A battery pack temperature estimation method based on measurable parameter difference information

Abstract

The application discloses a battery pack temperature estimation method based on measurable parameter difference information and relates to the technical field of lithium ion batteries. A reference battery is selected to calculate the Euclidean distance of terminal voltage between the reference battery and all other single batteries provided with temperature sensors, so as to determine a target battery; input features are constructed and collected; the reference battery is repeatedly selected to obtain a modeling data set; a temperature estimation data-driven model is established; the Euclidean distance of terminal voltage between a single battery not provided with a temperature sensor and all single batteries provided with temperature sensors is calculated, so as to obtain an index value; input features are calculated and collected based on the index value; the collection is input to the temperature estimation data-driven model, and the measured temperature of the target battery is taken as an estimated temperature. The input features are constructed based on the measurable parameter difference information, the modeling data set of the target battery index is obtained based on the distance index, and then the temperature estimation data-driven model is established, so that the robustness is improved, and the temperature estimation result is quickly predicted through the target battery index value.

Description

Technical Field

[0001] This invention relates to the field of lithium-ion battery technology, specifically a method for estimating battery pack temperature based on measurable parameter difference information. Background Technology

[0002] Accurate estimation and effective monitoring of the temperature of each individual cell in a battery pack are crucial for ensuring battery thermal safety, preventing thermal runaway, and ensuring safe driving of electric vehicles. However, most existing battery temperature estimation methods are based on individual cells, but a battery pack contains numerous cells in varying locations. Furthermore, the cooling system significantly impacts battery temperature estimation. Most importantly, due to limitations in cost, space, and wiring, only a limited number of temperature sensors are placed on the surfaces of critical individual cells within the battery pack, leaving the temperatures of other cells unmonitored. Therefore, there is an urgent need for an effective temperature estimation method for all individual cells within a battery pack.

[0003] Existing battery temperature estimation methods mainly fall into two categories: model-based and data-driven methods. Compared to model-based methods, which require consideration of complex heat generation and transfer mechanisms, battery location, battery pack structure, and whether the cooling system is operational to build a complex temperature estimation model, data-driven methods only need to establish a mapping relationship between parameters such as voltage, current, and state of charge and temperature. Furthermore, with the development of technologies such as artificial intelligence, big data, and vehicle-to-everything (V2X) communication, they have gained wider acceptance. However, traditional data-driven methods rely on large amounts of modeling data for their data-driven models. Although limited measurable temperature information can further improve the temperature estimation results for the battery pack, the inconsistent heat generation and transfer mechanisms of individual cells in different locations within the battery pack lead to poor robustness of the data-driven models. Summary of the Invention

[0004] To address the shortcomings of the prior art, this invention provides a battery pack temperature estimation method based on measurable parameter difference information. It constructs input features based on measurable parameter difference information, obtains a modeling dataset of target battery indexes based on distance indicators, and then establishes a temperature estimation data-driven model, which helps to improve robustness. By setting the target battery index value of individual cells equipped with temperature sensors, the temperature estimation result can be quickly predicted.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a battery pack temperature estimation method based on measurable parameter difference information, comprising the following steps:

[0006] Step 1: Establishing a temperature estimation data-driven model

[0007] S1.1 Obtaining the modeling objective for temperature estimation data-driven model building

[0008] For battery packs with built-in temperature sensors at critical locations, select one individual cell with a temperature sensor as a reference cell. Calculate the Euclidean distance between the reference cell and all other individual cells with temperature sensors, using the following formula:

[0009]

[0010] In the formula, ED represents the calculation of Euclidean distance. This indicates the terminal voltage of a single cell equipped with a temperature sensor, in addition to the reference cell. This represents the terminal voltage of the reference battery, and m represents the number of individual batteries equipped with temperature sensors.

[0011] Then, the index value with the minimum distance difference is obtained by calculating the distance. tar The formula is as follows:

[0012]

[0013] Based on the index value tar A single cell equipped with a temperature sensor was selected as the target cell for building a temperature estimation data-driven model.

