Battery health state data acquisition method and device, equipment, medium and product

Abstract

The embodiment of the invention provides a battery health state data acquisition method and device, equipment, a medium and a product. The method comprises the steps that charging voltage, charging current and temperature data of each single battery in an energy storage cabin are collected in real time through preset sampling frequency, important analysis data can be obtained without relying on complete charging and discharging testing, and the problem that the period is too long due to the fact that a complete charging and discharging process is waited in the related technology is solved from the data collection source. An incremental capacity curve is generated based on charging voltage and current data, health feature data related to capacity attenuation are extracted, then a feature vector sequence is formed in combination with temperature data and input into a pre-trained neural network model, and battery health state data of each single battery is output through powerful feature learning of the neural network model. Therefore, the output accuracy of the battery health state data is effectively improved.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a method, apparatus, device, medium and product for acquiring battery health status data. Background Technology

[0002] As the global energy landscape shifts towards cleaner and lower-carbon energy, the large-scale application of new energy sources places higher demands on the stability and flexibility of the power grid. Energy storage modules, as key buffering and regulating equipment, store and release electrical energy through battery charging and discharging; their reliability directly impacts grid security and economic efficiency.

[0003] In related technologies, the State of Health (SOH) data of batteries in energy storage compartments is usually estimated based on data from complete charge-discharge tests.

[0004] However, the process of obtaining battery health status data in the energy storage compartment is lengthy and the data obtained is not accurate enough. Summary of the Invention

[0005] This application provides a method, apparatus, device, medium, and product for acquiring battery health status data, which aims to reduce the cycle of acquiring battery health status data in an energy storage compartment and improve the accuracy of the acquired data.

[0006] In a first aspect, embodiments of this application provide a method for obtaining battery health status data, including:

[0007] Based on a preset sampling frequency, charging voltage data, charging current data, and temperature data of each individual battery cell in the energy storage compartment are collected at different times.

[0008] Based on the charging voltage and charging current data of each individual cell at different times, the incremental capacity curve of each individual cell is generated.

[0009] Based on the incremental capacity curve of each individual cell, extract the health characteristic data related to the capacity decay of each individual cell.

[0010] Based on the health characteristic data related to the capacity decay of each individual battery cell and the corresponding temperature data, the feature vector sequence of each individual battery cell is obtained.

[0011] The feature vector sequence of each individual battery cell is input into a pre-trained neural network model, which outputs the health status data of each individual battery cell.

[0012] In one possible implementation, based on the charging voltage and charging current data of each individual cell at different times, an incremental capacity curve for each individual cell is generated, including:

[0013] Gaussian smoothing was performed on the charging voltage and charging current data of each individual cell at different times to obtain the processed charging voltage and charging current data of each individual cell at different times.

[0014] The charging voltage and charging current data of each individual battery at different times after processing are input into a preset incremental capacity algorithm to generate the incremental capacity curve of each individual battery.

[0015] In one possible implementation, based on the incremental capacity curve of each individual cell, health characteristic data related to capacity decay of each individual cell are extracted, including:

[0016] For any single cell's incremental capacity curve, identify the maximum value within the preset characteristic range of the incremental capacity curve.

[0017] The voltage range is expanded based on the maximum value and the preset extended value voltage to determine the expanded voltage range.

[0018] Within the extended voltage range, health characteristic data related to the capacity decay of individual cells are extracted.

[0019] In one possible implementation, based on the health characteristic data related to the capacity decay of each individual battery cell and the corresponding temperature data, a feature vector sequence for each individual battery cell is obtained, including:

[0020] The health characteristic data related to capacity decay of each individual cell at the same time are timestamped with the temperature data to generate a multi-dimensional time-series data set for each individual cell.

[0021] The multi-dimensional time-series data of each individual cell are processed to obtain the feature vector sequence of each individual cell.

