Improved method for calculating maximum available capacity of power battery pack

Abstract

The invention provides an improved method for calculating the maximum available capacity of a power battery pack, and the method comprises the construction of a vehicle state segmentation rule, the extraction and screening of charge and discharge fragments, and the construction of a global quartile method-local isolated forest abnormal value comprehensive detection algorithm. Furthermore, the ampere-hour integral method is improved in combination with the charging and discharging process and the charging and discharging efficiency of different regions and seasons, after resampling, data fluctuation is effectively smoothed, data information is fully utilized, good accuracy and applicability are achieved, the current maximum available capacity of the battery pack can be accurately monitored, and the maximum available capacity of the battery pack can be accurately monitored. And a solid foundation is laid for accurate estimation of the power battery health state.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing of power batteries for new energy vehicles, and in particular to an improved method for calculating the maximum available capacity of a power battery pack. Background Art

[0002] In the context of the current changes in the global energy and transportation fields, the rapid development of big data technology, new energy vehicles and power batteries are jointly driving the industry towards a greener and more efficient future. As the core component of pure electric vehicles (BEV), the performance of power batteries significantly affects the vehicle's cruising range, safety performance and overall efficiency. At present, most of the research on the estimation of the state of health (SOH) of lithium-ion batteries (LIBs) relies on data collected in the laboratory. In the laboratory environment, the capacity calibration test is usually performed every 50 or 100 cycles, which is usually called a periodic static capacity test. The corresponding real-time dynamic capacity test is a capacity acquisition method under the most ideal conditions, which can better reflect the battery's SOH. However, in the actual application of vehicles, vehicles cannot perform real-time dynamic capacity tests, nor will they perform periodic static capacity tests. Therefore, how to accurately calculate the current maximum available capacity of the battery pack through historical data is a challenge. Summary of the invention

[0003] The technical problem to be solved by the present invention is: in order to overcome the deficiencies in the prior art, the present invention provides an improved method for calculating the maximum available capacity of a power battery pack.

[0004] The technical solution adopted by the present invention to solve the technical problem is: an improved method for calculating the maximum available capacity of a power battery pack, comprising the following steps:

[0005] S1: Collect the first data set X

[0006] The vehicle terminal collects the data of the whole vehicle and each component in real time through the battery management system BMS at a certain sampling frequency to obtain the first data set X.

[0007] S2: Clean the first data set X

[0008] Based on the first data set X, redundant samples are removed to obtain the second data set X′; based on the second data set X′, the time series is corrected to obtain the third data set X″, ensuring data quality and improving the credibility of the analysis results, which specifically includes the following steps:

[0009] S2.1: Eliminate redundant samples

[0010] Due to repeated records in the data collection process or signal fluctuations in the data transmission process, the data set contains repeated or similar sample records. Therefore, it is necessary to first eliminate the redundant data and use filtering to delete the repeated or similar sample records in the first data set X to obtain the second data set X′.

[0011] S2.2: Corrected time series

[0012] In the process of real-time data transmission or recording, time disorder may occur due to data transmission delay, time synchronization of different devices and other issues. In order to solve the time disorder problem in the data set, the format of the data time in the second data set X′ is unified by the time formatting function, arranged in chronological order, and the timestamp field of each data is obtained. The timestamp field is sorted by the sorting function, and the time series of the data in the second data set X′ is corrected to obtain the third data set X″.

[0013] S3: Extracting charging and discharging fragments

[0014] Based on the time characteristics, a charging and discharging segmentation process is constructed, and the charging segment (i.e., charging cycle) and discharging segment (i.e., discharging cycle) of the power battery are accurately and reliably identified from the third data set X″ data. Specifically, the following steps are included:

[0015] S3.1: According to the vehicle state and charging state, a charging and discharging sample data set is selected from the third data set X″, and data in non-charging and non-driving states are removed to obtain a fourth data set, which is recorded as D=[X c ,X d ];

[0016] S3.2: Determine the fourth data set D = [X c ,X d ], whether the time interval between two adjacent rows of data exceeds the set charge-discharge cycle threshold θ. If so, it is regarded as a new charge-discharge cycle, and the number of charge-discharge cycles is recorded. According to the above process, the fourth data set D = [X c ,X d ], denoted as M and N respectively, thus obtaining M charging segments X c (i), i∈[1,M] and N discharge segments X d (i), i∈[1,N], denoted as the fifth data set D′=[X c (i)|i∈[1,M],X d (i)|i∈[1,N]];

