A method, system, device and medium for monitoring the state of an energy storage power station

Abstract

The application discloses a kind of energy storage power station state monitoring method, system, equipment and medium, belong to energy storage power station state monitoring technical field, comprising: constructing cross-event response dataset;Second coordination feature and third coordination feature of battery monomer in the battery monomer set are extracted to cross-event response dataset;The multidimensional feature of battery monomer in the battery monomer set is constructed;Mismatch measure is carried out to the multidimensional feature of battery monomer in the battery monomer set, obtains the mismatch characteristic quantity of battery monomer in the battery monomer set;Based on series electrical connection relationship, according to the distribution of mismatch characteristic quantity of battery monomer in the battery monomer set in topological space, identify the degradation source battery monomer.The application obtains mismatch characteristic quantity by mismatch measure, and identifies degradation source based on the distribution of mismatch characteristic quantity in topological space.Degradation battery influences the coordination of adjacent battery, and mismatch characteristic quantity forms local extreme value at degradation source position.

Description

Technical Field

[0001] This invention relates to the field of energy storage power station condition monitoring technology, specifically to an energy storage power station condition monitoring method, system, equipment, and medium. Background Technology

[0002] Energy storage power stations typically employ large-scale series-connected battery packs to meet high voltage requirements. In this series connection configuration, the degradation of individual cells can affect the performance of the entire battery pack.

[0003] Existing battery state monitoring technologies employ tensor data structures to process multidimensional time-series data. For example, some solutions organize the device's time-series data and feature parameters into a three-dimensional tensor and use tensor decomposition (Tucker decomposition) to complete missing data for subsequent energy scheduling optimization. Other solutions construct the battery pack's time-series state data into a three-dimensional tensor and build a graph structure based on the topological connections between batteries, utilizing graph neural networks for health status assessment.

[0004] However, the aforementioned existing technologies have limitations when applied to the identification of degradation sources in series-connected battery packs. The Tucker decomposition method for data completion fails to consider the unique topological constraints of series systems, resulting in a factor matrix that lacks the ability to represent the coordination between series-connected batteries. While the graph neural network approach utilizes the topological graph structure, its analysis relies on extensive labeled data for training, and the graph structure only reflects static electrical connections, failing to extract the evolutionary characteristics of battery response from multiple operating condition changes.

[0005] Furthermore, existing technologies typically analyze battery state from a single perspective, such as focusing only on voltage response or frequency domain characteristics, making it difficult to comprehensively capture the manifestation of battery degradation across different physical processes. In series systems, degraded batteries disrupt the overall response coordination in the topological space, but existing technologies lack quantitative assessment methods for this disruption. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention provides a method, system, equipment and medium for monitoring the status of energy storage power stations.

[0007] Therefore, the technical problem solved by this invention is: how to accurately identify the deteriorating battery cells that disrupt the overall coordination by comprehensively considering the coordination characteristics of multiple dimensions from the response data under multiple operating conditions during the operation of a series battery pack.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for monitoring the status of an energy storage power station, comprising,

[0009] Acquire electrical parameter response data of a battery cell assembly connected in series under multiple operating condition change events, and construct a cross-event response dataset from the electrical parameter response data under multiple operating condition change events.

[0010] Based on the series electrical connection relationship, the first coordination feature of the battery cells in the battery cell set is extracted from the cross-event response dataset in the first dimension. The second coordination feature and the third coordination feature of the battery cells in the battery cell set are extracted from the cross-event response dataset in the second and third dimensions, respectively.

[0011] The first coordination feature, the second coordination feature, and the third coordination feature are constructed as multidimensional features of the battery cells in the battery cell set;

[0012] Mismatch measurement is performed on the multidimensional features of the battery cells in the battery cell set to obtain the mismatch feature quantity of the battery cells in the battery cell set;

[0013] Based on the series electrical connection relationship, the deterioration source battery cells are identified according to the distribution of the mismatch characteristics of the battery cells in the topological space.

[0014] As a preferred embodiment of the energy storage power station condition monitoring method of the present invention, the step of extracting the first coordination feature of the battery cells in the battery cell set based on the series electrical connection relationship in the first dimension of the cross-event response dataset includes:

[0015] Construct a topological adjacency matrix based on series electrical connections;

[0016] For each battery cell in the battery cell set, obtain the topological adjacency position index of each battery cell from the topological adjacency matrix;

[0017] The response data of each battery cell and the topological adjacency position of each battery cell are extracted from the cross-event response dataset, and a topological space vector field structure is constructed in the topological space.