[0022] In one possible implementation, the pre-trained neural network model includes: a temporal convolutional network algorithm, an attention algorithm, and a long short-term memory network algorithm;

[0023] The feature vector sequence of each individual battery cell is input into a pre-trained neural network model, which outputs the health status data of each individual battery cell, including:

[0024] The spatial features of each individual battery cell are learned by using a temporal convolutional network algorithm to obtain the learned feature vector sequence of each individual battery cell.

[0025] The feature vector sequence of each individual battery cell is weighted by the attention algorithm to obtain the weighted feature vector sequence of each individual battery cell.

[0026] The health status data of each individual battery cell is obtained by calculating the weighted feature vector sequence of each individual battery cell using the Long Short-Term Memory Network algorithm.

[0027] In one possible implementation, after outputting the health status data of each individual battery cell, the method further includes:

[0028] Based on the health status data of each individual battery cell, determine the maximum difference in health status data;

[0029] Obtain at least two preset difference threshold intervals;

[0030] Based on the maximum difference and at least two preset difference threshold intervals, determine the target difference threshold interval to which the maximum difference belongs;

[0031] Output early warning information corresponding to the difference threshold range from the target.

[0032] Secondly, embodiments of this application provide a device for acquiring battery health status data, comprising:

[0033] The data acquisition module is used to collect charging voltage data, charging current data, and temperature data of each individual battery in the energy storage compartment at different times, based on a preset sampling frequency.

[0034] The generation module is used to generate the incremental capacity curve of each individual battery cell based on the charging voltage and charging current data of each individual battery cell at different times.

[0035] The extraction module is used to extract health characteristic data related to the capacity decay of each individual battery cell based on the incremental capacity curve of each individual battery cell.

[0036] The processing module is used to obtain the feature vector sequence of each individual battery cell based on the health characteristic data related to the capacity decay of each individual battery cell and the corresponding temperature data.

[0037] The processing module is also used to input the feature vector sequence of each individual battery cell into a pre-trained neural network model and output the health status data of each individual battery cell.

[0038] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0039] The memory stores instructions that the computer executes;

[0040] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0041] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0042] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0043] The battery health status data acquisition method, apparatus, device, medium, and product provided in this application collect charging voltage, charging current, and temperature data of each individual battery cell in the energy storage compartment in real time through a preset sampling frequency. This allows for the acquisition of crucial analytical data without relying on a complete charge-discharge test, thus avoiding the excessively long cycle time caused by waiting for a complete charge-discharge process in related technologies. Incremental capacity curves are generated based on the charging voltage and current data, and health feature data related to capacity decay are extracted. These are then combined with temperature data to form a feature vector sequence, which is input into a pre-trained neural network model. Through the powerful feature learning capabilities of the neural network model, the battery health status data of each individual battery cell is output, thereby effectively improving the accuracy of the battery health status data output. Attached Figure Description

[0044] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0045] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application;

[0046] Figure 2 A flowchart illustrating a method for acquiring battery health status data provided in an embodiment of this application;

[0047] Figure 3 A flowchart illustrating a method for obtaining health status data of individual cells provided in an embodiment of this application;

[0048] Figure 4 This is a schematic diagram of health status data provided in an embodiment of this application;

[0049] Figure 5 A flowchart illustrating a method for outputting early warning information provided in an embodiment of this application;

[0050] Figure 6 A schematic diagram of the structure of the battery health status data acquisition device provided in the embodiments of this application;

[0051] Figure 7This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0052] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0053] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0054] As the global energy structure shifts towards cleaner and lower-carbon energy, the large-scale application of new energy power generation places higher demands on the stability and flexibility of the power grid. Energy storage modules, as key equipment for energy buffering and power regulation in new energy systems, store and release electrical energy through the charging and discharging of battery packs. Their operational reliability directly affects the safety and economy of the power grid.

[0055] However, as the core component of an energy storage module, the health of batteries gradually deteriorates with usage time, charge / discharge cycles, and ambient temperature, leading to decreased capacity, increased internal resistance, and even the risk of thermal runaway. Therefore, rapidly and accurately acquiring the SOH data of multiple batteries connected in series or parallel within the energy storage module, and establishing an early warning mechanism based on this data, is a core requirement for ensuring the safe operation of the energy storage module.

