Method for testing state of charge of energy storage battery pack

Abstract

The invention discloses a method for testing the state of charge of an energy storage battery pack, and the method comprises the following steps: determining the type of a single body test signal based on the real-time temperature, accumulated cycle index and historical operation data of a target single body; under the condition that the energy storage battery pack is kept running, applying a single test signal to the corresponding single battery through any single controller to perform cycle life test, acquiring a response result based on the single test signal, and synchronously offsetting power fluctuation generated by the test of the target single battery; acquiring the voltage variation, the current variation, the temperature gradient, the vibration frequency and the environment temperature of the target monomer corresponding to the single response result; and calculating the state of charge of the target monomer by combining the response result and the voltage variation, the current variation, the temperature gradient, the vibration frequency and the environment temperature of the target monomer corresponding to the single response result with a pre-trained neural network model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery testing, in particular to a state of charge testing method for an energy storage battery pack. BACKGROUND

[0002] The operation safety and service life of an energy storage battery pack, which is a core energy storage unit in the fields of new energy power generation, electric vehicles and emergency power supply, directly depend on the accurate monitoring of the state of charge (SOC), and the calculation accuracy of the SOC is closely related to the impedance characteristics of the single battery. The cell impedance is not only a core parameter reflecting the SOC, state of health (SOH), state of power (SOP) and safety state (SOS), but also an early warning indicator for cell performance degradation and safety hazards, which is of great significance for the fine management of the battery pack.

[0003] However, in the state of charge testing method of the existing energy storage battery pack, the existing technology either adopts an offline testing mode, which requires the target single battery to be disassembled and separated from the battery pack, thereby interrupting the normal operation of the battery pack and causing the SOC equalization problem when the offline cell is reconnected, and the operation is complex and inefficient; or although online testing is attempted, the test signal is not adapted to the real-time temperature, cumulative cycle number and historical operation stability of the single battery, and only a single signal type is used for excitation, so it is difficult to accurately capture the electrochemical response of the single battery under different states, and the power fluctuation generated in the testing process will interfere with the overall operation stability of the battery pack, further reducing the SOC calculation accuracy. In addition, the existing online testing technology cannot effectively integrate the impedance response and multi-dimensional state parameters such as voltage variation, temperature gradient and vibration frequency, resulting in a large SOC prediction error, which is difficult to meet the application requirements of high accuracy and high stability, and limits the operation and maintenance level and service life of the battery pack. SUMMARY

[0004] In view of the above existing technology, the present application provides a state of charge testing method for an energy storage battery pack, which mainly solves the technical problems in the background art.

[0005] To achieve the above purpose, the technical solution of the embodiment of the present application is as follows: a state of charge testing method for an energy storage battery pack, the testing method comprising the following steps: determining the type of single battery test signal based on the real-time temperature, cumulative cycle number and historical operation data of the target single battery; under the condition that the energy storage battery pack remains in operation, applying a single battery test signal to the corresponding single battery through any single battery controller for cycle life testing, and obtaining a response result based on the single battery test signal, and simultaneously offsetting the power fluctuation of the target single battery caused by the test; acquire the voltage variation, the current variation, the temperature gradient, the vibration frequency and the ambient temperature of the target cell corresponding to the single response result; combine the response result and the voltage variation, the current variation, the temperature gradient, the vibration frequency and the ambient temperature of the target cell corresponding to the single response result with the pre-trained neural network model to calculate the state of charge of the target cell.

[0006] Optionally, based on the real-time temperature, the cumulative cycle number and the historical operation data of the target cell, the cell test signal type is determined, specifically including: based on the real-time temperature of the target cell, determining the preliminary screening signal type, adjusting the amplitude and frequency parameters of the preliminary screening signal based on the aging degree reflected by the cumulative cycle number; in combination with the power stability characteristics embodied in the historical operation data, the adjusted signal is optimized to finally determine the test signal type adapted to the current working condition of the target cell.

[0007] Optionally, the test signal type includes a linear signal, a step signal, a sinusoidal signal and a pulse signal.

[0008] Optionally, the response result based on the cell test signal is acquired, specifically including: real-time acquisition of the electrical data of the target cell under the action of the test signal, and calculation of the impedance value of the target cell in combination with the test signal type as the response result based on the cell test signal.