[0017] S3.3: Eliminate the abnormal segments with interference and clutter in the charging and discharging process in the fifth data set D′ to obtain the sixth data set, which is recorded as D″=[X′ c (i)|i∈[1,M′],X′ d (i)|i∈[1,N′]], M′≤M, N′≤N;

[0018] S3.4: Correction of outliers and missing values

[0019] For extreme value fields (cell maximum voltage, maximum temperature, etc.), it is assumed that there are no outliers, which reflects the real situation of the system, so no processing is done. For other key fields (cell voltage, cell temperature, etc.), a comprehensive outlier detection method of global quartile method-local isolation forest method is constructed, and the outlier comprehensive detection method is used to identify the outliers in the sixth data set D″, and the outliers are set as missing values. Then, the missing values ​​are interpolated and filled based on the interpolation algorithm to obtain the seventh data set, which is recorded as D″′=[X″ c (i)|i∈[1,m],X″ d (i)|i∈[1,n]], m≥M′, n≥N′, where X″ c (i)|i∈[1,m] is the final extracted charging fragment, X″ d (i)|i∈[1,n] is the final extracted discharge segment.

[0020] Among them, the comprehensive outlier detection method of the global quartile method-local isolation forest method is as follows:

[0021] First, the key fields are divided into two categories: global fields and local fields. The global fields refer to the attribute fields that describe the entire data set or vehicle system, and do not change with changes in road conditions, environment, and driving behavior; the local fields refer to the fields that are independently related to each charging and discharging segment in the data set, which contain specific battery status information of the charging and discharging segments and change with changes in road conditions, environment, and driving behavior.

[0022] Secondly, the quartile method is used to perform global detection on key fields. For global and local fields with outliers, the isolation forest method is used to detect the global and local fields of each charging and discharging segment respectively, and the outliers are set as missing values ​​to facilitate subsequent interpolation filling.

[0023] S4: Calculate the maximum available capacity

[0024] The improved ampere-hour integration method is used for the charging segment and the discharging segment respectively, thereby obtaining the maximum available capacity of the charging process and the discharging process; specifically, the following steps are included:

[0025] S4.1: Determine the maximum available capacity formula for the improved charge and discharge process

[0026] The maximum available capacity is calculated as:

[0027]

[0028] Among them, C max is the maximum available capacity corresponding to a certain charge and discharge segment, ΔQ is the change in charge and discharge capacity, and ΔSOC is the change in charge and discharge capacity;

[0029] Based on the ampere-hour integration method, the above C max The inverse formula is:

[0030]

[0031] After further discretization, the improved maximum available capacity formula of the charge and discharge process is obtained:

[0032]

[0033] In the formula, SOC(t0) is the SOC value corresponding to the time t0 of a certain charge and discharge segment; SOC(t) is the SOC value corresponding to the time t when the charge and discharge of this segment ends; I(τ) is the battery current; η is the charge and discharge efficiency, which is usually between 0.98 and 1; k is the total number of samples of this charge and discharge segment, and △t i is the sampling interval between two adjacent rows of the charging segment, I i is the current of the charge and discharge segment.

[0034] S4.2: C based on factors that affect the maximum available capacity max Make adjustments

[0035] S4.2.1: Determine the charge and discharge efficiency η according to different regions and seasons, then:

[0036] The discharge efficiency in spring, summer and autumn in the south is η d,S ;

[0037] The discharge efficiency in spring, summer and autumn in the north is η d,N ;

[0038] The charging efficiency in spring, summer and autumn in the south is η c,s =η d,S -b S ;

[0039] The charging efficiency in the north in spring, summer and autumn is η c,N =η d,N -b N ;

[0040] The charge and discharge efficiency in winter is η winter =η state , region -b;

[0041] Among them, η is the charge and discharge efficiency, c and d represent the charge state and discharge state respectively, S and N represent the south and north respectively, b S 、b N , b are the bias constants corresponding to different regions and seasons, η syate，region It is the charging and discharging efficiency in spring, summer and autumn of a certain area.