[0014] S1.2, Construction of modeling features for temperature estimation data-driven model building

[0015] Input feature 1 and input feature 2 are constructed using the heat generation to represent the mean and standard deviation of the heat generation difference sequence between individual cells, respectively, as follows:

[0016] Fea1=mean(Q p -Q b (3)

[0017] Fea2=std(Q p -Q b (4)

[0018] In the formula, Q p Q represents the heat generated by the target battery. b The heat generated by the reference battery;

[0019] Input features 3 and 4 are constructed using the terminal voltage to represent the average and standard deviation of the terminal voltage difference sequence between individual cells, respectively, as follows:

[0020] Fea3=mean(U t,p -U t,b (5)

[0021] Fea4=std(U t,p -U t,b (6)

[0022] Input feature 5 and input feature 6 are constructed using the ratio between the terminal voltages and are expressed as follows:

[0023] Fea5=mean(U t,p / U t,b (7)

[0024] Fea6=std(U t,p / U t,b (8)

[0025] In the formula, U t,p U is the terminal voltage of the target battery. t,b The terminal voltage of the reference battery;

[0026] The constructed input feature set Fea = [Fea1, Fea2, Fea3, Fea4, Fea5, Fea6] is used as the modeling features for establishing a temperature estimation data-driven model.

[0027] S1.3 Construction of the modeling dataset for temperature estimation data-driven model building

[0028] Select the remaining individual cells in the battery pack equipped with temperature sensors as reference cells, and repeat steps S1.1 and S1.2 to obtain all modeling objectives and features in the battery pack used for temperature estimation data-driven model building, as shown in the following formula:

[0029]

[0030] In the formula, D mod For modeling datasets;

[0031] S1.4 Establishment of a temperature estimation data-driven model

[0032] Based on the acquired modeling dataset, a data-driven model for temperature estimation is established using a data-driven algorithm, as shown in the following formula:

[0033] M data =f(D mod (10)

[0034] In the formula, M data For temperature estimation data-driven models, f is the data-driven algorithm;

[0035] Step 2: Obtaining the index value of a single battery cell without a temperature sensor

[0036] S2.1 Obtaining the index value of a single cell equipped with a temperature sensor and having minimal distance difference.

[0037] The formula for calculating the Euclidean distance between the terminal voltages of a single cell without a temperature sensor and all single cells in the battery pack equipped with temperature sensors is as follows:

[0038]

[0039] In the formula, The terminal voltage of a single battery cell without a temperature sensor. The voltage at the terminal of a single cell equipped with a temperature sensor is s, where s is the number of single cells without a temperature sensor.

[0040] Then, the index value with the minimum distance difference is obtained by calculating the distance. tar,n The formula is as follows:

[0041]

[0042] By obtaining the index value of the minimum distance difference between a single cell without a temperature sensor and a single cell with a temperature sensor, a basis is provided for subsequent temperature estimation of the single cell without a temperature sensor.

[0043] S2.2 Input Feature Calculation for Temperature Estimation Based on Measurable Parameter Difference Information

[0044] Based on the index value of the single cell equipped with a temperature sensor that has the smallest distance difference obtained in step S2.1, the input characteristics between the target cell corresponding to the obtained index value and the single cell without a temperature sensor are calculated as follows:

[0045] Fea'1=mean(Q n -Q p (13)

[0046] Fea'2=std(Q n -Q p (14)

[0047] Fea'3=mean(U t,n -U t,p (15)

[0048] Fea'4=std(U t,n -U t,p (16)

[0049] Fea'5=mean(U t,n / U t,p (17)

[0050] Fea'6=std(U t,n / U t,p (18)

[0051] In the formula, Q n For the heat generated by a single battery cell without a temperature sensor, U t,n The terminal voltage of a single battery cell without a temperature sensor;

[0052] The calculated set of input features, Fea' = [Fea'1, Fea'2, Fea'3, Fea'4, Fea'5, Fea'6], is used as the input features for estimating the temperature of a single cell without a temperature sensor.

[0053] Step 3: Temperature estimation of individual cells without temperature sensors

[0054] Input the set Fea' calculated in step S2.2 into the temperature estimation data-driven model established in step S1.4 to obtain the target battery index prediction value for estimating the temperature of a single cell without a temperature sensor, as shown in the following formula:

[0055] index pre,n =M data (Fea') (19)

[0056] When performing temperature estimation, index pre,n The measured temperature of the target battery is used as the estimated temperature of a single cell without a temperature sensor.