[0056] In related technologies, when obtaining the SOH data of batteries in an energy storage compartment, it is usually estimated based on data from a complete charge-discharge test.

[0057] However, the above methods rely on complete charge and discharge data, requiring data collection when the battery is fully charged or fully discharged, necessitating system shutdown and resulting in a long acquisition cycle for battery health status data in the energy storage compartment. Furthermore, in scenarios involving non-full charge and discharge, capacity decay characteristics are difficult to extract accurately, leading to significant errors in the acquired SOH data.

[0058] Therefore, addressing the issue of long SOH data acquisition cycles caused by the need for a complete charge-discharge cycle during charging in related technologies, this application employs a preset sampling frequency to collect voltage, current, and temperature data in real time during the charging process, breaking the cycle limitation through non-stop data acquisition. To address the difficulty in accurately capturing capacity decay characteristics in non-complete charge-discharge scenarios, leading to large SOH errors, this application generates incremental capacity curves from the collected voltage and current data and extracts health feature data. It then integrates temperature data to construct a feature vector sequence and finally utilizes the feature learning capabilities of a trained neural network to achieve accurate calculations.

[0059] To facilitate understanding of the method in this application, an exemplary application scenario is provided below. Please refer to [link / reference]. Figure 1 , Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of the present application. In this application scenario, electronic device 01 and energy storage compartment 02 may be included.

[0060] Electronic device 01 establishes a communication connection with energy storage compartment 02. When the individual batteries in energy storage compartment 02 are charging, electronic device 01 collects the charging voltage, current, and temperature data of each individual battery in energy storage compartment 02 in real time according to a preset sampling frequency. Based on the voltage and current data, electronic device 01 generates incremental capacity curves for each individual battery, extracts health feature data related to capacity decay from the incremental capacity curves, constructs a feature vector sequence by combining it with temperature data, and then performs calculations through a built-in pre-trained neural network model to finally output the health status data of each individual battery, thereby achieving efficient monitoring of the batteries in energy storage compartment 02.

[0061] It is understood that the above examples are for illustrative purposes only and do not limit this application. The specific details can be determined based on the actual application situation.

[0062] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0063] Please see Figure 2 , Figure 2This is a flowchart illustrating a method for acquiring battery health status data according to an embodiment of this application. The execution entity of this method can be a battery health status data acquisition device. This battery health status data acquisition device can be implemented through a computer program, or through a medium storing the relevant computer program, such as a USB flash drive and / or optical disc, or through a physical device integrating or installing the relevant computer program, such as a chip or electronic device. The electronic device can be a server, server cluster, smart terminal, etc. The method may include:

[0064] S201. Based on the preset sampling frequency, collect charging voltage data, charging current data, and temperature data of each individual battery in the energy storage compartment at different times.

[0065] An energy storage compartment is a device used for energy storage and regulation in a new energy system. It is usually composed of multiple individual batteries connected in series or in parallel.

[0066] Voltage sensors, current sensors, and temperature sensors are deployed on the positive and negative terminals and surface of each individual battery cell within the energy storage chamber. All sensors have built-in preset control logic and are interconnected. Based on the battery type and monitoring accuracy requirements, a fixed sampling frequency is set through the sensor's preset control logic, for example, 1Hz, which means sampling once per second.

[0067] Once the energy storage battery enters the charging state, the sensors automatically and synchronously start collecting data at a preset frequency, capturing the charging voltage, current, and cell temperature of each individual battery in real time, and assigning a unique timestamp to each set of data.

[0068] S202. Based on the charging voltage and charging current data of each individual cell at different times, generate the incremental capacity curve of each individual cell.

[0069] The incremental capacity (IC) curve is a curve obtained by differentiating the charging voltage with respect to the current, reflecting the electrochemical changes inside the battery.

[0070] Gaussian smoothing was applied to the charging voltage and charging current data of each individual cell at different times to obtain the processed charging voltage and charging current data of each individual cell at different times.