[0009] Optionally, the power fluctuation of the target cell caused by the test is synchronously offset, specifically including: designating other cells, outputting a complementary signal corresponding to the test signal of the target cell through the cell controller of the other cells to offset the power fluctuation of the target cell caused by the test.

[0010] Optionally, the other cells are designated, specifically including: based on the real-time measurable state parameters of each cell in the battery pack, screening a plurality of complementary cells, and in the screened plurality of complementary cells, selecting the complementary cells in the same branch or adjacent position as the target cell as the final designated cells.

[0011] Optionally, the complementary signal corresponding to the test signal of the target cell is output through the cell controller, specifically including: selecting one or more other cells with stable state as compensation cells according to the current power of the target cell and the power variation of the test signal; based on the instantaneous power fluctuation value caused by the test signal of the target cell, calculating the complementary signal power required to be output by the compensation cells; through the cell controller of the compensation cells, the complementary signal is output in real time to dynamically offset the test interference of the target cell, and ensure that the overall voltage and power output fluctuation of the battery pack is controlled within a preset range.

[0012] Optionally, the complementary signal power is equal in size and opposite in direction to the instantaneous power fluctuation value.

[0013] Optionally, the response result and corresponding voltage change amount, current change amount, temperature gradient, vibration frequency, and ambient temperature are preprocessed to form a multi-dimensional feature vector, and the multi-dimensional feature vector is taken as an input of a pre-trained neural network model to output a state of charge value of the target single body.

[0014] The application has the advantages that the test signal application and impedance response acquisition are completed while the battery pack maintains normal charging and discharging operation, the complementary signal output by the compensation single body is matched to dynamically offset the power fluctuation of the target single body caused by the test, the test process does not affect the overall stability of the system, and the problems of low operation and maintenance efficiency and difficulty in reconnection and balancing caused by offline test are avoided, and secondly, the linear, step, sinusoidal or pulse signal is selected according to the real-time temperature, cumulative cycle number and historical operation data of the target single body, the adaptive matching design of the test signal greatly improves the effectiveness of the response data, and reduces the risk of battery damage. In addition, the response result and corresponding voltage change amount, current change amount, temperature gradient, vibration frequency, and ambient temperature are preprocessed to form a multi-dimensional feature vector, and the multi-dimensional feature vector is taken as an input of a pre-trained neural network model to output a state of charge value of the target single body. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 FIG. 1 is a flowchart of a battery state of charge test method according to an embodiment of the application. DETAILED DESCRIPTION

[0016] The technical solutions of the application are further described in detail below in combination with the accompanying drawings and specific embodiments. Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which the application belongs. The terms used in the specification of the application are only for the purpose of describing specific embodiments and are not intended to limit the application. In the following description, the expression "some embodiments" describes a subset of all possible embodiments, but it should be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0017] In the following description, a large number of specific details are given in order to provide a more thorough understanding of the application. However, it is obvious to those skilled in the art that the application can be implemented without one or more of these details. In other cases, some technical features known in the art are not described in order to avoid obscuring the application.

[0018] It is to be understood that the application can assume various alternative embodiments, and should not be limited to the examples described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the application to those skilled in the art. Also, the terminology used here is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising", when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0019] It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising", when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0020] For a thorough understanding of the application, reference will be made to the following detailed description, in which reference will be made to the drawings. Other embodiments of the application will be described, but the application is not limited to these embodiments. When considered in light of the disclosure provided herein, embodiments of the present application encompass other implementations beyond the specific examples disclosed.

[0021] The energy storage battery pack in the present application is composed of a plurality of battery nodes in series, each battery node containing one single battery, all nodes are interconnected through a bus, and the bus is provided with battery pack terminals at both ends for connecting external loads or power supply equipment to realize power input and output. The single battery serves as an energy storage and output unit, and its voltage and current can be independently regulated by a single controller.

[0022] The energy storage battery pack in the present application also contains a single controller and a battery pack controller. The single controller is the core execution unit of single control and testing, and each battery node is configured with one, which is electrically coupled with all single batteries in the node and is connected in series with other single controllers through a bus. The battery pack controller is the central control unit of the entire battery pack, which is communicatively coupled with all single controllers.