[0057] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention constructs input features based on measurable parameter difference information, obtains a modeling dataset of target battery indexes based on distance indicators, and then establishes a temperature estimation data-driven model, which helps to improve robustness. It no longer focuses on traditional direct data-driven modeling and temperature estimation, but instead uses the target battery index value of a single battery equipped with a temperature sensor to quickly predict the temperature estimation result of a single battery without a temperature sensor. This has the following advantages:

[0058] 1. It can maximize the use of temperature information from the limited number of individual cells in the battery pack that are equipped with temperature sensors, and realize temperature estimation for all individual cells in the pack that are not equipped with temperature sensors.

[0059] 2. It can be quickly deployed and applied to achieve efficient temperature estimation and monitoring of all individual cells within the battery pack;

[0060] 3. No need to consider complex heat generation and heat transfer mechanisms, no need for a large amount of operating data modeling, the temperature of all individual cells in the battery pack can be estimated quickly and effectively based on the relationship between temperature and measurable parameters.

[0061] 4. It eliminates the need to consider complex information such as the actual structure of the battery pack and whether the cooling system is on, making it more suitable for practical applications;

[0062] 5. It can provide a basis for predicting driving range and estimating aging status, and is also beneficial to the thermal safety and safe driving of electric vehicles. Attached Figure Description

[0063] Figure 1 This is a flowchart of the present invention;

[0064] Figure 2 This is a comparison chart of the temperature estimation results of individual cells 1-8 in the embodiment;

[0065] Figure 3 This is a comparison chart of the temperature estimation results of individual cells 9-16 in the embodiment. Detailed Implementation

[0066] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0067] like Figure 1 As shown, a battery pack temperature estimation method based on measurable parameter difference information includes the following steps:

[0068] Step 1: Establishing a temperature estimation data-driven model

[0069] S1.1 Obtaining the modeling objective for temperature estimation data-driven model building

[0070] For battery packs with built-in temperature sensors at critical locations, a single cell equipped with a temperature sensor is selected as a reference cell. Since there is a strong correlation between the terminal voltage and temperature of a single cell, and the Euclidean distance index can reflect the magnitude of this correlation, the Euclidean distance between the terminal voltages of the reference cell and all other single cells equipped with temperature sensors is calculated using the following formula:

[0071]

[0072] In the formula, ED represents the calculation of Euclidean distance. This indicates the terminal voltage of a single cell equipped with a temperature sensor, in addition to the reference cell. This represents the terminal voltage of the reference battery, and m represents the number of individual batteries equipped with temperature sensors.

[0073] Then, the index value with the minimum distance difference is obtained by calculating the distance. tar The formula is as follows:

[0074]

[0075] Based on the index value tar A single cell equipped with a temperature sensor was selected as the target cell for building a temperature estimation data-driven model.

[0076] S1.2, Construction of modeling features for temperature estimation data-driven model building

[0077] Input feature 1 and input feature 2 are constructed using the heat generation to represent the mean and standard deviation of the heat generation difference sequence between individual cells, respectively, as follows:

[0078] Fea1=mean(Q p -Q b (3)

[0079] Fea2=std(Q p -Q b (4)

[0080] In the formula, Q p Q represents the heat generated by the target battery. b The heat generated by the reference battery.

[0081] Input features 3 and 4 are constructed using the terminal voltage to represent the average and standard deviation of the terminal voltage difference sequence between individual cells, respectively, as follows:

[0082] Fea3=mean(U t,p -U t,b (5)

[0083] Fea4=std(U t,p -U t,b (6)

[0084] Furthermore, the ratio between the terminal voltages of different individual cells can also be used as a corresponding statistical indicator. Finally, input feature 5 and input feature 6 are constructed using the ratio between terminal voltages and are expressed as follows:

[0085] Fea5=mean(U t,p / U t,b (7)

[0086] Fea6=std(U t,p / U t,b (8)

[0087] In the formula, U t,p U is the terminal voltage of the target battery. t,b This is the terminal voltage of the reference battery.

[0088] The constructed input feature set Fea = [Fea1, Fea2, Fea3, Fea4, Fea5, Fea6] is used as the modeling features for establishing a temperature estimation data-driven model.