[0018] The coordination features of the topological space vector field structure are extracted to obtain the first coordination features of each battery cell.

[0019] As a preferred embodiment of the energy storage power station status monitoring method of the present invention, wherein: constructing the first coordination feature, the second coordination feature, and the third coordination feature into multidimensional features of the battery cells in the battery cell set includes:

[0020] In the multidimensional feature space, the first coordination feature of each battery cell in the battery cell set is determined as the component value of the first spatial dimension, the second coordination feature is determined as the component value of the second spatial dimension, and the third coordination feature is determined as the component value of the third spatial dimension.

[0021] Based on the component values ​​of the first spatial dimension, the component values ​​of the second spatial dimension, and the component values ​​of the third spatial dimension, the spatial position coordinates of each battery cell are determined in the multidimensional feature space.

[0022] As a preferred embodiment of the energy storage power station condition monitoring method of the present invention, the step of measuring the mismatch of the multidimensional characteristics of the battery cells in the battery cell set to obtain the mismatch characteristic quantity of the battery cells in the battery cell set includes:

[0023] In the multidimensional feature space, the spatial position of each battery cell is determined based on the multidimensional features of each battery cell in the battery cell set.

[0024] Spatial position deviation pattern recognition is performed on the spatial position of each battery cell in the battery cell set. Spatial position deviation pattern recognition is achieved by analyzing the distribution deviation characteristics of the spatial position of each battery cell in a multi-dimensional feature space.

[0025] The mismatch characteristics of each battery cell are obtained from the spatial position deviation pattern recognition results.

[0026] As a preferred embodiment of the energy storage power station condition monitoring method of the present invention, the step of identifying the deterioration source battery cells based on the distribution of mismatch characteristics of battery cells in the topological space according to the series electrical connection relationship includes:

[0027] The topological location index of each battery cell in the battery cell set is determined from the series electrical connection relationship;

[0028] The mismatch feature quantities of each battery cell in the battery cell set are arranged according to the topological position index to construct a topological distribution sequence of mismatch feature quantities;

[0029] The topological distribution sequence of the mismatch feature quantity is subjected to extreme value identification in the topological location direction, and the topological location index where the mismatch feature quantity exhibits local extreme values ​​is identified.

[0030] The cell in which the topological location index of the mismatch feature quantity shows a local extremum is identified as the cell in which the degradation source cell is located.

[0031] As a preferred embodiment of the energy storage power station condition monitoring method of the present invention, after identifying the deteriorating battery cell, the method further includes:

[0032] The multiple operating condition change events are divided into multiple time windows according to their chronological order.

[0033] For each of the multiple time windows, the following processes are executed: cross-event response dataset construction, first coordination feature extraction, second coordination feature extraction, third coordination feature extraction, multi-dimensional feature construction, and mismatch feature quantity acquisition, to obtain the mismatch feature quantity of the degraded source battery cell in each time window.

[0034] The mismatch characteristics of the degraded battery cells in the multiple time windows are arranged in chronological order to construct a temporal evolution sequence of mismatch characteristics.

[0035] As a preferred embodiment of the energy storage power station condition monitoring method of the present invention, after constructing the time-series evolution sequence of the mismatch characteristic quantity, the method further includes:

[0036] A degradation process prediction model is established based on the temporal evolution sequence of the mismatch features;

[0037] The degradation process prediction model is used to predict the mismatch characteristics of future time windows.

[0038] Set a failure threshold for mismatch features, and identify the time window that reaches the failure threshold from the predicted mismatch features;

[0039] The time window when the failure threshold of the mismatch characteristic quantity is reached is determined as the predicted failure time node of the deterioration source battery cell.

[0040] This invention provides a status monitoring system for energy storage power stations.

[0041] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a condition monitoring system for an energy storage power station, comprising:

[0042] The data acquisition unit is used to acquire the electrical parameter response data of a set of battery cells connected in series under multiple operating condition change events, and to construct the electrical parameter response data under multiple operating condition change events into a cross-event response dataset.

[0043] The coordination feature extraction unit is used to extract the first coordination feature of the battery cells in the battery cell set based on the series electrical connection relationship in the first dimension of the cross-event response dataset, and to extract the second coordination feature and the third coordination feature of the battery cells in the battery cell set in the second dimension and the third dimension, respectively.