[0071] During the charging process, the cell voltage of a single battery cell does not increase in a strictly monotonically; there will be outliers. Therefore, Gaussian smoothing is required to filter the data.

[0072] Gaussian smoothing is a filtering algorithm based on normal distribution weights. In this embodiment, it is used to eliminate noise in the charging voltage and charging current data to ensure the accuracy of the incremental capacity curve.

[0073] The charging voltage and charging current data of each individual battery at different times after processing are input into a preset incremental capacity algorithm to generate the incremental capacity curve of each individual battery.

[0074] By performing incremental capacity analysis, flat voltage data can be transformed into peak data in the incremental capacity curve.

[0075] Optionally, the preset incremental capacity algorithm can be as shown in the following formula (1):

[0076]

[0077] in, Indicates incremental capacity; It represents the differential of voltage with respect to current; Indicates time b; Indicates time a; This represents the current value during the time interval from time b to time a. This represents the voltage value at time b; This represents the voltage value at time a.

[0078] S203. Based on the incremental capacity curve of each individual cell, extract the health characteristic data related to the capacity decay of each individual cell.

[0079] For any single cell's incremental capacity curve, the maximum value within the preset characteristic range of the incremental capacity curve is identified.

[0080] Through research and analysis of a large amount of battery cell data, it was found that the second peak interval of the incremental capacity curve changes with the battery capacity decay, such as the change in peak height and the voltage corresponding to the peak. Therefore, the second peak interval can be selected as the preset characteristic interval.

[0081] Gaussian smoothing was applied to the incremental capacity curve, and local maxima were identified within the second peak interval.

[0082] The range is expanded based on the maximum value and the preset extended voltage value to determine the expanded voltage range. Within the expanded voltage range, health characteristic data related to the capacity decay of individual cells are extracted.

[0083] The voltage range is expanded with the maximum point as the center, for example, the voltage range of ±0.05V is expanded with the local maximum point as the center, and this range is used as the feature extraction window to extract health feature data related to the capacity decay of each individual battery cell.

[0084] Health characteristic data include, but are not limited to: peak height, peak area, peak voltage, and sub-peak energy.

[0085] S204. Based on the health characteristic data related to the capacity decay of each individual cell and the corresponding temperature data, obtain the feature vector sequence of each individual cell.

[0086] The health characteristic data related to capacity decay of each individual battery at the same time are timestamped with the temperature data to generate a multi-dimensional time series data set for each individual battery. The multi-dimensional time series data set of each individual battery is then processed to obtain the feature vector sequence of each individual battery.

[0087] The health characteristic data and temperature data of each individual battery cell are aggregated, ensuring that each data point carries a collection timestamp. Based on the timestamp, health characteristic data and temperature data at the same time are matched for each individual battery cell, combined into a multi-dimensional data unit of "characteristic + temperature", and arranged in chronological order to generate a multi-dimensional time-series data group for each individual battery cell.

[0088] By employing standardization or normalization methods, health characteristic data of different dimensions in a multi-dimensional time series dataset are converted to eliminate numerical scale differences, thereby obtaining the feature vector sequence of the single battery cell.

[0089] S205. Input the feature vector sequence of each individual battery cell into the pre-trained neural network model and output the health status data of each individual battery cell.

[0090] In this embodiment, a neural network is pre-trained. The network model may include Temporal Convolutional Network (TCN), Attention Mechanism (AM), and Long Short-Term Memory (LSTM) algorithms.

[0091] The feature vector sequence of each individual battery cell is input into the neural network model mentioned above, and the health status data of each individual battery cell is finally output.

[0092] In the above embodiments of this application, the charging voltage, charging current, and temperature data of each individual battery cell in the energy storage chamber are collected in real time by a preset sampling frequency. This allows for the acquisition of crucial analytical data without relying on a complete charge-discharge test, thus avoiding the excessively long cycle time caused by waiting for a complete charge-discharge process in related technologies. Incremental capacity curves are generated based on the charging voltage and current data, and health feature data related to capacity decay are extracted. These are then combined with temperature data to form a feature vector sequence, which is input into a pre-trained neural network model. Through the powerful feature learning capabilities of the neural network model, the battery health status data of each individual battery cell is output, thereby effectively improving the accuracy of the battery health status data output.

