A charging SOC correction method and system based on dynamic time warping

Abstract

The present invention provides a charging SOC correction method and system based on dynamic time warping in the field of lithium battery online detection technology, the method comprising: step S10, obtaining a large number of historical charging orders of new energy vehicles and corresponding charging process data; step S20, classifying each of the historical charging orders into high SOC interval orders and low SOC interval orders based on SOC; step S30, performing SOC correction on each of the high SOC interval orders through the inverse ampere-hour integration method to obtain a first corrected SOC value; step S40, constructing a reference charging curve set based on the first corrected SOC value and charging process data; step S50, based on the dynamic time warping algorithm, performing SOC correction on the low SOC interval order through the reference charging curve set to obtain a second corrected SOC value; step S60, generating a SOC correction curve based on each of the first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC. The advantage of the present invention is that it greatly improves the accuracy of charging SOC correction.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery online detection, and in particular to a charging SOC correction method and system based on dynamic time regularization. Background Art

[0002] With the continuous development of lithium battery fast charging technology, it is increasingly important to correctly estimate the SOC of lithium batteries during the charging stage. However, with the continuous increase in charging rate and the complexity of charging conditions (changes in temperature and rate), the estimation of SOC has become more difficult.

[0003] When the traditional OCV-SOC curve is applied to SOC estimation, the lithium battery needs to be left to rest for a period of time after charging to obtain the open circuit voltage OCV, so that the SOC estimation based on the OCV-SOC curve is difficult to achieve during the charging stage of the lithium battery. The method based on the equivalent circuit model + Kalman filter has become the mainstream method for SOC estimation during the charging stage, but due to factors such as the inaccuracy of the initial SOC value, the changes in the total battery capacity Q, the charging rate, and the charging temperature, the real-time SOC estimation often has a certain deviation. The evaluation and prediction of the lithium battery status depends on the SOC during the charging stage, so the SOC needs to be corrected.

[0004] For the correction of SOC under charging status, traditionally, calibration points are generally selected in the high SOC area to correct the SOC. This means that before the SOC correction, the SOC value obtained during the charging process may not be correct, and new energy vehicles may not be charged to the high SOC area during actual charging. Due to the randomness of charging, each charge may only obtain a charging fragment of a part of the SOC area, which makes it impossible to completely perform SOC correction of the entire charging curve based on the SOC correction value of the BMS. If the SOC in the low SOC area cannot be corrected, but the feature extraction and state estimation / prediction of the lithium battery are directly performed, it will lead to deviations in the estimation results / prediction results.

[0005] Therefore, how to provide a charging SOC correction method and system based on dynamic time regularization to improve the accuracy of charging SOC correction has become a technical problem that needs to be solved urgently. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a charging SOC correction method and system based on dynamic time regularization to improve the accuracy of charging SOC correction.

[0007] In a first aspect, the present invention provides a charging SOC correction method based on dynamic time warping, comprising the following steps:

[0008] Step S10, obtaining a large number of historical charging orders for new energy vehicles and corresponding charging process data;

[0009] Step S20, classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC;

[0010] Step S30, performing SOC correction on each of the high SOC interval orders by an inverse ampere-hour integration method to obtain a first corrected SOC value;

[0011] Step S40, constructing a reference charging curve set based on the first corrected SOC value and charging process data;

[0012] Step S50: Based on the dynamic time warping algorithm, perform SOC correction on the low SOC interval order through the reference charging curve set to obtain a second corrected SOC value;

[0013] Step S60: Generate an SOC correction curve based on the first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC.

[0014] Furthermore, the step S10 is specifically as follows:

[0015] Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained;

[0016] The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

[0017] Furthermore, the step S20 is specifically as follows:

[0018] When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders;

[0019] When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

[0020] Furthermore, in step S30, the calculation formula of the first corrected SOC value is:

[0021]

[0022] Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment factor.

