere-hours.

3. The method for correcting battery capacity in a severe data loss scenario according to claim 1, characterized in that: The initial value of the capacity in step 1 is set to the rated capacity or 0.

4. The method for correcting battery capacity in a severe data loss scenario according to claim 1, characterized in that: The data loss time length threshold a in step 2 is 2-5 minutes.

5. The method for correcting battery capacity in a severe data loss scenario according to claim 1, wherein: The data loss time length threshold a in step 2 can be adjusted according to the current characteristics of the actual scene. The adjustment method is: use the method of artificially creating data loss to analyze the compensation accuracy of the proximity method compensation data under different data loss lengths, and use the data loss time length when the proximity method error reaches an undesirable level as the threshold a.

6. The method for correcting battery capacity in a severe data loss scenario according to claim 1, characterized in that: The clustering algorithm in step 4 is specifically for the actual denominator VD of SOC. 0 'Implement 3-category clustering, take the group with the largest average value in the clustering results as the target class array, and find the value with the largest probability density distribution or the average value for the target class data.

7. The method for correcting battery capacity in a severe data loss scenario according to claim 1, characterized in that: The clustering algorithm adopts Kmean or Dbscan.

8. The method for correcting battery capacity in a severe data loss scenario as claimed in claim 1, characterized in that: The correction method is applied to new energy vehicle battery systems or energy storage battery systems.

Images