A SOH estimation method for series battery packs in mobile energy storage systems based on transfer learning

Abstract

The present invention discloses a method for estimating the SOH of a series battery pack of a mobile energy storage system based on transfer learning, including: allowing the battery pack to charge and discharge cyclically and suspend discharge for multiple times, collecting data during the period to construct an original data set; extracting features according to the original data set, and forming a feature data set with the original data; using an information entropy method to obtain feature weights, performing weighted summation to obtain a consistency estimation result, and merging it with the feature data set to form a new feature data set; constructing an LSTM model, and pre-training with the new feature data set to obtain a basic model; retraining with a small amount of data from other battery packs and a transfer learning strategy to obtain a TL‑LSTM model; and weighted summing the output results of the two models to obtain an SOH estimation result. The present invention does not require a specific voltage range as an estimation condition, and does not require the collection of a large amount of data for retraining in the face of different battery packs. By fusing the output results of the two models, the estimation accuracy and generalization ability are effectively improved, and the negative effects of negative migration are reduced.

Description

Technical Field

[0001] The present invention relates to the field of battery energy storage technology, and in particular to a method for estimating the SOH of a series battery pack of a mobile energy storage system based on transfer learning. Background Art

[0002] Battery mobile energy storage systems generally connect multiple batteries in series to form a battery pack, and then connect multiple battery packs in series to meet voltage requirements. In order to reduce the safety risks of battery mobile energy storage systems and ensure the operating performance of mobile energy storage systems, it is necessary to perform real-time health status (SOH) estimation on each battery pack of the mobile energy storage system. Mobile energy storage system battery packs usually do not fully discharge or set a specific voltage point as the discharge stop threshold. Existing estimation methods based on a specific voltage range or full discharge as features are difficult to apply directly. Secondly, due to the differences in parameters of each battery cell in the battery pack, inconsistencies will occur in the battery pack, which will cause the battery pack SOH to be different from the SOH of each battery cell.

[0003] Existing methods do not fully consider the consistency of battery packs. They directly treat the battery pack as a "big battery" for SOH estimation, which makes it difficult to ensure the estimation accuracy of the battery pack SOH. In addition, when faced with battery packs composed of different numbers of batteries, the commonly used deep learning-based methods need to re-collect a large amount of data and spend a lot of time retraining the model. Existing estimation methods based on transfer learning do not fully consider invalid migration and negative migration factors, which may lead to a decrease in estimation accuracy in practical applications. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a SOH estimation method for a series battery pack of a mobile energy storage system based on transfer learning, which can solve the problems of insufficient estimation accuracy and insufficient model generalization ability faced by the existing SOH estimation method for a battery pack of a mobile energy storage system.

[0007] In order to solve the above technical problems, the present invention provides the following technical solution, a method for estimating SOH of a series battery pack of a mobile energy storage system based on transfer learning, comprising:

[0008] Collect the current and voltage data of the battery pack during the whole process of cyclic charging and discharging and multiple discharge pauses to build the original data set;

[0009] Preprocessing the original data set to obtain a feature data set;

[0010] The information entropy method is used to obtain feature weights, and weighted summation is performed to obtain a consistency estimation result, and the result is combined with the feature data set to form a new feature data set;

[0011] Construct an LSTM model and perform pre-training on the new feature data set to obtain a basic prediction model;

[0012] A small amount of data from other battery packs and a transfer learning strategy were used to retrain the TL-LSTM model.

[0013] The battery pack data to be measured is input into the above two models, and the output results of the above two models are weighted and summed to obtain the SOH estimation result.

[0014] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, wherein: the original data set includes:

[0015] Note that the number of battery cells in the battery pack is P, and the discharge current of the battery pack is increased from 0 to the maximum rated current I max Select N current values ​​arithmetically to obtain a discharge current data set;

[0016] The battery pack is discharged from the cut-off voltage U pack-dis To charging cut-off voltage U pack-cha M voltage values ​​are selected arithmetically as voltage points at which the battery pack suspends discharging, and a voltage data set including the M voltage values ​​is obtained.