[0093] Furthermore, based on the above embodiments, the following embodiments illustrate the process of inputting the feature vector sequence of each individual battery cell into a pre-trained neural network model to output the health status data of each individual battery cell.

[0094] Please see Figure 3 , Figure 3 This application provides a flowchart illustrating a method for obtaining health status data of individual battery cells, which may include the following steps:

[0095] S301. Using a temporal convolutional network algorithm, spatial feature learning is performed on the feature vector sequence of each individual battery cell to obtain the learned feature vector sequence of each individual battery cell.

[0096] In this embodiment, the temporal convolutional network algorithm is an improved model that introduces causal convolution and dilated convolution on the basis of convolutional neural networks. It has the advantages of efficient parallel computing capabilities and long sequence data processing, and is very suitable for the analysis of large-scale battery data.

[0097] In this case, the causal convolution of TCN ensures that the temporal correlation of the feature vector sequence of a single battery cell is not disturbed by future data by constraining the output at time t to depend only on the input at the current and previous times, which conforms to the temporal evolution law of battery operation data.

[0098] Dilated convolution, also known as dilated convolution or attenuated convolution, uses a hole structure to increase the sensitivity during the convolution process, allowing the convolution kernel to cover a wider range of temporal information in the sequence and improving the ability to capture local correlation patterns in individual battery data.

[0099] Its feature transformation process follows the formula (2):

[0100]

[0101] in, This represents the sequence of feature vectors learned over time step t. The weights of the convolution kernel are represented by K; K represents the kernel size. d represents the size of the k-th convolutional kernel; d represents the size of the inflation factor.

[0102] Meanwhile, the residual connection structure of TCN effectively alleviates the gradient vanishing and information loss problems of deep networks, retains the key information in the original feature vectors, and enables the model to learn the spatial features contained in the feature vector sequence while efficiently processing large-scale battery data by leveraging the advantages of parallel computing, and finally outputs a feature vector sequence enhanced with spatial features.

[0103] S302. Using the attention algorithm, the feature vector sequences learned by each individual battery cell are weighted to obtain the weighted feature vector sequences of each individual battery cell.

[0104] This step leverages the global dependency capture capability of the self-attention mechanism to overcome the limitations of TCN in mining global associations in long sequences. First, each element in the feature vector sequence output by S301 is converted into three vector representations: query (Q), key (K), and value (V). Then, the dot product of the query and all keys is calculated and divided by... ,in, To avoid excessively large dot product results, the weights of each value are obtained after normalization using the softmax probability function, with the key vector dimension as the reference dimension.

[0105] This weight can directly reflect the importance of features at different times in the sequence to the current battery health status assessment. For example, the higher the correlation between the capacity decay feature of a battery at a certain time and the historical key decay node features, the greater the corresponding Value weight.

[0106] Finally, based on the following formula (3):

[0107]

[0108] in, This indicates the dimension of the key vector.

[0109] The weighted Value is summed according to formula (3) to strengthen the core features that play a decisive role in battery health assessment in the feature vector sequence, while weakening the secondary or noisy features, thereby generating a weighted feature vector sequence that focuses on key information, laying the foundation for subsequent accurate calculations.

[0110] S303. Using a long short-term memory network algorithm, the weighted feature vector sequence of each individual battery cell is processed to obtain the health status data corresponding to each individual battery cell.

[0111] This step leverages the advantages of LSTM in handling long sequence dependencies to complete the final output. LSTM uses a collaborative control mechanism of input gate, forget gate, and output gate to dynamically filter information and model long-term features from the weighted feature vector sequence output by S302.