[0023] The data acquisition in the energy storage battery pack in the present application is realized by a sensor network, which includes battery pack level sensors and single cell level sensors. The single cell level sensors mainly include voltage probes (electrically coupled to each single cell terminal), current sensors (such as Hall effect sensors, current shunts), which collect real-time voltage and current data of the single cell, providing a basis for impedance calculation; the battery pack level sensors include thermocouples (thermally coupled to the single cell or node), pressure sensors, gas sensors, ultrasonic sensors, etc., which monitor the single cell temperature, battery pack internal pressure, electrolyte leakage, etc., and the temperature data can be used for correlation impedance analysis. All sensors are connected to the single cell controller and the battery pack controller through wired or wireless means, ensuring real-time data transmission.

[0024] Please refer to the attached Figure 1 The present application provides a kind of energy storage battery pack state-of-charge test method, test mode includes the following steps: S1, based on the real-time temperature of target single cell, cumulative cycle number and historical operation data, determine the type of single cell test signal; S2, in the case where the energy storage battery pack keeps running, the corresponding single cell battery is applied to the corresponding single cell by any single cell controller to carry out cycle life test, and the response result based on the single cell test signal is obtained, and the power fluctuation of the target single cell due to test is offset simultaneously; S3, obtain the voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature of the target single cell corresponding to the single response result; S4, combine the response result and the voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature of the target single cell corresponding to the single response result with the pre-trained neural network model to calculate the state-of-charge of the target single cell.

[0025] The energy storage battery pack state of charge testing method disclosed by the embodiments of the present application can realize accurate measurement of the target single cell state of charge while avoiding interference with the overall operation of the battery pack during the testing process on the premise that the battery pack maintains normal charging and discharging operation. Specifically, during the testing, the real-time temperature, cumulative cycle number and historical operation data of the target single cell are collected to determine the adapted single cell testing signal type from three dimensions of the current thermal state, aging degree and past operation stability of the battery, so as to ensure that the testing signal can effectively stimulate the electrochemical response of the single cell and will not cause damage to the battery or data distortion due to signal mismatch. After the testing signal is determined, the target single cell is applied with the testing signal through the single cell controller corresponding to the target single cell, and a complementary signal is output by other single cells in a stable state to dynamically offset the power fluctuation of the target single cell caused by the testing, so as to ensure the stability of the overall voltage and power output of the battery pack and obtain the impedance response result of the target single cell under the action of the testing signal. In order to improve the accuracy of the state of charge calculation, multi-dimensional state data corresponding to the response result are further collected, including the voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature of the target single cell, to form a feature set that comprehensively reflects the electrochemical characteristics and external working conditions of the single cell. The impedance response result and the above multi-dimensional state data are integrated and processed, and input into a neural network model that is trained based on the full working condition data in advance, so as to accurately output the state of charge of the target single cell by using the self-learning and mapping capability of the model for the complex nonlinear relationship between multiple physical quantities.

[0026] In some embodiments, the single cell testing signal type is determined based on the real-time temperature, cumulative cycle number and historical operation data of the target single cell, specifically including: collecting the real-time temperature of the target single cell, determining the preliminary screening signal type, adjusting the amplitude and frequency parameters of the preliminary screening signal based on the aging degree reflected by the cumulative cycle number; and optimizing the adjusted signal in combination with the power stability characteristics embodied in the historical operation data to finally determine the testing signal type adapted to the current working condition of the target single cell, wherein the testing signal type includes linear signal, step signal, sinusoidal signal and pulse signal.

[0027] For example, if the temperature is ≤0℃, a low-amplitude pulse signal or a sinusoidal signal is preliminarily screened; if 0℃<temperature<45℃, a step signal or a linear signal is preliminarily screened; and if the temperature is ≥45℃, a low-frequency sinusoidal signal is preliminarily screened. In combination with the cumulative cycle number: if the cycle number is ≤500 times, the signal type preliminarily screened is retained; if 500 times<cycle number<1500 times, the amplitude of the preliminarily screened signal is reduced by 10%-20%; and if the cycle number is ≥1500 times, a sinusoidal signal is preferentially selected. Finally, the power fluctuation amplitude of the last three charging and discharging in the historical operation data is referred to: if the fluctuation amplitude is ≤3%, the above screening result is adopted; and if the fluctuation amplitude is >3%, a pulse signal is replaced, and the pulse width is shortened to 1 / 2 of the original scheme, so as to determine the final single cell testing signal type.