[0023] Furthermore, the step S40 is specifically as follows:

[0024] Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve;

[0025] The first charging curve is a curve showing that the charging voltage changes with the first corrected SOC value; the second charging curve is a curve showing that the charging current changes with the first corrected SOC value; and the third charging curve is a curve showing that the temperature changes with the first corrected SOC value;

[0026] The step S50 specifically includes:

[0027] Step S51: select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set based on the charging rate, charging current or charging date as the reference charging curve sequence S. benchmark ;

[0028] Step S52: Using 1% SOC as the resolution interval and 5% SOC as the search interval, obtain the correction candidate set SOC of the starting SOC of the order to be corrected. start_n And the correction candidate set SOC of the end SOC end_n ;

[0029] SOC start_n = {SOC star -5,...,SOC star -1,SOC star ,SOC star +1,...,SOC star +5,};

[0030] SOC end_n ={SOC end -5,...,SOC end -1,SOC end ,SOC end +1,...,SOC end+5,};

[0031] Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC;

[0032] Step S53: traverse the SOC start_n and SOC end_n , perform linear transformation on the SOC time series data of the order to be corrected:

[0033]

[0034] in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation;

[0035] Based on the above Constructing a sequence of transformation charging curves

[0036] Step S54: convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S benchmark and Construct a matrix A of size N*M using the Euclidean distance between any two points in:

[0037]

[0038] Among them, d MN Indicates S benchmark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity;

[0039] Step S55: Calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,...,W k};

[0040] Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W k represents the kth shortest distance;

[0041] Step S56: Based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value;

[0042] The step S60 is specifically as follows:

[0043] Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

[0044] In a second aspect, the present invention provides a charging SOC correction system based on dynamic time warping, comprising the following modules:

[0045] The historical charging order acquisition module is used to obtain a large number of historical charging orders for new energy vehicles and the corresponding charging process data;

[0046] A historical charging order classification module, used for classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC;

[0047] A high SOC interval order correction module, configured to perform SOC correction on each of the high SOC interval orders by an inverse ampere-hour integration method to obtain a first corrected SOC value;

[0048] A reference charging curve set building module, configured to build a reference charging curve set based on the first corrected SOC value and charging process data;

[0049] A low SOC interval order correction module, configured to perform SOC correction on the low SOC interval order through the reference charging curve set based on a dynamic time warping algorithm to obtain a second corrected SOC value;

[0050] The SOC correction curve generating module is used to generate an SOC correction curve based on the first correction SOC value and the second correction SOC value to complete the correction of the charging SOC.

[0051] Furthermore, the historical charging order acquisition module is specifically used for:

[0052] Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained;

[0053] The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

[0054] Furthermore, the historical charging order classification module is specifically used for:

[0055] When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders;

[0056] When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

[0057] Furthermore, in the high SOC interval order correction module, the calculation formula of the first corrected SOC value is:

[0058]

[0059] Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment factor.

[0060] Furthermore, the reference charging curve set construction module is specifically used for:

[0061] Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve;

[0062] The first charging curve is a curve showing that the charging voltage changes with the first corrected SOC value; the second charging curve is a curve showing that the charging current changes with the first corrected SOC value; and the third charging curve is a curve showing that the temperature changes with the first corrected SOC value;

[0063] The low SOC interval order correction module specifically includes:

[0064] A reference charging curve sequence matching unit is used to select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set as the reference charging curve sequence S based on the charging rate, charging current or charging date. benchmark ;

[0065] A correction candidate set construction unit is used to obtain a correction candidate set SOC of the starting SOC of the order to be corrected, using 1% SOC as a resolution interval and 5% SOC as a search interval. start_n And the correction candidate set SOC of the end SOC end_n ;

[0066] SOC start_n ={SOC star -5,...,SOC star -1,SOC star ,SOC star +1,…,SOC star +5,};

[0067] SOC end_n ={SOC end -5,…,SOC end -1,SOC end ,SOC end +1,…,SOC end +5,};

[0068] Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC;

[0069] Linear transformation unit for traversing the SOC start_n and SOC end_n , perform linear transformation on the SOC time series data of the order to be corrected:

[0070]

[0071] in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation;