[0112] In this structure, the input gate filters out key health features from the weighted features, allowing them to enter the cell state. The forget gate selectively discards redundant or outdated features, such as temporary data reflecting normal battery fluctuations. The output gate generates a valid output based on the current cell state and updates the hidden state. This structure enables LSTM to accurately capture the long-term dynamic change patterns of the battery, i.e., the trend of capacity decay, contained in the weighted feature vector sequence. It fully explores the potential correlations in the data and reduces the gradient problem in long sequence processing of traditional recurrent networks. Through stepwise computation and state updates on the weighted feature vector sequence, LSTM transforms the abstract weighted feature vector sequence into specific battery health state data, achieving the final mapping from feature input to result output. Furthermore, leveraging its excellent performance in long sequence processing, it ensures that the health state data reflects the true state of the battery after long-term operation.

[0113] The final output health status data for each individual battery cell is as follows: Figure 4 As shown, Figure 4 This is a schematic diagram of health status data provided in an embodiment of this application. The y-axis represents the health status data of a single battery cell, which represents the percentage of battery health status; typically, 100% represents a brand new battery. The x-axis represents the number of battery cycles.

[0114] In the embodiments described above, spatial feature learning is performed on the feature vector sequence using a temporal convolutional network algorithm. This allows for targeted mining of spatial correlation features within the sequence, providing high-quality foundational data for subsequent analysis. Furthermore, an attention algorithm is used to weight the learned feature vector sequence, highlighting core features crucial for battery health assessment and reducing interference from irrelevant information. Finally, a long short-term memory network algorithm is used to process the weighted feature vector sequence, effectively handling the sequence data and outputting health status data. This enables the neural network model to fully utilize the effective information in the feature vector sequence, ultimately achieving accurate output of health status data.

[0115] This application can also output corresponding early warning information based on the health status data of each individual battery cell.

[0116] Please see Figure 5 , Figure 5 This application provides a flowchart illustrating a method for outputting early warning information, which may include the following steps:

[0117] S501. Determine the maximum difference in health status data based on the health status data of each individual battery cell.

[0118] Based on the type of energy storage compartment equalization system, determine the health status data of individual batteries.

[0119] If the energy storage compartment equalization system is a passive equalization system, then the minimum value of the health status data of all individual batteries is directly used as the reference benchmark.

[0120] If the energy storage compartment balancing system is an active balancing system, then the minimum health status data of a single battery cell is used as the basis, and the ratio of the active balancing capability to the rated capacity of the cell during a single charge is added to obtain the corrected minimum health status data of the single battery cell, which is then used as a reference value.

[0121] Then, the difference between the maximum and minimum values ​​of the health status data of all batteries in the cabin is calculated to obtain the maximum difference of the health status data. This difference can reflect the consistency level of the health status data of the batteries in the cabin.

[0122] S502. Obtain at least two preset difference threshold intervals.

[0123] In this embodiment, at least two difference threshold ranges are preset, such as a difference threshold range greater than 10% and less than 20% and a difference threshold range greater than 20%.

[0124] S503. Based on the maximum difference and at least two preset difference threshold intervals, determine the target difference threshold interval to which the maximum difference belongs.

[0125] The maximum difference obtained in S501 is matched with at least two preset difference threshold intervals. If the maximum difference is between 10% and 20%, the threshold interval corresponding to the first-level warning is matched.

[0126] If the maximum difference exceeds 20%, then the threshold range corresponding to the Level 2 warning is matched.

[0127] If the maximum difference does not exceed 10%, it is within the normal range and no warning needs to be triggered.

[0128] S504. Output the warning message corresponding to the difference threshold range between the target and the target.

[0129] When the maximum difference falls within the range of 10%-20%, a Level 1 warning message is output to remind maintenance personnel to pay attention to changes in battery consistency and prepare for monitoring.

[0130] When the range of the maximum difference is greater than 20%, a level 2 early warning message is output to alert maintenance personnel to take immediate intervention measures, such as battery replacement or equalization system debugging.

[0131] If it falls within the normal range, output a message indicating good consistency.