[0028] Further, the real-time temperature of the target monomer is closely related to the electrochemical activity and temperature of the battery. For example, in a low-temperature environment, the internal resistance of the battery increases, and if a high-amplitude test signal is applied, it may exacerbate the polarization phenomenon. In a high-temperature environment, the reaction is intense, and a high-frequency signal can easily cause additional heat generation. Based on the interval in which the real-time temperature is located (such as low temperature, normal temperature, and high temperature), the type of signal waveform that is suitable is preliminarily screened to ensure that the signal will not cause internal damage to the battery or distortion of the test data due to temperature characteristics; The cumulative cycle number directly reflects the aging degree of the battery. New batteries with fewer cycles have high structural stability and strong resistance to signals. However, aged batteries with many cycles may have problems such as electrode material falling off and electrolyte degradation, and are more sensitive to signal amplitude and frequency. Therefore, according to the aging level corresponding to the cumulative cycle number, the signal parameters (such as amplitude and frequency) that are preliminarily screened are adaptively adjusted. For example, the signal amplitude is appropriately reduced for batteries with high cycle numbers to avoid excessive stimulation of the decaying electrochemical system. For historical operation data, the power fluctuation rules and voltage response characteristics in the historical operation data are extracted. If the historical data shows that the monomer has large power fluctuations during past charging and discharging processes, it indicates that its state stability is weak, and a signal type with less disturbance (such as a pulse signal with a shortened pulse width) needs to be selected. If the historical response characteristics show that the response to step signals is more significant, then step signal types can be preferentially retained.

[0029] Further, the core feature of the linear signal is that the current or voltage of the monomer battery changes linearly at a fixed rate. For example, the current increases linearly from 0A at a rate of 0.33A / s to 10A (for 30 seconds), or the voltage increases linearly from 3.9V at a rate of 0.003V / s to 4.0V (for 33 seconds). The characteristic of the step signal is that the monomer current or voltage suddenly changes to the target value in a very short time, for example, in less than 0.1 seconds, and then remains stable for a period of time. For example, the current suddenly increases from 25A to 40A and then remains for 10 seconds, or the voltage suddenly decreases from 4.0V to 3.8V and then remains for 20 seconds.

[0030] The current or voltage of the sinusoidal signal changes periodically according to the sinusoidal wave rule over time. For example, a sinusoidal current with a frequency of 1kHz and an amplitude of ±10A, or a sinusoidal voltage with a frequency of 500Hz and an amplitude of ±0.2V. The frequency range covers 1Hz to 1MHz. The pulse signal is characterized by a single or a few short-time current / voltage spikes, without continuous periodicity, such as a single pulse current with an amplitude of 20 A, a width of 0.5 seconds, and a rise time of <0.05 seconds, or a double pulse voltage with an amplitude of 0.5 V and a width of 1 second. The pulse width is usually controlled between 0.5 seconds and 1 minute to avoid long-time high-amplitude pulses causing local overheating of the monomer, and the rise / fall time is extremely short to quickly excite the transient impedance response of the monomer.

[0031] In some embodiments, the response result based on the monomer test signal is obtained, specifically including: collecting the electrical data of the target monomer under the action of the test signal in real time, and calculating the impedance value of the target monomer in combination with the type of the test signal as the response result based on the monomer test signal.

[0032] Specifically, for a linear signal, the equivalent impedance is directly derived by linear fitting of ΔV / ΔI using the total amount of voltage change and the total amount of current change during the signal duration; for a step signal, the ratio of the voltage steady value after the step to the current step amplitude is analyzed, and the dynamic response curve in the voltage transition process is combined to separate the static impedance and dynamic impedance components; for a sinusoidal signal, the voltage and current signals in the time domain are converted into frequency domain signals by means of Fourier transform, and the impedance modulus and phase angle at different frequencies are extracted to realize accurate analysis of multi-band impedance; for a pulse signal, if it is a short pulse, the impedance is calculated by extracting the ratio of the voltage change amount at the peak value time to the current pulse amplitude, and if it is a long pulse, the impedance is calculated by Fourier decomposition of the response signal.