[0017] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack of the present invention, the preprocessing includes:

[0018] The formula for calculating the standard deviation rate is:

[0019]

[0020] Among them, x pj represents the jth feature of the pth battery, P represents the number of battery cells, represents the average value of the j-th feature.

[0021] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, the characteristic data set includes extracting the terminal voltage value of all battery cells at the sampling moment before the discharge is suspended, the instantaneous change value of the terminal voltage of all battery cells at the moment of suspension of discharge, the voltage recovery rate of the whole process of the voltage of all battery cells recovering to the steady-state value after the discharge is suspended, and the steady-state terminal voltage value of all battery cells as features.

[0022] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, the information entropy method includes:

[0023] The calculation formula of information entropy is:

[0024]

[0025] Among them, S j represents the information entropy of the jth feature, k is a constant, y pj represents the normalized value of the jth feature of the pth battery, and J represents the number of features.

[0026] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, the information entropy method also includes feature weights,

[0027] The weight calculation formula is:

[0028]

[0029] Among them, ω j represents the corresponding weight of the j-th feature.

[0030] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, the basic prediction model includes inputting the new feature data set into the LSTM network for pre-training to obtain the basic parameters of the model.

[0031] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, wherein: the basic prediction model also includes:

[0032] i t =σ(W i [h (t-1) ,x t ]+b i )

[0033] f t =σ(W f [h (t-1) ,x t ]+b f )

[0034] o t =σ(W o [h (t-1) ,x t ]+b o )

[0035] qt =tanh(W q [h (t-1) ,x t ]+b q )

[0036] c t =f t *c (t-1) +i t *q t

[0037] h t =o t *tanh(c t )

[0038] Among them, g t represents the output of the input node at the current moment, x t represents the input of the network at the current moment, h (t-1) 、h t Represent the output at the previous moment and the current moment respectively, W i , W f , W o , W q Respectively represent the weight parameters of the input gate, forget gate, output gate, and input node, b i 、b f 、b o 、b g They represent the bias parameters of the input gate, forget gate, output gate, and input node, respectively. t It represents the state value retained from the unit state of the previous moment to the current moment, and σ is the sigmoid function.

[0039] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, the weighted sum of the output results of the two models includes that the weights of the estimation results of the two networks are both taken as 0.5, and the weighted sum is performed to obtain the SOH of the battery pack.

[0040] As a preferred solution of the SOH estimation method of the mobile energy storage system series battery pack described in the present invention, wherein: the LSTM model includes, the LSTM model is composed of an input layer, a hidden layer and an output layer, the number of neurons in the input layer is determined according to the number of features of the new feature data set, the hidden layer includes a long short-term memory neural network neuron layer, a Dropout layer and a fully connected layer, the number of neurons in the output layer is 1, and the output result is the SOH value estimated by the neural network.

[0041] Beneficial effects of the present invention: The present invention proposes a method for estimating the SOH of a series battery pack of a mobile energy storage system based on transfer learning. The battery SOH estimation method proposed by the present invention does not require a specific voltage range as a use condition, and can meet the requirements for estimating the SOH of a battery pack in the actual operating conditions of a mobile energy storage system battery pack. The consistency factor of the battery pack is taken into consideration, and the consistency evaluation result of the battery pack is used as one of the features. At the same time, by extracting the effective information of all battery cells in the battery pack, the accuracy of the SOH estimation of the battery pack is fully improved; by combining the estimation results of the long short-term memory neural network model based on transfer learning and the basic long short-term memory neural network model, the generalization ability of the model is improved, the negative effects of invalid migration and negative migration are reduced, and effective guarantee is provided for the estimation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0043] Figure 1 A flow chart of a method for estimating SOH of a series-connected battery pack of a mobile energy storage system based on transfer learning provided in one embodiment of the present invention; DETAILED DESCRIPTION

[0044] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0047] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0048] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0049] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] Example 1