[0132] In the above embodiments of this application, by calculating the maximum difference in the health status data of all individual batteries, and combining it with at least two preset difference threshold intervals to determine the target interval and output corresponding early warning information, the accurate assessment and graded early warning of the consistency of the battery health status of the energy storage compartment can be achieved, thereby ensuring the safety and stability of the operation of the energy storage compartment.

[0133] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of the battery health status data acquisition device provided in the embodiments of this application, as shown below. Figure 6 As shown, the battery health status data acquisition device provided in this embodiment includes:

[0134] The acquisition module 601 is used to acquire charging voltage data, charging current data and temperature data of each individual battery in the energy storage compartment at different times based on a preset sampling frequency.

[0135] The generation module 602 is used to generate the incremental capacity curve of each individual battery cell based on the charging voltage data and charging current data of each individual battery cell at different times.

[0136] The extraction module 603 is used to extract health characteristic data related to the capacity decay of each individual cell based on the incremental capacity curve of each individual cell.

[0137] The processing module 604 is used to obtain the feature vector sequence of each individual battery cell based on the health characteristic data related to the capacity decay of each individual battery cell and the corresponding temperature data.

[0138] The processing module 604 is also used to input the feature vector sequence of each individual battery cell into a pre-trained neural network model and output the health status data of each individual battery cell.

[0139] In one possible implementation, the generation module 602 is specifically used for:

[0140] Gaussian smoothing was applied to the charging voltage and charging current data of each individual cell at different times to obtain the processed charging voltage and charging current data of each individual cell at different times.

[0141] The charging voltage and charging current data of each individual battery at different times after processing are input into a preset incremental capacity algorithm to generate the incremental capacity curve of each individual battery.

[0142] In one possible implementation, the extraction module 603 is specifically used for:

[0143] For any single cell's incremental capacity curve, the maximum value within the preset characteristic range of the incremental capacity curve is identified.

[0144] The voltage range is expanded based on the maximum value and the preset extended value voltage to determine the expanded voltage range.

[0145] Within the extended voltage range, health characteristic data related to the capacity decay of individual cells are extracted.

[0146] In one possible implementation, processing module 604 is specifically used for:

[0147] The health characteristic data related to capacity decay of each individual cell at the same time point are timestamped with the temperature data to generate a multi-dimensional time-series data set for each individual cell.

[0148] The multi-dimensional time-series data of each individual cell are processed to obtain the feature vector sequence of each individual cell.

[0149] In one possible implementation, the pre-trained neural network model includes temporal convolutional network algorithms, attention algorithms, and long short-term memory network algorithms. The processing module 604 is specifically used for:

[0150] By using a temporal convolutional network algorithm, spatial feature learning is performed on the feature vector sequence of each individual battery cell to obtain the learned feature vector sequence of each individual battery cell.

[0151] The attention algorithm is used to weight the feature vector sequences learned by each individual cell to obtain the weighted feature vector sequence of each individual cell.

[0152] The health status data of each individual battery cell is obtained by calculating the weighted feature vector sequence of each individual battery cell using the Long Short-Term Memory Network algorithm.

[0153] In one possible implementation, after outputting the health status data of each individual battery cell, the processing module 604 is specifically used for:

[0154] Based on the health status data of each individual battery cell, determine the maximum difference in health status data.

[0155] Obtain at least two preset difference threshold intervals.

[0156] Based on the maximum difference and at least two preset difference threshold intervals, determine the target difference threshold interval to which the maximum difference belongs.

[0157] Output early warning information corresponding to the difference threshold range from the target.

[0158] The battery health status data acquisition device provided in this embodiment can execute the battery health status data acquisition method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0159] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 7 As shown, the electronic device provided in this embodiment includes at least one processor 701 and a memory 702. Optionally, the device further includes a communication component 703. The processor 701, memory 702, and communication component 703 are connected via a bus 704.

[0160] In a specific implementation, at least one processor 701 executes computer execution instructions stored in memory 702, causing at least one processor 701 to perform the above-described method.