[0072] Based on the above Constructing a sequence of transformation charging curves

[0073] A matrix A constructing unit, used to convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S benchmark and Construct a matrix A of size N*M using the Euclidean distance between any two points in:

[0074]

[0075] Among them, d MN Indicates Sbenchmark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity;

[0076] The optimal planning path calculation unit is used to calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,...,W k};

[0077] Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W k represents the kth shortest distance;

[0078] The correction unit for the order to be corrected is used to correct the order based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value;

[0079] The SOC correction curve generation module is specifically used for:

[0080] Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

[0081] The advantages of the present invention are:

[0082] By acquiring a large number of historical charging orders of new energy vehicles and the corresponding charging process data, each historical charging order is classified into a high SOC interval order and a low SOC interval order based on SOC, and SOC correction is performed on each high SOC interval order through the inverse ampere-hour integration method to obtain a first corrected SOC value; then a benchmark charging curve set is constructed based on the first corrected SOC value and the charging process data; based on the dynamic time warping algorithm, SOC correction is performed on the low SOC interval order through the benchmark charging curve set to obtain a second corrected SOC value; finally, an SOC correction curve is generated based on each first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC; that is, historical charging orders are classified into high SOC interval orders and low SOC interval orders, SOC correction is performed on the high SOC interval orders and a benchmark charging curve set is constructed, and then SOC correction is performed on the low SOC interval orders based on the dynamic time warping algorithm and the benchmark charging curve set, so as to realize the correction of the SOC of any charging SOC segment, thereby greatly improving the accuracy of the charging SOC correction, and providing more accurate sequence input for downstream lithium battery state estimation and prediction tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] The present invention will be further described below in conjunction with embodiments with reference to the accompanying drawings.

[0084] Figure 1 It is a flow chart of a charging SOC correction method based on dynamic time regularization of the present invention.

[0085] Figure 2 It is a structural schematic diagram of a charging SOC correction system based on dynamic time regularization of the present invention. DETAILED DESCRIPTION

[0086] The technical solution in the embodiment of the present application has the following overall idea: historical charging orders are classified into high SOC interval orders and low SOC interval orders, SOC correction is performed on the high SOC interval orders and a benchmark charging curve set is constructed, and then SOC correction is performed on the low SOC interval orders based on the dynamic time warping algorithm and the benchmark charging curve set, so as to realize SOC correction of any charging SOC segment and improve the accuracy of charging SOC correction.

[0087] Please refer to Figure 1 to Figure 2 As shown, a preferred embodiment of a charging SOC correction method based on dynamic time warping of the present invention includes the following steps:

[0088] Step S10, obtaining a large number of historical charging orders for new energy vehicles and corresponding charging process data;

[0089] Step S20, classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC;

[0090] Step S30, performing SOC correction on each of the high SOC interval orders by the inverse ampere-hour integration method to obtain a first corrected SOC value; that is, taking the end SOC of the high SOC interval order as the starting point, reversely calculating the charged power;

[0091] Step S40, constructing a reference charging curve set based on the first corrected SOC value and charging process data;

[0092] Step S50, based on a dynamic time warping (DTW) algorithm, performing SOC correction on the low SOC interval order through the reference charging curve set to obtain a second corrected SOC value;

[0093] Step S60: Generate an SOC correction curve based on the first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC.

[0094] The step S10 is specifically as follows:

[0095] Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained;

[0096] The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

[0097] The step S20 is specifically as follows:

[0098] When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders;

[0099] When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

[0100] It can be considered that the SOC of the high SOC interval order is accurate, and the SOC of the low SOC interval order is corrected by using the SOC of the high SOC interval order.

[0101] In step S30, the calculation formula of the first corrected SOC value is:

[0102]

[0103] Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment coefficient, which is obtained by looking up the table. Due to differences in charging efficiency at different charging rates and temperatures due to battery aging, the SOC adjustment coefficient is set.