[0044] A multi-dimensional feature construction unit is used to construct the first coordination feature, the second coordination feature, and the third coordination feature into multi-dimensional features of the battery cells in the battery cell set;

[0045] The mismatch measurement unit is used to measure the mismatch of the multidimensional features of the battery cells in the battery cell set and obtain the mismatch feature quantity of the battery cells in the battery cell set.

[0046] The degradation source identification unit is used to identify the degradation source battery cells based on the series electrical connection relationship and the distribution of the mismatch characteristics of the battery cells in the battery cell set in the topological space.

[0047] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the energy storage power station status monitoring method.

[0048] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the energy storage power station status monitoring method.

[0049] The beneficial effects of this invention are as follows: This invention extracts coordination features based on series electrical connections in the first dimension, analyzes the response correlation between each battery and its topologically adjacent batteries, and transforms series topological constraints into quantifiable first coordination features. The deterioration of a battery and the coordination relationship with adjacent batteries cause deviations in feature values.

[0050] Existing technologies employ graph neural networks that rely on labeled data for training and fail to extract response evolution features from multiple operating condition changes. This invention constructs a cross-event response dataset from multiple operating condition change events to analyze response evolution patterns. Healthy battery responses are repeatable, while degraded battery responses deviate from this repeatability. The cross-event dataset captures evolutionary deviation features without requiring labeled data.

[0051] Existing technologies, analyzing from a single perspective, struggle to comprehensively capture degradation performance. This invention constructs a multi-dimensional feature from three dimensions of compatibility characteristics. Internal resistance degradation primarily affects topological compatibility and frequency domain compatibility, while capacity degradation primarily affects topological compatibility and temporal evolution compatibility. Different degradation mechanisms exhibit different distributions in the multi-dimensional feature space, and mismatch metrics identify degradation and distinguish mechanisms.

[0052] Existing technologies lack quantitative assessment of the disruption of coordination in series systems. This invention obtains mismatch characteristic quantities through mismatch measurement and identifies degradation sources based on the distribution of these mismatch characteristic quantities in the topological space. Degraded cells affect the coordination of adjacent cells, and the mismatch characteristic quantities form local extrema at the locations of degradation sources. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 The above is a general flowchart of a method for monitoring the status of an energy storage power station, provided as an embodiment of the present invention.

[0055] Figure 2 The flowchart of step S2 of a condition monitoring method for an energy storage power station provided in an embodiment of the present invention is shown.

[0056] Figure 3 The flowchart of step S3 of a condition monitoring method for an energy storage power station provided in an embodiment of the present invention is shown. Detailed Implementation

[0057] To make the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0058] Example 1, referring to Figures 1-3 This is one embodiment of the present invention, which provides a method for monitoring the status of an energy storage power station, including:

[0059] S1: Obtain the electrical parameter response data of a set of battery cells connected in series under multiple operating condition change events, and construct a cross-event response dataset from the electrical parameter response data under multiple operating condition change events.

[0060] It should be noted that in step S1, this embodiment applies to the state monitoring of battery packs in an energy storage power station. The battery packs in the energy storage power station are connected in series, which forces each battery cell in the battery cell set to have a uniform current. During operation, changes in load power or battery management strategies can cause the topology management system to perform topology switching operations. Topology switching operations alter the current path of the series circuit, causing changes in the operating current of each battery cell in the battery cell set, which in turn leads to changes in the terminal voltage response. This embodiment identifies the deteriorating battery cells from the evolution pattern of the terminal voltage response under multiple topology switching.

[0061] In some embodiments, the electrical parameter response data are the terminal voltage response data of each battery cell in the battery cell set. The terminal voltage directly reflects the transient response characteristics of each battery cell in the battery cell set during topology switching.

[0062] Specifically, topology switching causes a change in the total impedance of the series circuit, resulting in a change in the circuit current. This change in circuit current leads to changes in the ohmic voltage drop and polarization voltage of each cell in the battery cell assembly, manifested as a transient response in the terminal voltage. The transient response characteristics of each cell in the battery cell assembly are affected by its internal resistance and polarization characteristics. Degraded cells exhibit altered internal resistance and polarization characteristics, resulting in a difference in transient response compared to healthy cells.

[0063] In some embodiments, three consecutive topology switching events that occur in chronological order are selected from multiple topology switching events monitored during the continuous operation of the energy storage power station.