[0033] In some embodiments, the power fluctuation of the target monomer caused by the test is synchronously offset, specifically including: specifying other monomers, outputting a complementary signal corresponding to the test signal of the target monomer through the monomer controller of the other monomers, the power of the complementary signal being equal in magnitude and opposite in direction to the instantaneous power fluctuation value, so as to offset the power fluctuation of the target monomer caused by the test.

[0034] Further, the complementary signal corresponding to the test signal of the target monomer is output through the monomer controller, specifically including: selecting one or more other monomers in a stable state as compensation monomers according to the current power of the target monomer and the power change amount of the test signal; calculating the power of the complementary signal to be output by the compensation monomers based on the instantaneous power fluctuation value caused by the test signal of the target monomer; and outputting the complementary signal in real time through the monomer controller of the compensation monomers to dynamically offset the test interference of the target monomer, so as to ensure that the overall voltage and power output fluctuation of the battery pack is controlled within a preset range.

[0035] Specifically, for linear signals, the complementary signals of the compensation monomers need to change in the opposite direction at the same slope to ensure that the power change rates of the two match; for step signals, the complementary signals need to perform a reverse step at the same time, and the step amplitude is equal to the step amplitude of the target monomer test signal; for sinusoidal signals, the complementary signals need to maintain a 180° phase difference with the target monomer test signal, and the amplitude is distributed in proportion to the number of compensation monomers; for pulse signals, short pulses can be temporarily offset by bus capacitance, and long pulses are offset by the compensation monomers outputting pulse signals of the same width and in the opposite direction, with the amplitude being consistent with the target monomer pulse amplitude, so as to realize the offset of the power fluctuation of the target monomer caused by the test.

[0036] In some embodiments, other monomers are specified, specifically including: based on the real-time measurable state parameters of each monomer in the battery pack, wherein the measurable state parameters include real-time temperature, state of health SOH, recent voltage fluctuation amplitude, current working current and rated power, a plurality of complementary monomers are selected by setting a screening threshold, for example, complementary monomers with selected temperature in the normal working interval of the battery, SOH≥80%, voltage fluctuation amplitude≤5% and current working current lower than 60% of its rated power, among the plurality of complementary monomers selected, complementary monomers with the same branch or adjacent position as the target monomer are selected as the final specified monomers, by taking advantage of the consistency of monomers with the same branch or adjacent position in signal transmission path and response timing, the transmission delay of the compensation signal is shortened, the synchronization of the complementary signal and the target monomer test signal is ensured, and the accuracy of power fluctuation offset is improved.

[0037] In some embodiments, the response results and corresponding voltage change, current change, temperature gradient, vibration frequency and ambient temperature are preprocessed to form a multi-dimensional feature vector, and the multi-dimensional feature vector is taken as the input of the pre-trained neural network model to output the state of charge value of the target monomer.

[0038] Specifically, the neural network model in the application selects a hybrid neural network model combining convolutional neural network and long short-term memory network. The model has the local feature extraction capability of CNN and the time sequence dependence capturing advantage of LSTM, and is suitable for the nonlinear and time sequence characteristics of the battery state parameters. The model structure includes: an input layer that receives a preprocessed multi-dimensional feature vector, the number of neurons is consistent with the dimension of the feature vector; two convolutional layers that use 3x1 size convolution kernels to extract local correlation features between parameters through a sliding window, and the activation function is ReLU to enhance the nonlinear fitting capability of the model; a pooling layer that uses maximum pooling to reduce the dimension of the convolutional features, retaining key information while reducing computational complexity; an LSTM layer that sets 64 hidden units to capture the dynamic dependence of the feature vector in the time sequence and solve the information decay problem of long-term data; two fully connected layers that fuse the local features after pooling and the time sequence features output by the LSTM, and map to a high-dimensional feature space through a weight matrix; and an output layer that uses a Sigmoid activation function to output continuous values in the 0-1 interval corresponding to the state of charge of the target monomer During the training process, monomer test data under different environmental temperatures (-20℃ to 60℃), different cumulative cycle numbers (0 to 2000 times), and different SOC intervals (0% to 100%) are collected, covering response results and multi-dimensional state parameters corresponding to four signal types of linear, step, sine, and pulse, and the real SOC value is used as a label. The data set is divided into a training set, a validation set, and a test set in a ratio of 7:2:1. During training, the mean square error (MSE) is used as the loss function, the Adam optimizer is used to dynamically adjust the learning rate, and the back propagation algorithm is used to update the weights of each layer of the model. During the training process, an early stopping mechanism is set to stop training when the validation set loss value does not decrease for 10 consecutive rounds to avoid model overfitting. Finally, the generalization ability of the model is verified by the test set to ensure that the prediction accuracy is maintained in unseen working condition samples, and the construction of the pre-trained model is completed.