[0089] S1.3 Construction of the modeling dataset for temperature estimation data-driven model building

[0090] Select the remaining individual cells in the battery pack equipped with temperature sensors as reference cells, and repeat steps S1.1 and S1.2 to obtain all modeling objectives and features in the battery pack used for temperature estimation data-driven model building, as shown in the following formula:

[0091]

[0092] In the formula, D mod For modeling datasets;

[0093] S1.4 Establishment of a temperature estimation data-driven model

[0094] Based on the acquired modeling dataset, a data-driven model for temperature estimation is established using a data-driven algorithm, as shown in the following formula:

[0095] M data =f(D mod (10)

[0096] In the formula, M data For temperature estimation data-driven models, f is the data-driven algorithm;

[0097] Step 2: Obtaining the index value of a single battery cell without a temperature sensor

[0098] S2.1 Obtaining the index value of a single cell equipped with a temperature sensor and having minimal distance difference.

[0099] The formula for calculating the Euclidean distance between the terminal voltages of a single cell without a temperature sensor and all single cells in the battery pack equipped with temperature sensors is as follows:

[0100]

[0101] In the formula, The terminal voltage of a single battery cell without a temperature sensor. The value is the terminal voltage of a single cell equipped with a temperature sensor, and s is the number of single cells without a temperature sensor.

[0102] Then, the index value with the minimum distance difference is obtained by calculating the distance. tar,n The formula is as follows:

[0103]

[0104] By obtaining the index value of the minimum distance difference between a single cell without a temperature sensor and a single cell with a temperature sensor, a basis is provided for subsequent temperature estimation of the single cell without a temperature sensor.

[0105] S2.2 Input Feature Calculation for Temperature Estimation Based on Measurable Parameter Difference Information

[0106] Based on the index value of the single cell equipped with a temperature sensor that has the smallest distance difference obtained in step S2.1, the input characteristics between the target cell corresponding to the obtained index value and the single cell without a temperature sensor are calculated based on the measurable parameter difference information as follows:

[0107] Fea'1=mean(Q n -Q p (13)

[0108] Fea'2=std(Q n -Q p (14)

[0109] Fea'3=mean(U t,n -U t,p (15)

[0110] Fea'4=std(U t,n -U t,p (16)

[0111] Fea'5=mean(U t,n / U t,p (17)

[0112] Fea'6=std(U t,n / U t,p (18)

[0113] In the formula, Q n For the heat generated by a single battery cell without a temperature sensor, U t,n This represents the terminal voltage of a single cell without a temperature sensor.

[0114] The calculated set of input features, Fea' = [Fea'1, Fea'2, Fea'3, Fea'4, Fea'5, Fea'6], is used as the input features for estimating the temperature of a single cell without a temperature sensor, so as to obtain the temperature estimate of the single cell without a temperature sensor.

[0115] Step 3: Temperature estimation of individual cells without temperature sensors

[0116] Input the set Fea' calculated in step S2.2 into the temperature estimation data-driven model established in step S1.4 to obtain the target battery index prediction value for estimating the temperature of a single cell without a temperature sensor, as shown in the following formula:

[0117] index pre,n =M data (Fea') (19)

[0118] When performing temperature estimation, index pre,n The measured temperature of the target battery is used as the estimated temperature of a single cell without a temperature sensor.

[0119] Example

[0120] This embodiment verifies the method based on low-temperature heating battery pack data. The experiment is based on a battery pack with a built-in temperature sensor, specifically a 133Ah ternary lithium-ion battery. The sliding window size is set to 200. The operating condition data used for modeling is the battery pack experimental data at -20℃, and the test operating condition data used is the battery pack experimental data at -30℃. The temperature estimation results of different individual cells obtained by the method of this invention are shown in Table 1.

[0121] Table 1 Temperature estimation results for different individual cells

[0122]

[0123]

[0124] Specific temperature estimation results comparison chart Figures 2-3 As shown in the figure, the reference value curve is the actual temperature curve measured by the temperature sensor, and the estimated value curve is the estimated temperature curve obtained by the method of the present invention. It can be seen from the above experimental results that the method proposed in this invention can accurately estimate the temperature of individual cells in the battery pack.