[0104] The step S40 is specifically as follows:

[0105] Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve for subsequent DWT calculation;

[0106] The first charging curve is a curve showing that the charging voltage varies with the first corrected SOC value; the second charging curve is a curve showing that the charging current varies with the first corrected SOC value; and the third charging curve is a curve showing that the temperature varies with the first corrected SOC value;

[0107] The dynamic time warping algorithm is a method that measures the similarity of two time series of unequal lengths according to the principle of nearest distance. It is a nonlinear warping technology. It is generally believed that the aging process of lithium batteries is relatively slow, so in the short term, if the charging conditions are similar (such as charging rate strategy, charging temperature is similar), then the voltage curves in the same SOC interval should be as consistent as possible. Therefore, SOC correction is implemented for low SOC interval orders based on the benchmark charging curve set.

[0108] Due to the dynamic time warping algorithm's evaluation of the similarity of two sequences, the default sequence is aligned at the beginning and end, and the starting SOC and ending SOC of the charging data to be corrected may not be correct. However, the SOC estimation error of a general BMS is not too large. Therefore, the starting and ending SOCs are evaluated multiple times with 1% SOC as the resolution interval and ±5% as the retrieval interval, and the minimum DTW calculated value is finally selected as the charging curve corresponding to the corrected SOC.

[0109] The step S50 specifically includes:

[0110] Step S51: select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set based on the charging rate, charging current or charging date as the reference charging curve sequence S. benchmark ;

[0111] Step S52: Using 1% SOC as the resolution interval and 5% SOC as the search interval, obtain the correction candidate set SOC of the starting SOC of the order to be corrected. start_n And the correction candidate set SOC of the end SOC end_n ;

[0112] SOC start_n ={SOC star -5,...,SOC star -1,SOC star ,SOC star +1,...,SOC star +5,};

[0113] SOC end_n = {SOC end -5,…,SOC end -1,SOC end ,SOCend +1,…,SOC end +5,};

[0114] Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC;

[0115] Step S53: traverse the SOC start_n and SOC end_n , the SOC time series data of the order to be corrected is linearly transformed, even if the SOC range of the order to be corrected is from [SOC star ,SOC end ] to [SOC star_n ,SOC end_n ]:

[0116]

[0117] in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation;

[0118] Based on the above Constructing a sequence of transformation charging curves

[0119] Step S54: convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S benchmark and Construct a matrix A of size N*M using the Euclidean distance between any two points in:

[0120]

[0121] Among them, d MN Indicates S benchmark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity;

[0122] Step S55: Calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,...,W k};

[0123] Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W k represents the kth shortest distance;

[0124] Step S56: Based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value;

[0125] The step S60 is specifically as follows:

[0126] Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

[0127] A preferred embodiment of a charging SOC correction system based on dynamic time warping of the present invention includes the following modules:

[0128] The historical charging order acquisition module is used to obtain a large number of historical charging orders for new energy vehicles and the corresponding charging process data;

[0129] A historical charging order classification module, used for classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC;

[0130] A high SOC interval order correction module, used to perform SOC correction on each of the high SOC interval orders by an inverse ampere-hour integration method to obtain a first corrected SOC value; that is, taking the end SOC of the high SOC interval order as a starting point, reversely calculating the charged power;

[0131] A reference charging curve set building module, configured to build a reference charging curve set based on the first corrected SOC value and charging process data;

[0132] A low SOC interval order correction module, configured to perform SOC correction on the low SOC interval order through the reference charging curve set based on a dynamic time warping algorithm (Dynamic Time Warping, DTW), to obtain a second corrected SOC value;

[0133] The SOC correction curve generating module is used to generate an SOC correction curve based on the first correction SOC value and the second correction SOC value to complete the correction of the charging SOC.

[0134] The historical charging order acquisition module is specifically used for:

[0135] Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained;

[0136] The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

[0137] The historical charging order classification module is specifically used for:

[0138] When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders;

[0139] When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

[0140] It can be considered that the SOC of the high SOC interval order is accurate, and the SOC of the low SOC interval order is corrected by using the SOC of the high SOC interval order.