[0039] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. The protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of testing the state of charge of an energy storage battery pack, characterized by, The test method comprises the following steps: Based on the real-time temperature, cumulative cycle number and historical operation data of the target monomer, the monomer test signal type is determined; While the energy storage battery pack is kept running, the monomer test signal is applied to the corresponding monomer battery through any monomer controller to perform cycle life test, and the response result based on the monomer test signal is obtained, and the power fluctuation of the target monomer caused by the test is synchronously offset; The voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature of the target monomer corresponding to the single response result are obtained; The response result and the voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature of the target monomer corresponding to the single response result are combined with the pre-trained neural network model to calculate the state of charge of the target monomer.

2. The method of claim 1, wherein, Based on the real-time temperature, cumulative cycle number and historical operation data of the target monomer, the monomer test signal type is determined, which specifically comprises: based on the real-time temperature of the target monomer, the preliminary screening signal type is determined, the amplitude and frequency parameters of the preliminary screening signal are adjusted based on the aging degree reflected by the cumulative cycle number; the adjusted signal is optimized in combination with the power stability characteristics embodied in the historical operation data, and finally the test signal type suitable for the current working condition of the target monomer is determined.

3. The method of claim 1, wherein, The test signal type includes linear signal, step signal, sinusoidal signal and pulse signal.

4. The method of claim 1, wherein, The response result based on the monomer test signal is obtained, specifically including: real-time acquisition of the electrical data of the target monomer under the action of the test signal, and combination of the test signal type to calculate the impedance value of the target monomer as the response result based on the monomer test signal.

5. The method of claim 1, wherein, The power fluctuation of the target monomer caused by the test is synchronously offset, specifically including: specifying other monomers, outputting complementary signals corresponding to the test signal of the target monomer through the monomer controllers of the other monomers to offset the power fluctuation of the target monomer caused by the test.

6. A method of testing the state of charge of an energy storage battery pack according to claim 5, wherein, Specifying other monomers specifically includes: based on the real-time measurable state parameters of each monomer in the battery pack, screening a plurality of complementary monomers, and in the screened plurality of complementary monomers, selecting the complementary monomers in the same branch or adjacent position as the target monomer as the final specified monomers.

7. A method of testing the state of charge of an energy storage battery pack according to claim 6, wherein, Outputting the complementary signal corresponding to the test signal of the target monomer through the monomer controller specifically includes: selecting one or more other monomers with stable state as compensation monomers according to the current power of the target monomer and the power variation of the test signal; calculating the complementary signal power to be output by the compensation monomers based on the instantaneous power fluctuation value caused by the test signal of the target monomer; and outputting the complementary signal in real time through the monomer controller of the compensation monomer to dynamically offset the test interference of the target monomer, so as to ensure that the overall voltage and power output fluctuation of the battery pack is controlled within a preset range.

8. A method of testing the state of charge of an energy storage battery pack according to claim 7, wherein, The complementary signal power is equal in size and opposite in direction to the instantaneous power fluctuation value.

9. The method of claim 1, wherein, The response result and the corresponding voltage variation, current variation, temperature gradient, vibration frequency and ambient temperature are preprocessed to form a multi-dimensional feature vector, and the multi-dimensional feature vector is taken as the input of the pre-trained neural network model to output the state of charge value of the target monomer.

Images