[0051] Reference Figure 1 , which is the first embodiment of the present invention, and provides a method for estimating SOH of a series battery pack of a mobile energy storage system based on transfer learning, comprising:

[0052] S1: Collect the current and voltage data of the battery pack during cyclic charging and discharging and multiple discharge pauses to build the original data set;

[0053] Furthermore, let the number of battery cells in the battery pack be P, and the discharge current of the battery pack is increased from 0 to the maximum rated current I max Select N current values ​​​​arbitrarily and obtain the discharge current data set as follows:

[0054] I=[I1,I2,...,I N ]

[0055] Furthermore, the battery pack is discharged from the cut-off voltage U pack-dis To charging cut-off voltage U pack-cha M voltage values ​​are selected arithmetically as the voltage points at which the battery pack suspends discharging, and a voltage data set including the M voltage values ​​is obtained as follows:

[0056] U=[U1,U2,...,U M ]

[0057] Furthermore, in each cycle, the battery pack selects a current from the obtained discharge current data set for constant current discharge, and suspends discharge M times according to the obtained M discharge suspension voltage points, and when the battery pack reaches the discharge cut-off voltage in each cycle, the SOH of the battery pack is calculated according to the total discharge amount.

[0058] It should be noted that each time the discharge is suspended, the battery pack terminal voltage and discharge current value are recorded, and the voltage data of all battery cells in the battery pack from the sampling moment before the discharge is suspended to the voltage recovery to the steady-state value at the end of the discharge suspension are collected;

[0059] Furthermore, the acquired data are merged to construct the original data set D ori .

[0060] S2: preprocessing the original data set to obtain a feature data set;

[0061] Furthermore, according to the original data set D collected by S1 ori , calculate the range and standard deviation rate of the terminal voltage value of all battery cells at the sampling moment before the suspension of discharge, calculate the standard deviation rate of the instantaneous change value of the terminal voltage of all battery cells at the moment of suspension of discharge, calculate the standard deviation rate of the voltage recovery rate of the whole process of the voltage recovery to the steady-state value after the suspension of discharge of all battery cells, calculate the range and standard deviation rate of the steady-state terminal voltage value of all battery cells, where the calculation formula of the standard deviation rate is:

[0062]

[0063] Among them, x pj represents the jth feature of the pth battery, P represents the number of battery cells, represents the average value of the j-th feature.

[0064] Furthermore, the obtained data is compared with the original feature data set D ori Merge to form feature dataset D f , and normalized.

[0065] S3: using the information entropy method to obtain feature weights, performing weighted summation to obtain a consistency estimation result, and merging the result with the feature data set to form a new feature data set;

[0066] Furthermore, the information entropy method is used to assign weights to each of the features. The calculation formula of information entropy is:

[0067]

[0068] Among them, Sj represents the information entropy of the jth feature, k is a constant, ypj represents the normalized value of the jth feature of the pth battery, and J represents the number of features.

[0069] The weight calculation formula is:

[0070]

[0071] Among them, ω j represents the corresponding weight of the j-th feature.

[0072] It should be noted that according to the obtained feature data set D f , extract the terminal voltage value of all battery cells at the sampling moment before the suspension of discharge, the instantaneous change value of the terminal voltage of all battery cells at the moment of suspension of discharge, the voltage recovery rate of the whole process of the voltage of all battery cells recovering to the steady-state value after the suspension of discharge, and the steady-state terminal voltage value of all battery cells as features;

[0073] Furthermore, a weighted sum calculation is performed based on each feature weight obtained to obtain the battery pack consistency evaluation result, and the consistency evaluation result is combined with the obtained feature data set D f Merge to form a new feature dataset D new .