[0064] It should be noted that the selection of three consecutive topology switching events addresses a deficiency in existing technologies: To address the problem that existing technologies cannot distinguish between transient fluctuations and true degradation from a single measurement, this embodiment selects consecutively occurring topology switching events. Battery degradation is a gradual, time-evolving process. Under continuous changes in operating conditions, the response of healthy battery cells remains stable and repeatable, while the response of degraded battery cells gradually deviates from the normal pattern over time. The transient response of a single measurement is affected by the current operating point, ambient temperature, and measurement noise, misinterpreting occasional fluctuations as degradation. The short time intervals between consecutive operating condition changes result in a relatively gradual change in the degree of battery degradation, exhibiting a continuous response evolution, which allows for the identification of the gradual evolution characteristics of degradation.

[0065] Furthermore, addressing the issue of existing technologies relying on large amounts of labeled data to train models, this embodiment constructs second-order evolutionary features of response changes by selecting three consecutive events, eliminating the need for labeled data. Two topology switches yield the first-order change in response, i.e., the difference in response between the first and second events, but this first-order change may involve random fluctuations. Three topology switches yield the second-order change in response: the difference in response among individual cells in the battery cell set between the first and second topology switches is the first-order change; the difference in response among individual cells in the battery cell set between the second and third topology switches is the second-order change; the difference between the two first-order changes is the second-order change. The response of healthy battery cells remains repeatable under continuous operating conditions, exhibiting randomness in the first-order change and near-zero second-order change; the response of degraded battery cells gradually deviates with changing operating conditions, showing a trend in the first-order change and deviating from zero in the second-order change. The second-order change feature distinguishes between random fluctuations and degradation trends, and this judgment is based on physical laws rather than data-driven model training.

[0066] Furthermore, for the three selected topology switching events, the transient response of the terminal voltage of each battery cell in the battery cell set is collected. The battery management system triggers terminal voltage data acquisition when it detects a topology switching signal, with the start time of data acquisition being the moment the topology switching signal is emitted. It can be understood that the transient response of a lithium-ion battery includes two stages: ohmic response and polarization response. The ohmic response is a voltage jump, while the polarization response is a gradual voltage change. The acquisition duration covers both the ohmic and polarization response stages, ensuring that the acquired terminal voltage data encompasses the complete transient response process. The sampling frequency satisfies the Nyquist sampling theorem.

[0067] Furthermore, the terminal voltage values ​​of each battery cell in the battery cell set collected during the first topology switching event are organized into a first operating condition change event response matrix according to the positional and temporal order of the battery cells in the series topology. It should be noted that the row indices of the matrix strictly follow the topological positional order of the battery cells in the battery cell set, reflecting the series electrical connection relationship. Adjacent rows in the matrix represent adjacent battery cells on the series link. The column indices of the matrix correspond to the sampling times in the acquisition time series. The second and third operating condition change event response matrices are obtained using the same method.

[0068] Furthermore, the response matrices for the first, second, and third operating condition change events are stacked along the third dimension to construct a cross-event response dataset. This cross-event response dataset is a three-dimensional tensor structure. The first modality index corresponds to the topological position of each battery cell in the battery cell set, the second modality index corresponds to the time sampling point, and the third modality index corresponds to the sequence of operating condition change events.

[0069] It is easy to understand that, unlike existing technologies that treat topological relationships merely as external constraints or neural network inputs, this embodiment directly embeds the series topological relationships into the first modality index structure across the event response dataset. In the first modality direction, the adjacency relationships of each battery cell in the battery cell set on the series topology are directly reflected as the adjacency relationships of the tensor index. Based on the adjacency of the first modality index, topologically adjacent battery cells can be directly identified without the need for additional topological relationship matrices or graph structures as auxiliary inputs. In the second modality direction, the tensor is expanded along the time sampling point index, allowing for time-domain analysis or frequency-domain transformation of the response data of each battery cell in the battery cell set. In the third modality direction, the tensor is expanded along the event sequence index, allowing for analysis of the evolution trend of the response of each battery cell in the battery cell set between multiple topological switching, and extraction of the aforementioned second-order change features. The three-dimensional tensor structure of the cross-event response dataset unifies the three physical dimensions of serial topological constraints, time-frequency domain characteristics, and cross-event evolution into the three modal indices of the tensor. This avoids treating topological relationships as external constraints separately and avoids treating time-frequency domain characteristics and cross-event evolution as independent analysis steps.