[0042] S4.2.2: Local Isolation Forest Outlier Detection and Correction for Total Current and SOC

[0043] The SOC and total current of each charging segment and discharging segment obtained in step S3 are respectively detected and corrected for outliers, and the formula is as follows:

[0044]

[0045] In the formula, I(t) and SOC(t) are the current and SOC values ​​corresponding to time t, respectively. φ(.) are the local isolation forest outlier detection function and the outlier correction function (i.e., interpolation function), respectively. I′(t) and SOC′(t) are the corrected current and SOC values ​​at time t, respectively.

[0046] S4.2.3: Screening for effectiveness of charge and discharge depth

[0047] Based on the battery charge depth DOC, the charging segments corrected by S4.2.2 are screened, and the DOC value is less than the battery charge state threshold ε c K c fragments;

[0048] Similarly, based on the battery's discharge depth DOD, the discharge segments corrected in S4.2.2 are screened, and the discharge segments with DOD values ​​less than the battery's discharge depth threshold ε are eliminated. d K d fragments.

[0049] S4.2.4: Considering the continuity of charge and discharge current, during data collection, the sampling frequency is generally 10s, 20s or other. The actual value of the current I′(t) in the sampling interval is unknown. Therefore, consider dividing the adjacent sampling intervals (I′ j ,I′ j+1 ) average current As the current within a sampling period (10s / 20s), rather than a sampling period △t i The first and last values ​​of .

[0050] S4.2.5: The factors affecting the maximum available capacity determined in steps S4.2.1-S4.2.4 shall be used to calculate the C max To make adjustments:

[0051] The extracted charging fragments are processed to obtain the maximum available capacity of the charging process:

[0052]

[0053] Similarly, the extracted discharge fragments are processed to obtain the maximum available capacity of the discharge process:

[0054]

[0055] Among them, η j The value is η d,S , η d,N , η c，s , η c，N , η winter ;

[0056] Then the maximum available capacity data set during the charge and discharge process is recorded as the eighth data set C = [C cmax (t), C dmax (t)].

[0057] S4.3: Resampling

[0058] With a certain resampling frequency f, the eighth data set C of a single charge and discharge at different time resolutions is converted into a time series data set with a fixed time frequency, which is recorded as the ninth data set [C max (t)]:

[0059] C max (t) = R f (C cmax (t),C dmax (t))

[0060] In the formula, C max (t) is the maximum available charge and discharge capacity after resampling, R f Represents the resampling function R(.) with resampling frequency f.

[0061] S4.4: Fill in empty values

[0062] Based on the interpolation algorithm, the ninth data set [C max (t)] to fill in the empty values ​​in the image to obtain the nonlinear representation of the current maximum available capacity, i.e., the tenth data set C′ max (t):

[0063] C′ max (t) = φ(C max (t))

[0064] Where φ(.) is the interpolation function.

[0065] S5: The maximum available capacity calculated by the nonlinear representation of the current maximum available capacity is used to accurately and effectively control the battery.

[0066] The calculation accuracy of the maximum available capacity of a battery in the prior art is relatively limited. However, by adopting the method of the present invention, a more accurate maximum available capacity can be obtained, so that the battery can be used accurately and better controlled, thereby being able to more accurately monitor the current maximum available capacity of the battery pack, laying a solid foundation for accurate estimation of the health status of the power battery.

[0067] The local isolation forest outlier detection function in the present invention The interpolation function φ(.), i.e., the outlier correction function, the resampling function R(.), the quartile method and the isolation forest method are all known functions and methods in the prior art, and therefore, their specific contents are not described in detail in the present invention.

[0068] The beneficial effects of the present invention are:

[0069] (1) Combined with the charging and discharging process and the charging and discharging efficiency in different regions and seasons, the maximum available capacity of the power battery pack is calculated, and the impact of different stages on the battery is considered, with good accuracy and applicability;

[0070] (2) After the resampling technology, the influence of outliers is significantly reduced, the data fluctuations are effectively smoothed, the data information is fully utilized, and it has good accuracy and generalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0072] Figure 1 It is a flow chart of the improved method for calculating the maximum available capacity of a power battery pack according to the present invention.