[0074] S4: construct an LSTM model, and perform pre-training on the new feature data set to obtain a basic prediction model;

[0075] It should be noted that the long short-term memory neural network model is composed of an input layer, a hidden layer and an output layer. The number of neurons in the input layer is based on the new feature data set D obtained in step S3. new The number of features is determined, the hidden layer includes the long short-term memory neural network neuron layer, the Dropout layer and the fully connected layer, the number of neurons in the output layer is 1, and the output result is the SOH value estimated by the neural network. The neurons of the long short-term memory neural network are composed of the long-term state c t and the short-term state h t is composed of and depends on the forget gate f t , input gate i t and output gate o t , the specific calculation formula is as follows:

[0076] i t =σ(W i [h (t-1) ,x t ]+b i )

[0077] f t =σ(W f [h (t-1) ,x t ]+bf )

[0078] o t =σ(W o [h (t-1) ,x t ]+b o )

[0079] q t =tanh(W q [h (t-1) ,x t ]+b q )

[0080] c t =f t *c (t-1) +i t *q t

[0081] h t =o t *tanh(c t )

[0082] Among them, g t represents the output of the input node at the current moment, x t represents the input of the network at the current moment, h (t-1) 、h t Represent the output at the previous moment and the current moment respectively, W i , W f , W o , W q Respectively represent the weight parameters of the input gate, forget gate, output gate, and input node, b i 、b f 、b o 、b g They represent the bias parameters of the input gate, forget gate, output gate, and input node, respectively. t It represents the state value retained from the unit state at the previous moment to the current moment, and σ is the sigmoid function;

[0083] Furthermore, the new feature dataset D new Input into the LSTM network for pre-training to obtain the basic parameters of the model.

[0084] S5: Retrain a small amount of data from other battery packs and transfer learning strategies to obtain the TL-LSTM model;

[0085] It should be noted that the TL-LSTM model is designed for the final weighted summation.

[0086] S6: Input the battery pack data to be measured into the above two models, perform weighted summation on the output results of the above two models, and obtain the SOH estimation result.

[0087] Furthermore, the two long short-term memory neural network models obtained were imported into the battery management system of the mobile energy storage system;

[0088] It should be noted that the mobile energy storage system battery management system monitors and collects the operating status of all battery packs and all battery cells contained therein in real time. When a battery pack is detected to exit the discharge operation, the discharge current and terminal voltage of the battery pack are recorded, and the voltage data of the whole process of the terminal voltage of all battery cells in the battery pack returning to the steady-state value during the exit discharge operation is collected. The method is used to obtain the feature data set D from the original data. f , the battery pack is evaluated for consistency using the evaluation method, and a new feature data set D is formed new ;

[0089] Furthermore, the obtained new feature dataset D new The results are input into the two long short-term memory neural networks for estimation. The weights of the estimation results of the two networks are both taken as 0.5, and the weighted sum is taken to obtain the SOH of the battery pack.

[0090] Example 2

[0091] Reference Figure 1 , which is an embodiment of the present invention, provides a SOH estimation method for a series battery pack of a mobile energy storage system based on transfer learning. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0092] Table 1 compares the advantages and disadvantages of the present invention with the battery pack SOH estimation method based on partial voltage intervals and based on IC / DV curves

[0093]

[0094] As can be seen from Table 1, the improved battery pack SOH estimation method does not require data from the entire charging process, its characteristics cover multiple operating voltage ranges, and it has no requirements for the charging current size. It has significant advantages over the other two inventions and achieves the innovative purpose of this patent.

[0095] Table 2 compares the advantages and disadvantages of the present invention and the battery pack SOH estimation method based on no transfer learning and based on no transfer learning

[0096]