[0141] In the high SOC interval order correction module, the calculation formula of the first corrected SOC value is:

[0142]

[0143] Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment coefficient, which is obtained by looking up the table. Due to differences in charging efficiency at different charging rates and temperatures due to battery aging, the SOC adjustment coefficient is set.

[0144] The reference charging curve set building module is specifically used for:

[0145] Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve for subsequent DWT calculation;

[0146] The first charging curve is a curve showing that the charging voltage changes with the first corrected SOC value; the second charging curve is a curve showing that the charging current changes with the first corrected SOC value; and the third charging curve is a curve showing that the temperature changes with the first corrected SOC value;

[0147] The dynamic time warping algorithm is a method that measures the similarity of two time series of unequal lengths according to the principle of nearest distance. It is a nonlinear warping technology. It is generally believed that the aging process of lithium batteries is relatively slow, so in the short term, if the charging conditions are similar (such as charging rate strategy, charging temperature is similar), then the voltage curves in the same SOC interval should be as consistent as possible. Therefore, SOC correction is implemented for low SOC interval orders based on the benchmark charging curve set.

[0148] Due to the dynamic time warping algorithm's evaluation of the similarity of two sequences, the default sequence is aligned at the beginning and end, and the starting SOC and ending SOC of the charging data to be corrected may not be correct. However, the SOC estimation error of a general BMS is not too large. Therefore, the starting and ending SOCs are evaluated multiple times with 1% SOC as the resolution interval and ±5% as the retrieval interval, and the minimum DTW calculated value is finally selected as the charging curve corresponding to the corrected SOC.

[0149] The low SOC interval order correction module specifically includes:

[0150] A reference charging curve sequence matching unit is used to select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set as the reference charging curve sequence S based on the charging rate, charging current or charging date. benchmark ;

[0151] A correction candidate set construction unit is used to obtain a correction candidate set SOC of the starting SOC of the order to be corrected, using 1% SOC as a resolution interval and 5% SOC as a search interval. start_n And the correction candidate set SOC of the end SOC end_n ;

[0152] SOC start_n ={SOC star -5,...,SOC star -1,SOC star ,SOC star +1,...,SOC star +5,};

[0153] SOC end_n ={SOC end -5,...,SOC end -1,SOCend ,SOC end +1,...,SOC end +5,};

[0154] Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC;

[0155] Linear transformation unit, used to traverse the SOC start_n and SOC end_n , the SOC time series data of the order to be corrected is linearly transformed, even if the SOC range of the order to be corrected is from [SOC star ,SOC end ] to [SOC star_n ,SOC end_n ]:

[0156]

[0157] in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation;

[0158] Based on the above Constructing a sequence of transformation charging curves ;

[0159] A matrix A constructing unit, used to convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S bench mark and Construct a matrix A of size N*M using the Euclidean distance between any two points in:

[0160]

[0161] Among them, d MN Indicates S bench mark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity;

[0162] The optimal planning path calculation unit is used to calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,...,W k};

[0163] Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W krepresents the kth shortest distance;

[0164] The correction unit for the order to be corrected is used to correct the order based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value;

[0165] The SOC correction curve generating module is specifically used for:

[0166] Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

[0167] In summary, the advantages of the present invention are:

[0168] By acquiring a large number of historical charging orders of new energy vehicles and the corresponding charging process data, each historical charging order is classified into a high SOC interval order and a low SOC interval order based on SOC, and SOC correction is performed on each high SOC interval order through the inverse ampere-hour integration method to obtain a first corrected SOC value; then a benchmark charging curve set is constructed based on the first corrected SOC value and the charging process data; based on the dynamic time warping algorithm, SOC correction is performed on the low SOC interval order through the benchmark charging curve set to obtain a second corrected SOC value; finally, an SOC correction curve is generated based on each first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC; that is, historical charging orders are classified into high SOC interval orders and low SOC interval orders, SOC correction is performed on the high SOC interval orders and a benchmark charging curve set is constructed, and then SOC correction is performed on the low SOC interval orders based on the dynamic time warping algorithm and the benchmark charging curve set, so as to realize the correction of the SOC of any charging SOC segment, thereby greatly improving the accuracy of the charging SOC correction, and providing more accurate sequence input for downstream lithium battery state estimation and prediction tasks.