[0073] Figure 2 It is a flow chart of the charging and discharging segment segmentation rules.

[0074] Figure 3 This is a flowchart of outlier processing.

[0075] Figure 4 It is the comparison of SOH obtained by charging, discharging, splicing and resampling. DETAILED DESCRIPTION

[0076] The present invention is now described in detail in conjunction with the accompanying drawings. The accompanying drawings are simplified schematic diagrams, which only illustrate the basic process of the present invention in an illustrative manner, and therefore only show the components related to the present invention. In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0077] like Figure 1 As shown, an improved method for calculating the maximum available capacity of a power battery pack of the present invention comprises the following steps:

[0078] S1: Collect the first data set X

[0079] The vehicle terminal collects the data of the whole vehicle and each component in real time through the battery management system BMS at a certain sampling frequency to obtain the first data set X.

[0080] S2: Clean the first data set X

[0081] Due to the limitations of sensor acquisition accuracy and the inevitable noise interference during data acquisition and transmission, the battery data sets stored in the database usually have problems such as data redundancy, anomalies, and missing data. Therefore, it is necessary to clean the collected data X to ensure data quality and improve the credibility of the analysis results.

[0082] Based on the first data set X, redundant samples are removed to obtain the second data set X′; based on the second data set X′, the time series is corrected to obtain the third data set X″, ensuring data quality and improving the credibility of the analysis results, which specifically includes the following steps:

[0083] S2.1: Eliminate redundant samples

[0084] Due to repeated records in the data collection process or signal fluctuations in the data transmission process, the data set contains repeated or similar sample records. Therefore, it is necessary to first eliminate the redundant data and use filtering to delete the repeated or similar sample records in the first data set X to obtain the second data set X′.

[0085] S2.2: Corrected time series

[0086] In the process of real-time data transmission or recording, time disorder may occur due to data transmission delay, time synchronization of different devices and other issues. In order to solve the time disorder problem in the data set, the format of the data time in the second data set X′ is unified by the time formatting function, arranged in chronological order, and the timestamp field of each data is obtained. The timestamp field is sorted by the sorting function, and the time series of the data in the second data set X′ is corrected to obtain the third data set X″.

[0087] S3: Extracting charging and discharging fragments

[0088] like Figure 2As shown, based on the time characteristics, a charging and discharging segment segmentation process is constructed, and the charging segment (i.e., charging cycle) and discharging segment (i.e., discharging cycle) of the power battery are accurately and reliably identified from the third data set X″ data, which specifically includes the following steps:

[0089] S3.1: According to the vehicle state and charging state, a charging and discharging sample data set is selected from the third data set X″, and data in non-charging and non-driving states are removed to obtain a fourth data set, which is recorded as D=[X c ,X d ];

[0090] S3.2: Determine the fourth data set D = [X c ,X d ], whether the time interval between two adjacent rows of data exceeds the set charge-discharge cycle threshold θ. If so, it is regarded as a new charge-discharge cycle, and the number of charge-discharge cycles is recorded. According to the above process, the fourth data set D = [X c ,X d ], denoted as M and N respectively, thus obtaining M charging segments X c (i), i∈[1,M] and N discharge segments X d (i), i∈[1,N], denoted as the fifth data set D′=[X c (i)|i∈[1,M],X d (i)|i∈[1,N]];

[0091] S3.3: Eliminate the abnormal segments with interference and clutter in the charging and discharging process in the fifth data set D′ to obtain the sixth data set, which is recorded as D″=[X′ c (i)|i∈[1,M′],X′ d (i)|i∈[1,N′]], M′≤M, N′≤N;

[0092] S3.4: Correction of outliers and missing values

[0093] For extreme value fields (cell maximum voltage, maximum temperature, etc.), it is assumed that there are no outliers, which reflects the real situation of the system, so no processing is done. For other key fields (cell voltage, cell temperature, etc.), a comprehensive outlier detection method of global quartile method-local isolation forest method is constructed, and the outlier comprehensive detection method is used to identify the outliers in the sixth data set D″, and the outliers are set as missing values. Then, the missing values ​​are interpolated and filled based on the interpolation algorithm to obtain the seventh data set, which is recorded as D″′=[X″ c (i)|i∈[1,m],X″ d (i)|i∈[1,n]],m≥M ′, n≥N′, where X″ c (i)|i∈[1,m] is the final extracted charging fragment, X″ d (i)|i∈[1,n] is the final extracted discharge segment.