[0070] S2: Based on the series electrical connection relationship, extract the first coordination feature of the battery cells in the battery cell set across the event response dataset in the first dimension. In the second and third dimensions, extract the second coordination feature and the third coordination feature of the battery cells in the battery cell set across the event response dataset, respectively.

[0071] It should be noted that step S2 extracts coordination features from three different dimensions. The first dimension is the topological space dimension, which extracts coordination features in the first modal direction using the series electrical connection relationship. The second dimension is the frequency domain space dimension, which extracts frequency domain coordination features in the second modal direction through time-frequency domain transformation. The third dimension is the time evolution dimension, which extracts coordination features across events in the third modal direction.

[0072] Furthermore, in some embodiments, step S2 includes S21-S23.

[0073] S21: Extract the first coordination feature of battery cells in the battery cell set based on the series electrical connection relationship in the first dimension across the event response dataset.

[0074] In some embodiments, S21 includes S211-S214.

[0075] S211: Construct a topological adjacency matrix based on series electrical connection relationships.

[0076] In some embodiments, a topological adjacency matrix is ​​constructed based on the series electrical connection relationship and the topological position order of the first modal index across the event response dataset. It is understood that in step S1, the first modal index across the event response dataset is arranged strictly according to the topological position order of each battery cell in the battery cell set. The series electrical connection relationship determines that battery cells at adjacent positions on the first modal index are adjacent on the series link.

[0077] Specifically, the topological adjacency matrix is ​​a square matrix with both rows and columns equal to the number of battery cells in the battery cell set. The element in the i-th row and j-th column of the topological adjacency matrix indicates whether a battery cell at the i-th topological position is adjacent to a battery cell at the j-th topological position. In the series connection mode, the adjacent battery cells of the battery cell at the i-th topological position are the battery cells at the (i-1)-th and (i+1)-th topological positions. If the i-th and j-th topological positions are adjacent on the series link, i.e., |ij|=1, then the element in the i-th row and j-th column of the topological adjacency matrix is ​​assigned a value of 1; otherwise, it is assigned a value of 0.

[0078] It should be noted that, unlike existing technologies that use the topology graph structure as an external input or construct the topology graph independently using static electrical connection information, the topology adjacency matrix in this embodiment is directly constructed based on the first modal index structure across the event response dataset and the series electrical connection relationships. The first modal index across the event response dataset has been organized according to the topological position order, and the topology adjacency matrix is ​​a matrix representation of this order relationship under the series connection constraint.

[0079] S212: For each battery cell in the battery cell set, obtain the topological adjacency position index of each battery cell from the topological adjacency matrix.

[0080] Furthermore, for each battery cell in the battery cell set, the topological adjacency index of that battery cell is found in the topological adjacency matrix. Specifically, for the battery cell at the i-th topological position, the column index with a value of 1 in the i-th row of the topological adjacency matrix is ​​found. This column index is the topological adjacency index of the battery cell at the i-th topological position.

[0081] For example, for a battery cell at the k-th topological position, if k is neither equal to 1 nor equal to the total number of battery cells in the set, its topological adjacency index includes the (k-1)-th and (k+1)-th positions. For a battery cell at the 1-th topological position, its topological adjacency index is the 2-th position. For a battery cell whose topological position is the total number of battery cells in the set, its topological adjacency index is the total number minus 1 position.

[0082] S213: Extract the response data of each battery cell and the topological adjacency position of each battery cell from the cross-event response dataset, and construct the topological space vector field structure in the topological space.

[0083] Furthermore, based on the topological adjacency location index obtained in step S212, response data is extracted from the cross-event response dataset.

[0084] Specifically, for the battery cell at the k-th topological position in the battery cell set, the response data of the battery cell at the k-th topological position in the battery cell set under multiple operating condition change events is extracted from the cross-event response dataset. It is easy to understand that the cross-event response dataset is a three-dimensional tensor structure, where the first modality index corresponds to the topological position, the second modality index corresponds to the time sampling point, and the third modality index corresponds to the operating condition change event. Fixing the first modality index at the k-th position, slicing is performed on the second and third modalities to obtain an event response evolution matrix with dimensions of [number of time sampling points × number of operating condition change events].