[0169] Although the specific implementation modes of the present invention are described above, those skilled in the art should understand that the specific implementation modes described are only illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A charging SOC correction method based on dynamic time warping, characterized in that: The steps include: Step S10, obtaining a large number of historical charging orders for new energy vehicles and corresponding charging process data; Step S20, classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC; Step S30, performing SOC correction on each of the high SOC interval orders by an inverse ampere-hour integration method to obtain a first corrected SOC value; Step S40, constructing a reference charging curve set based on the first corrected SOC value and charging process data; Step S50: Based on the dynamic time warping algorithm, perform SOC correction on the low SOC interval order through the reference charging curve set to obtain a second corrected SOC value; Step S60: Generate an SOC correction curve based on the first corrected SOC value and the second corrected SOC value to complete the correction of the charging SOC.

2. The charging SOC correction method based on dynamic time warping according to claim 1, characterized in that: The step S10 is specifically as follows: Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained; The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

3. The charging SOC correction method based on dynamic time warping according to claim 1, characterized in that: The step S20 is specifically as follows: When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders; When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

4. The charging SOC correction method based on dynamic time warping according to claim 1, characterized in that: In step S30, the calculation formula of the first corrected SOC value is: Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment factor.

5. The charging SOC correction method based on dynamic time warping according to claim 1, characterized in that: The step S40 is specifically as follows: Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve; The first charging curve is a curve showing that the charging voltage changes with the first corrected SOC value; the second charging curve is a curve showing that the charging current changes with the first corrected SOC value; and the third charging curve is a curve showing that the temperature changes with the first corrected SOC value; The step S50 specifically includes: Step S51: select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set based on the charging rate, charging current or charging date as the reference charging curve sequence S. benchmark ; Step S52: Using 1% SOC as the resolution interval and 5% SOC as the search interval, obtain the correction candidate set SOC of the starting SOC of the order to be corrected. start_n And the correction candidate set SOC of the end SOC end_n ; SOCIETY start_n ={SOC star -5,...,SOC star -1,SOC star ,SOC star +1,…,SOC star +5,}; SOCIETY end_n ={SOC end -5,…,SOC end -1,SOC end ,SOC end +1,…,SOC end +5,}; Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC; Step S53: traverse the SOC start_n and SOC end_n , perform linear transformation on the SOC time series data of the order to be corrected: in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation; Based on the above Constructing a sequence of transformation charging curves Step S54: convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S benchmark and Construct a matrix A of size N*M using the Euclidean distance between any two points in: Among them, d MN Indicates S benchmark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity; Step S55: Calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,…,W k }; Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W k represents the kth shortest distance; Step S56: Based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value; The step S60 is specifically as follows: Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

6. A charging SOC correction system based on dynamic time warping, characterized in that: Includes the following modules: The historical charging order acquisition module is used to obtain a large number of historical charging orders for new energy vehicles and the corresponding charging process data; A historical charging order classification module, used for classifying each of the historical charging orders into a high SOC interval order and a low SOC interval order based on SOC; A high SOC interval order correction module, configured to perform SOC correction on each of the high SOC interval orders by an inverse ampere-hour integration method to obtain a first corrected SOC value; A reference charging curve set building module, configured to build a reference charging curve set based on the first corrected SOC value and charging process data; A low SOC interval order correction module, configured to perform SOC correction on the low SOC interval order through the reference charging curve set based on a dynamic time warping algorithm to obtain a second corrected SOC value; The SOC correction curve generating module is used to generate an SOC correction curve based on the first correction SOC value and the second correction SOC value to complete the correction of the charging SOC.