[0094] Among them, the comprehensive outlier detection method of the global quartile method-local isolation forest method is as follows:

[0095] First, the key fields are divided into two categories: global fields and local fields. The global fields refer to the attribute fields that describe the entire data set or vehicle system, and do not change with changes in road conditions, environment, and driving behavior; the local fields refer to the fields that are independently related to each charging and discharging segment in the data set, which contain specific battery status information of the charging and discharging segments and change with changes in road conditions, environment, and driving behavior.

[0096] Secondly, the quartile method is used to perform global detection on key fields. For global and local fields with outliers, the isolation forest method is used to detect the global and local fields of each charging and discharging segment respectively, and the outliers are set as missing values ​​to facilitate subsequent interpolation filling.

[0097] S4: Calculate the maximum available capacity

[0098] The maximum available capacity of a power battery refers to the amount of electricity that the battery can store and release under certain conditions. This capacity will gradually decrease with the use and aging of the battery. Therefore, accurate estimation of the maximum available capacity has important practical significance for battery management, usage optimization, safety improvement, and economic benefit improvement.

[0099] The improved ampere-hour integration method is used for the charging segment and the discharging segment respectively, thereby obtaining the maximum available capacity of the charging process and the discharging process; specifically, the following steps are included:

[0100] S4.1: Determine the maximum available capacity formula for the improved charge and discharge process

[0101] The maximum available capacity is calculated as:

[0102]

[0103] Among them, C max is the maximum available capacity corresponding to a certain charge and discharge segment, ΔQ is the change in charge and discharge capacity, and ΔSOC is the change in charge and discharge capacity;

[0104] Based on the ampere-hour integration method, the above C max The inverse formula is:

[0105]

[0106] After further discretization, the improved maximum available capacity formula of the charge and discharge process is obtained:

[0107]

[0108] In the formula, SOC(t0) is the SOC value corresponding to the time t0 of a certain charge and discharge segment; SOC(t) is the SOC value corresponding to the time t when the charge and discharge of this segment ends; I(τ) is the battery current; η is the charge and discharge efficiency, which is usually between 0.98 and 1; k is the total number of samples of this charge and discharge segment, and △t i is the sampling interval between two adjacent rows of the charging segment, I i is the current of the charge and discharge segment.

[0109] S4.2: C based on factors that affect the maximum available capacity max Make adjustments

[0110] Starting from the main factors affecting the maximum available capacity, the following factors are mainly considered:

[0111] S4.2.1: Consider the charge and discharge efficiency in different regions and seasons. The specific charge and discharge efficiency is shown in Table 1. Under low temperature conditions in winter, the chemical reaction rate of LIBs will drop significantly, so the charge and discharge efficiency in winter is relatively low.

[0112] Table 1 Charge and discharge efficiency η table

[0113]

[0114] In Table 1, the charge and discharge efficiency η is determined according to different regions and seasons, then:

[0115] The discharge efficiency in spring, summer and autumn in the south is η d,S ;

[0116] The discharge efficiency in spring, summer and autumn in the north is η d,N ;

[0117] The charging efficiency in the south in spring, summer and autumn is η c,s =η d,S -b S ;

[0118] The charging efficiency in the north in spring, summer and autumn is η c,N =η d,N -b N ;

[0119] The charge and discharge efficiency in winter is η winter =η state , region -b;

[0120] Among them, η is the charge and discharge efficiency, c and d represent the charge state and discharge state respectively, S and N represent the south and north respectively, b S 、b N , b are the bias constants corresponding to different regions and seasons, η state , region is the charging and discharging efficiency in spring, summer and autumn of a certain region.

[0121] S4.2.2: Local Isolation Forest Outlier Detection and Correction for Total Current and SOC

[0122] like Figure 3 As shown, outlier detection and correction are performed on the SOC and total current of each charging segment and discharging segment obtained in step S3, respectively, and the formula is as follows:

[0123]

[0124] In the formula, I(t) and SOC(t) are the current and SOC values ​​corresponding to time t, respectively. φ(.) are the local isolation forest outlier detection function and the outlier correction function (i.e., interpolation function), respectively. I′(t) and SOC′(t) are the corrected current and SOC values ​​at time t, respectively.