[0161] The specific implementation process of processor 701 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0162] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0163] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0164] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0165] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0166] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0167] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory, electrically erasable programmable read-only memory, erasable programmable read-only memory, programmable read-only memory, read-only memory, magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0168] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0169] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0170] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0171] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0172] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0173] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0174] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for acquiring battery health status data, characterized in that, include: Based on a preset sampling frequency, charging voltage data, charging current data, and temperature data of each individual battery cell in the energy storage compartment are collected at different times. Based on the charging voltage and charging current data of each individual battery at different times, an incremental capacity curve for each individual battery is generated. Based on the incremental capacity curve of each individual cell, extract the health characteristic data related to the capacity decay of each individual cell. Based on the health characteristic data related to the capacity decay of each individual battery and the corresponding temperature data, the feature vector sequence of each individual battery is obtained. The feature vector sequence of each individual battery cell is input into a pre-trained neural network model, which outputs the health status data of each individual battery cell.

2. The method according to claim 1, characterized in that, The step of generating the incremental capacity curve for each individual battery cell based on its charging voltage and charging current data at different times includes: Gaussian smoothing is performed on the charging voltage and charging current data of each individual battery at different times to obtain the processed charging voltage and charging current data of each individual battery at different times. The charging voltage and charging current data of each individual battery at different times after processing are input into a preset incremental capacity algorithm to generate the incremental capacity curve of each individual battery.

3. The method according to claim 1, characterized in that, The method involves extracting health characteristic data related to capacity decay of each individual cell based on its incremental capacity curve, including: For any single cell's incremental capacity curve, the maximum value within the preset characteristic range of the incremental capacity curve is identified. Based on the maximum value and the preset extended voltage value, the range is expanded to determine the expanded voltage range; Within the extended voltage range, health characteristic data related to the capacity decay of the individual battery cells are extracted.

4. The method according to claim 1, characterized in that, The step of obtaining the feature vector sequence of each individual battery cell based on the health characteristic data related to capacity decay and the corresponding temperature data includes: The health characteristic data related to capacity decay of each individual cell at the same time are timestamped with the temperature data to generate a multi-dimensional time-series data set for each individual cell. The multi-dimensional time-series data of each individual battery cell are processed to obtain the feature vector sequence of each individual battery cell.

5. The method according to claim 1, characterized in that, The pre-trained neural network model includes: temporal convolutional network algorithm, attention algorithm, and long short-term memory network algorithm; The step of inputting the feature vector sequence of each individual battery cell into a pre-trained neural network model and outputting the health status data of each individual battery cell includes: The temporal convolutional network algorithm is used to learn spatial features from the feature vector sequence of each individual battery cell to obtain the learned feature vector sequence of each individual battery cell. The attention algorithm is used to weight the feature vector sequence of each individual battery after learning, so as to obtain the weighted feature vector sequence of each individual battery. The long short-term memory network algorithm is used to calculate the weighted feature vector sequence of each individual battery cell to obtain the health status data of each individual battery cell.

6. The method according to any one of claims 1-5, characterized in that, After outputting the health status data of each individual battery cell, the following is also included: Based on the health status data of each individual battery cell, determine the maximum difference in health status data; Obtain at least two preset difference threshold intervals; Based on the maximum difference and the preset at least two difference threshold intervals, determine the target difference threshold interval to which the maximum difference belongs; Output early warning information corresponding to the target difference threshold range.

7. A device for acquiring battery health status data, characterized in that, include: The data acquisition module is used to collect charging voltage data, charging current data, and temperature data of each individual battery in the energy storage compartment at different times, based on a preset sampling frequency. The generation module is used to generate the incremental capacity curve of each individual battery cell based on the charging voltage data and charging current data of each individual battery cell at different times. The extraction module is used to extract health characteristic data related to the capacity decay of each individual battery cell based on the incremental capacity curve of each individual battery cell. The processing module is used to obtain the feature vector sequence of each individual battery cell based on the health characteristic data related to the capacity decay of each individual battery cell and the corresponding temperature data. The processing module is also used to input the feature vector sequence of each individual battery cell into a pre-trained neural network model and output the health status data of each individual battery cell.

8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-6.

Images