[0125] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0126] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for estimating battery pack temperature based on measurable parameter difference information, characterized in that: Includes the following steps: Step 1: Establishing a temperature estimation data-driven model S1.1 Obtaining the modeling objective for temperature estimation data-driven model building For battery packs with built-in temperature sensors at critical locations, select one individual cell with a temperature sensor as a reference cell. Calculate the Euclidean distance between the reference cell and all other individual cells with temperature sensors, using the following formula: In the formula, ED represents the calculation of Euclidean distance. This indicates the terminal voltage of a single cell equipped with a temperature sensor, in addition to the reference cell. This represents the terminal voltage of the reference battery, and m represents the number of individual batteries equipped with temperature sensors. Then, the index value with the minimum distance difference is obtained by calculating the distance. tar The formula is as follows: Based on the index value tar A single cell equipped with a temperature sensor was selected as the target cell for building a temperature estimation data-driven model. S1.2, Construction of modeling features for temperature estimation data-driven model building Input feature 1 and input feature 2 are constructed using the heat generation to represent the mean and standard deviation of the heat generation difference sequence between individual cells, respectively, as follows: Fea1=mean(Q p -Q b ) (3) Fea2=std(Q p -Q b ) (4) In the formula, Q p Q represents the heat generated by the target battery. b The heat generated by the reference battery; Input features 3 and 4 are constructed using the terminal voltage to represent the average and standard deviation of the terminal voltage difference sequence between individual cells, respectively, as follows: Fea3=mean(U t,p -U t,b ) (5) Fea4=std(U t,p -U t,b ) (6) Input feature 5 and input feature 6 are constructed using the ratio between the terminal voltages and are expressed as follows: Fea5=mean(U t,p / U t,b ) (7) Fea6=std(U t,p / U t,b ) (8) In the formula, U t,p U is the terminal voltage of the target battery. t,b The terminal voltage of the reference battery; The constructed input feature set Fea = [Fea1, Fea2, Fea3, Fea4, Fea5, Fea6] is used as the modeling features for establishing a temperature estimation data-driven model. S1.3 Construction of the modeling dataset for temperature estimation data-driven model building Select the remaining individual cells in the battery pack equipped with temperature sensors as reference cells, and repeat steps S1.1 and S1.2 to obtain all modeling objectives and features in the battery pack used for temperature estimation data-driven model building, as shown in the following formula: In the formula, D mod For modeling datasets; S1.4 Establishment of a temperature estimation data-driven model Based on the acquired modeling dataset, a data-driven model for temperature estimation is established using a data-driven algorithm, as shown in the following formula: M data =f(D mod ) (10) In the formula, M data For temperature estimation data-driven models, f is the data-driven algorithm; Step 2: Obtaining the index value of a single battery cell without a temperature sensor S2.1 The index value of the individual battery cell equipped with a temperature sensor with the minimum distance difference is obtained. The Euclidean distance between the terminal voltages of the individual battery cell without a temperature sensor and all individual batteries in the battery pack equipped with temperature sensors is calculated using the following formula: In the formula, The terminal voltage of a single battery cell without a temperature sensor. The voltage at the terminal of a single cell equipped with a temperature sensor is s, where s is the number of single cells without a temperature sensor. Then, the index value with the minimum distance difference is obtained by calculating the distance. tar,n The formula is as follows: By obtaining the index value of the minimum distance difference between a single cell without a temperature sensor and a single cell with a temperature sensor, a basis is provided for subsequent temperature estimation of the single cell without a temperature sensor. S2.2 Input Feature Calculation for Temperature Estimation Based on Measurable Parameter Difference Information Based on the index value of the single cell equipped with a temperature sensor that has the smallest distance difference obtained in step S2.1, the input characteristics between the target cell corresponding to the obtained index value and the single cell without a temperature sensor are calculated as follows: Fea1'=mean(Q n -Q p ) (13) Fea'2=std(Q n -Q p ) (14) Fea'3=mean(U t,n -U t,p ) (15) Fea'4=std(U t,n -U t,p ) (16) Fea'5=mean(U t,n / U t,p ) (17) Fea'6=std(U t,n / U t,p ) (18) In the formula, Q n For the heat generated by a single battery cell without a temperature sensor, U t,n The terminal voltage of a single battery cell without a temperature sensor; The calculated set of input features, Fea' = [Fea1', Fea'2, Fea'3, Fea'4, Fea'5, Fea'6], is used as the input features for estimating the temperature of a single cell without a temperature sensor. Step 3: Temperature estimation of individual cells without temperature sensors Input the set Fea' calculated in step S2.2 into the temperature estimation data-driven model established in step S1.4 to obtain the target battery index prediction value for estimating the temperature of a single cell without a temperature sensor, as shown in the following formula: index pre,n =M data (Fea') (19) When performing temperature estimation, index pre,n The measured temperature of the target battery is used as the estimated temperature of a single cell without a temperature sensor.

Images