[0125] S4.2.3: Screening for effectiveness of charge and discharge depth

[0126] Based on the battery charge depth DOC, the charging segments corrected by S4.2.2 are screened, and the DOC value is less than the battery charge state threshold ε c K c fragments;

[0127] Similarly, based on the battery's discharge depth DOD, the discharge segments corrected in S4.2.2 are screened, and the discharge segments with DOD values ​​less than the battery's discharge depth threshold ε are eliminated. d K d fragments.

[0128] S4.2.4: Considering the continuity of charge and discharge current, during data collection, the sampling frequency is generally 10s, 20s or other. The actual value of the current I′(t) in the sampling interval is unknown. Therefore, consider dividing the adjacent sampling intervals (I′ j ,I′ j+1 ) average current As the current within a sampling period (10s / 20s), rather than a sampling period △t i The first and last values ​​of .

[0129] S4.2.5: The factors affecting the maximum available capacity determined in steps S4.2.1-S4.2.4 shall be used to calculate the C max To make adjustments:

[0130] The extracted charging fragments are processed to obtain the maximum available capacity of the charging process:

[0131]

[0132] Similarly, the extracted discharge fragments are processed to obtain the maximum available capacity of the discharge process:

[0133]

[0134] Among them, η j The value is η d，s , η d,N , η c，s , η c，N , η winter ;

[0135] Then the maximum available capacity data set during the charge and discharge process is recorded as the eighth data set C = [C cmax (t), C dmax ((t)].

[0136] The charging process reflects the user's charging habits, and the various parameter indicators in this process are relatively stable; while the discharging process indirectly reflects the driver's driving style and traffic conditions, which is often accompanied by extreme instantaneous power, and the fluctuation range of various parameters is large. All these factors will have a comprehensive impact on the battery pack life through different degradation processes. Therefore, considering the impact of the charging and discharging process on the maximum available capacity, the eighth data set is also resampled and filled with null values. The specific steps are as follows.

[0137] S4.3: Resampling

[0138] Based on the resampling technology, the eighth data set C of single charge and discharge at different time resolutions is converted into a time series data set with a fixed time frequency at a certain resampling frequency f, which is recorded as the ninth data set [C max (t)]:

[0139] C max (t) = R f (C cmax (t),C dmax (t))

[0140] In the formula, C max (t) is the maximum available charge and discharge capacity after resampling, R f Represents the resampling function R(.) with resampling frequency f.

[0141] S4.4: Fill in empty values

[0142] Although the resampling process unified the data format, it also introduced the null value problem. Therefore, the resampled ninth data set [C max (t)] to fill in the empty values ​​in the image to obtain the nonlinear representation of the current maximum available capacity, i.e., the tenth data set C′ max (t):

[0143] C′ max (t) = φ(C max (t))

[0144] Where φ(.) is the interpolation function.

[0145] S5: The maximum available capacity calculated by the nonlinear representation of the current maximum available capacity is used to accurately and effectively control the battery.

[0146] like Figure 4 As shown, the left side (a) shows the SOH distribution of the charging and discharging segments of the original data collected from a vehicle with a mileage of 300,000-500,000 km, which is a common method in the prior art. Despite the data cleaning and DOC / DOD filtering techniques, there are still some outliers in the charging and discharging process, and there are more data with SOH greater than 100% and SOH less than 80%. Therefore, the calculated maximum available capacity of the battery pack is less accurate. After splicing the charging and discharging segment data in time series, these outliers are still inevitable, as shown in the black scattered points in the right side figure (b). The orange scattered points in the right figure (b) are the SOH distribution of the charging segment and the discharging segment after the original data with a mileage of 300,000-500,000 km collected is processed by the method of the present invention. It can be seen from the figure that compared with the original data, the SOH distribution of the data is concentrated between 100% and 80%, which makes full use of the charging and discharging segment data, effectively eliminates abnormal data, and makes the data more accurate, so that the current maximum available capacity of the battery pack can be monitored more accurately.