7. A charging SOC correction system based on dynamic time warping as claimed in claim 6, characterized in that: The historical charging order acquisition module is specifically used for: Based on the VIN code of new energy vehicles, a large number of historical charging orders of new energy vehicles and corresponding charging process data are obtained; The historical charging order at least carries the battery type; the charging process data at least includes the charging voltage, charging current, charging rate, temperature, SOC, charging time and charging date.

8. The charging SOC correction system based on dynamic time warping according to claim 6, characterized in that: The historical charging order classification module is specifically used for: When the battery type is a ternary lithium battery, the historical charging orders with SOC ≥ 93% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders; When the battery type is a lithium iron phosphate battery, the historical charging orders with SOC ≥ 97% at the end of charging are classified as high SOC interval orders, and among the remaining historical charging orders, the historical charging orders with charging increasing SOC ≥ 30% are classified as low SOC interval orders.

9. The charging SOC correction system based on dynamic time warping according to claim 6, characterized in that: In the high SOC interval order correction module, the calculation formula of the first corrected SOC value is: Among them, SOC i-caribration represents the first corrected SOC value at the i-th moment; SOC end Indicates the SOC at the end of charging; Q end Indicates the charged capacity at the end of charging; Q i represents the charging capacity at the i-th moment, Q i =∫i(t)dt, i(t) represents the charging current at the i-th moment; Q rated It represents the rated capacity of the lithium battery; γ represents the SOC adjustment factor.

10. The charging SOC correction system based on dynamic time warping according to claim 6, characterized in that: The reference charging curve set building module is specifically used for: Based on the first corrected SOC value and the charging process data, construct a reference charging curve set including a first charging curve, a second charging curve and a third charging curve; The first charging curve is a curve showing that the charging voltage changes with the first corrected SOC value; the second charging curve is a curve showing that the charging current changes with the first corrected SOC value; and the third charging curve is a curve showing that the temperature changes with the first corrected SOC value; The low SOC interval order correction module specifically includes: A reference charging curve sequence matching unit is used to select one of the historical charging orders from the low SOC interval orders as the order to be corrected, and select the closest curve from the reference charging curve set as the reference charging curve sequence S based on the charging rate, charging current or charging date. benchmark ; A correction candidate set construction unit is used to obtain a correction candidate set SOC of the starting SOC of the order to be corrected, using 1% SOC as a resolution interval and 5% SOC as a search interval. start_n And the correction candidate set SOC of the end SOC end_n ; SOCIETY start_n ={SOC star -5,...,SOC star -1,SOC star ,SOC star +1,…,SOC star +5,}; SOCIETY end_n ={SOC end -5,...,SOC end -1,SOC end ,SOC end +1,...,SOC end +5,}; Among them, SOC star Indicates the starting SOC; SOC end Indicates the end of SOC; Linear transformation unit for traversing the SOC start_n and SOC end_n , perform linear transformation on the SOC time series data of the order to be corrected: in, It represents the SOC value at the i-th moment after linear transformation; SOC i represents the SOC value at the i-th moment before the linear transformation; Based on the above Constructing a sequence of transformation charging curves A matrix A constructing unit, used to convert the reference charging curve sequence S benchmark And change the charging curve sequence As input, S benchmark and Construct a matrix A of size N*M using the Euclidean distance between any two points in: Among them, d MN Indicates S benchmark The Mth point in The Euclidean distance of the Nth point in , i.e., the similarity; The optimal planning path calculation unit is used to calculate the shortest distance from A(1,1) to A(N,M) according to the dynamic time warping algorithm to obtain the optimal planning path W best ={W1,W2,...,W k }; Among them, A(1,1) represents the value of the first row and first column in matrix A; A(N,M) represents the value of the Nth row and Mth column in matrix A; W k represents the kth shortest distance; The correction unit for the order to be corrected is used to correct the order based on the optimal planning path W best get With S benchmark and performing SOC correction on the order to be corrected based on the mapping relationship to obtain a second corrected SOC value; The SOC correction curve generation module is specifically used for: Based on the first corrected SOC value and the second corrected SOC value, a SOC correction curve is generated by a bilinear difference method to complete the correction of the charging SOC.

Images