[0097] As shown in Table 2, the improved battery pack SOH estimation method saves the time of retraining in practical applications. The model can be applied to battery packs with different numbers of batteries. By integrating the output results of the transfer learning model and the non-transfer learning model, the negative impact caused by negative transfer can be effectively reduced. Compared with the other two inventions, it has significant advantages and achieves the innovative purpose of this patent.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0100] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0101] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0102] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0104] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0105] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for estimating SOH of a series battery pack of a mobile energy storage system based on transfer learning, characterized in that: include, Collect the current and voltage data of the battery pack during the whole process of cyclic charging and discharging and multiple discharge pauses to build the original data set; Preprocessing the original data set to obtain a feature data set; The information entropy method is used to obtain feature weights, and weighted summation is performed to obtain a consistency estimation result, and the result is combined with the feature data set to form a new feature data set; Construct an LSTM model and perform pre-training on the new feature data set to obtain a basic prediction model; A small amount of data from other battery packs and a transfer learning strategy were used to retrain the TL-LSTM model. The battery pack data to be measured is input into the basic prediction model and the TL-LSTM model, and the output results of the basic prediction model and the TL-LSTM model are weighted and summed to obtain the SOH estimation result; The original data set includes: Note that the number of battery cells in the battery pack is P, and the discharge current of the battery pack is increased from 0 to the maximum rated current I max Select N current values ​​arithmetically to obtain a discharge current data set; The battery pack is discharged from the cut-off voltage U pack-dis To charging cut-off voltage U pack-cha Arithmetically selecting M voltage values ​​as voltage points at which the battery pack suspends discharging, and obtaining a voltage data set including the M voltage values; The pretreatment includes: The formula for calculating the standard deviation rate is: Among them, x pj represents the jth feature of the pth battery, P represents the number of battery cells, represents the average value of the jth feature; The characteristic data set includes extracting the terminal voltage value of all battery cells at a sampling time before the discharge is suspended, the instantaneous change value of the terminal voltage of all battery cells at the time of the discharge suspension, the voltage recovery rate of the whole process of the voltage of all battery cells recovering to the steady-state value after the discharge is suspended, and the steady-state terminal voltage value of all battery cells as features; The information entropy method includes: The calculation formula of information entropy is: Among them, S j represents the information entropy of the jth feature, k is a constant, y pj represents the normalized value of the jth feature of the pth battery, and J represents the number of features; The information entropy method also includes feature weights, The weight calculation formula is: Among them, ω j represents the corresponding weight of the j-th feature.

2. The SOH estimation method for a mobile energy storage system series battery pack based on transfer learning as claimed in claim 1, characterized in that: The basic prediction model includes inputting a new feature data set into an LSTM network for pre-training to obtain basic parameters of the model.

3. A method for estimating SOH of a series-connected battery pack of a mobile energy storage system based on transfer learning as claimed in claim 2, characterized in that: The basic prediction model also includes: i t =σ(W i [h (t-1) ,x t ]+b i ) f t =σ(W f [h (t-1) ,x t ]+b f ) the t =σ(W o [h (t-1) ,x t ]+b o ) q t =tanh(W q [h (t-1) ,x t ]+b q ) c t =f t *c (t-1) +i t *q t h t =o t *tanh(c t ) Among them, g t represents the output of the input node at the current moment, x t represents the input of the network at the current moment, h (t-1) 、h t Represent the output at the previous moment and the current moment respectively, W i , W f , W o , W q Respectively represent the weight parameters of the input gate, forget gate, output gate, and input node, b i , b f , b o , b g They represent the bias parameters of the input gate, forget gate, output gate, and input node, respectively. t It represents the state value retained from the unit state of the previous moment to the current moment, and σ is the sigmoid function.

4. The SOH estimation method for a mobile energy storage system series battery pack based on transfer learning as claimed in claim 3 is characterized in that: The weighted sum of the output results of the basic prediction model and the TL-LSTM model includes taking the weights of the estimation results of the two networks as 0.5, and performing weighted summation to obtain the SOH of the battery pack.

5. The SOH estimation method for a mobile energy storage system series battery pack based on transfer learning as claimed in claim 4 is characterized in that: The LSTM model includes, The LSTM model consists of an input layer, a hidden layer and an output layer. The number of neurons in the input layer is determined according to the number of features of the new feature data set. The hidden layer includes a long short-term memory neural network neuron layer, a Dropout layer and a fully connected layer. The number of neurons in the output layer is 1, and the output result is the SOH value estimated by the neural network.

Images