[0147] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the scope of the present invention through the above description. The technical scope of this invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An improved method for calculating the maximum available capacity of a power battery pack, characterized in that: The following steps are involved: S1: Collect the first data set X The vehicle terminal collects data of the vehicle and each component in real time through the battery management system BMS at a certain sampling frequency to obtain the first data set X; S2: Clean the first data set X Based on the first data set X, redundant samples are removed to obtain the second data set X′; based on the second data set X′, the time series is corrected to obtain the third data set X″; S3: Extracting charging and discharging fragments Based on the time characteristics, a charging and discharging segmentation process is constructed, and the charging segment and the discharging segment of the power battery are identified from the data of the third data set X″ by using the process; S4: Calculate the maximum available capacity The improved ampere-hour integration method is used for the charging segment and the discharging segment respectively, thereby obtaining the maximum available capacity of the charging process and the discharging process; S5: Use the calculated maximum available capacity to accurately and effectively control the battery.

2. The improved method for calculating the maximum available capacity of a power battery pack according to claim 1, characterized in that: Step S2 cleans the first data set X, specifically including the following steps: S2.1: Eliminate redundant samples By filtering, duplicate or similar sample records in the first data set X are deleted, thereby obtaining a second data set X′; S2.2: Corrected time series The time formatting function is used to unify the format of the data time in the second data set X′, arrange them in chronological order, obtain the timestamp field of each data, use the sorting function to sort the timestamp field, correct the time series of the data in the second data set X′, and obtain the third data set X″.

3. The improved method for calculating the maximum available capacity of a power battery pack as claimed in claim 2, characterized in that: Step S3 extracting the charge and discharge segments specifically includes the following steps: S3.1: According to the vehicle state and charging state, a charging and discharging sample data set is selected from the third data set X″, and data in non-charging and non-driving states are removed to obtain a fourth data set, which is recorded as D=[X c , X d ]; S3.2: Determine the fourth data set D = [X c , X d ], whether the time interval between two adjacent rows of data exceeds the set charge-discharge cycle threshold θ. If so, it is regarded as a new charge-discharge cycle, and the number of charge-discharge cycles is recorded. According to the above process, the fourth data set D = [X c , X d ], denoted as M and N respectively, thus obtaining M charging segments X c (i), i∈[1,M] and N discharge segments X d (i), i∈[1, N], denoted as the fifth data set D′=[X c (i)|i∈[1,M],X d (i)|i∈[1,N]]; S3.3: Eliminate the abnormal segments with interference and clutter in the charging and discharging process in the fifth data set D′ to obtain the sixth data set, which is recorded as D″=[X′ c (i)|i∈[1,M′],X′ d (i)|i∈[1, N′]], M′≤M, N′≤N; S3.4: Correction of outliers and missing values A comprehensive outlier detection method based on the global quartile method and the local isolation forest method is constructed. The outlier comprehensive detection method is used to identify the outliers in the sixth data set D", and the outliers are set as missing values. Then, the missing values ​​are interpolated and filled based on the interpolation algorithm to obtain the seventh data set, which is recorded as D"' = [X" c (i)|i∈[1,m],X″ d (i)|i∈[1,n]],m≥M′,n≥N′,where,X″ c (i)|i∈[1,m] is the charging segment finally extracted, and X″d(i)|i∈[1,n] is the discharging segment finally extracted.

4. The improved method for calculating the maximum available capacity of a power battery pack as claimed in claim 3, characterized in that: The comprehensive outlier detection method of the global quartile method-local isolation forest method in step S3.4 is specifically: First, the key fields are divided into two categories: global fields and local fields. The global fields refer to the attribute fields that describe the entire data set or vehicle system, which do not change with changes in road conditions, environment, and driving behavior; the local fields refer to the fields that are independently related to each charging and discharging segment in the data set, which contain specific battery status information of the charging and discharging segments and change with changes in road conditions, environment, and driving behavior; Secondly, the quartile method is used to perform global detection on key fields. For global and local fields with outliers, the isolation forest method is used to detect the global and local fields of each charging and discharging segment respectively, and the outliers are set as missing values ​​to facilitate subsequent interpolation filling.

5. The improved method for calculating the maximum available capacity of a power battery pack according to claim 1, characterized in that: Step S4 calculates the maximum available capacity and specifically includes the following steps: S4.1: Determine the maximum available capacity formula for the improved charge and discharge process The maximum available capacity is calculated as: Among them, C max is the maximum available capacity corresponding to a certain charge and discharge segment, ΔQ is the change in charge and discharge capacity, and ΔSOC is the change in charge and discharge capacity; Based on the ampere-hour integration method, the above C max The inverse formula is: After further discretization, the improved maximum available capacity formula of the charge and discharge process is obtained: In the formula, SOC(t0) is the SOC value corresponding to the time t0 of a certain charge and discharge segment; SOC(t) is the SOC value corresponding to the time t when the charge and discharge of this segment ends; I(τ) is the battery current; η is the charge and discharge efficiency, k is the total number of samples of this charge and discharge segment, △t i is the sampling interval between two adjacent rows of the charging segment, I i is the current of the charge and discharge segment; S4.2: C based on factors that affect the maximum available capacity max Make adjustments S4.2.1: Determine the charge and discharge efficiency η according to different regions and seasons, then: The discharge efficiency in spring, summer and autumn in the south is η d,S ; The discharge efficiency in spring, summer and autumn in the north is η d,N ; The charging efficiency in spring, summer and autumn in the south is η c，s =η d,S -b S ; The charging efficiency in the north in spring, summer and autumn is η c，N =η d,N -b N ; The charge and discharge efficiency in winter is η winter =η state，region -b; Among them, η is the charge and discharge efficiency, c and d represent the charge state and discharge state respectively, S and N represent the south and north respectively, b S , b N , b are the bias constants corresponding to different regions and seasons, η state，region The charge and discharge efficiency in spring, summer and autumn of a certain region; S4.2.2: Local Isolation Forest Outlier Detection and Correction for Total Current and SOC The SOC and total current of each charging segment and discharging segment obtained in step S3 are respectively detected and corrected for outliers, and the formula is as follows: In the formula, I(t) and SOC(t) are the current and SOC values ​​corresponding to time t, respectively. φ(.) are the local isolation forest outlier detection function and outlier correction function, respectively. I′(t) and SOC′(t) are the corrected current and SOC values ​​at time t, respectively. S4.2.3: Screening for effectiveness of charge and discharge depth Based on the battery charge depth DOC, the charging segments corrected by S4.2.2 are screened, and the DOC value is less than the battery charge state threshold ε c K c fragments; Similarly, based on the battery's discharge depth DOD, the discharge segments corrected in S4.2.2 are screened, and the discharge segments with DOD values ​​less than the battery's discharge depth threshold ε are eliminated. d K d fragments; S4.2.4: The adjacent sampling interval (I′ j , I′ j+1 ) average current As the current within a sampling period; S4.2.5: The factors affecting the maximum available capacity determined in steps S4.2.1-S4.2.4 shall be used to calculate the C max To make adjustments: The extracted charging fragments are processed to obtain the maximum available capacity of the charging process: Similarly, the extracted discharge fragments are processed to obtain the maximum available capacity of the discharge process: Among them, η j takes the value of η d,S , η d,N , η c，s , η c,N , η winter ; Then the maximum available capacity data set during the charge and discharge process is recorded as the eighth data set C = [C cmax (t), C dmax (t)].

6. The improved method for calculating the maximum available capacity of a power battery pack as claimed in claim 5, characterized in that: The following steps are also included: S4.3: Resampling With a certain resampling frequency f, the eighth data set C of a single charge and discharge at different time resolutions is converted into a time series data set with a fixed time frequency, which is recorded as the ninth data set [C max (t)]: C max (t)=R f (C cmax (t),C dmax (t)) In the formula, C max (t) is the maximum available charge and discharge capacity after resampling, R f represents the resampling function R(.) with resampling frequency f; S4.4: Fill in empty values Based on the interpolation algorithm, the ninth data set [C max (t)] to fill in the empty values ​​in the image to obtain the nonlinear representation of the current maximum available capacity, i.e., the tenth data set C′ max (t): C′ max (t)=φ(C max (t)) Where φ(.) is the interpolation function.

Images