[0085] Furthermore, column-wise extraction is performed from the event response evolution matrix of the battery cell at the k-th topological position in the battery cell set to obtain the time-domain response vectors of the battery cell at the k-th topological position in the first operating condition change event, the time-domain response vectors of the battery cell at the k-th topological position in the second operating condition change event, and the time-domain response vectors of the battery cell at the k-th topological position in the third operating condition change event. For each adjacent position in the topological adjacency index of the battery cell at the k-th topological position in the battery cell set, the time-domain response vectors of each topologically adjacent battery cell under the three operating condition change events are extracted from the cross-event response dataset in the same way.

[0086] Understandably, a topological space vector field structure is constructed within the topological space. This structure organizes the time-domain response vector of the battery cell at the k-th topological position in the battery cell set, along with the time-domain response vectors of its topologically adjacent battery cells, into a vector field under topological adjacency constraints. In this vector field structure, the time-domain response vector of the battery cell at the k-th topological position in the battery cell set serves as the center vector, and the time-domain response vectors of its topologically adjacent battery cells serve as the adjacency vectors.

[0087] It should be noted that the topological space vector field structure reflects the response coordination of series-connected batteries under the constraint of forced current uniformity. In the series connection mode, adjacent battery cells experience the same current change when operating conditions change, and the response of a healthy battery cell should be coordinated with the response of its adjacent battery cell. Degraded battery cells have altered internal resistance and polarization characteristics, causing their response to deviate from that of adjacent battery cells.

[0088] S214: Extract the coordination features of the topological space vector field structure to obtain the first coordination features of each battery cell.

[0089] Furthermore, coordination features are extracted from the topological space vector field structure. It can be understood that coordination feature extraction is based on series current consistency constraints, identifying whether the center vector and adjacent vectors maintain a coordination relationship.

[0090] In some embodiments, for a battery cell at the k-th topological position in the battery cell set, in the topological space vector field structure, the time-domain response vector of the battery cell at the k-th topological position under each operating condition change event is used as the center vector, and the time-domain response vectors of the topologically adjacent battery cells at the k-th topological position under the same operating condition change event are used as the adjacency vectors. The difference between the center vector and each adjacency vector is calculated, and the difference reflects the degree of response deviation between the battery cell at the k-th topological position and its topologically adjacent battery cells. The average difference between the center vector and the adjacency vectors under each operating condition change event is taken as the first coordination characteristic of the battery cell at the k-th topological position in the battery cell set.

[0091] It should be noted that, unlike existing technologies that use Tucker decomposition to decompose tensors without considering topological constraints, this embodiment directly analyzes the response coordination between adjacent battery cells in the first modal direction based on the series topological constraint relationship. The series current consistency constraint requires that the responses of adjacent battery cells be coordinated, and the first coordination characteristic quantifies the degree to which each battery cell satisfies this topological coordination.

[0092] S22: Extract the second coordination feature of battery cells in the battery cell set across the event response dataset in the second dimension.

[0093] In some embodiments, S22 includes S221-S223.

[0094] S221: Extract time-domain response data of each battery cell in the battery cell set under multiple operating condition change events from the cross-event response dataset.

[0095] In some embodiments, the time-domain response vectors of each battery cell in the battery cell set under multiple operating condition change events are extracted from the cross-event response dataset. Optionally, the time-domain response vectors extracted in step S213 can be used directly.

[0096] S222: Perform time-frequency domain transformation on the time-domain response data of each battery cell in the battery cell set under multiple operating condition change events to obtain the frequency domain response data of each battery cell in the battery cell set under multiple operating condition change events.

[0097] Furthermore, Fourier transforms are performed on the time-domain response vectors of each battery cell in the battery cell set under each operating condition change event, transforming the time-domain response vectors into frequency-domain response vectors. It can be understood that the Fourier transform converts the response data from the time domain to the frequency domain, and the frequency-domain response vector reflects the frequency component distribution of the response signal.

[0098] It should be noted that the transient response characteristics of a battery exhibit a characteristic frequency distribution in the frequency domain. The ohmic response corresponds to the high-frequency components, while the polarization response corresponds to the mid-to-low-frequency components. Changes in the polarization characteristics of degraded battery cells lead to changes in the frequency domain distribution characteristics.

[0099] S223: Based on the distribution characteristics of the frequency domain response data of each battery cell in the battery cell set under multiple operating condition change events in the frequency domain space, the second coordination characteristics of each battery cell are obtained.

[0100] Furthermore, based on the frequency domain response vectors of each battery cell in the battery cell set under multiple operating condition change events, the consistency of the frequency domain response of each battery cell among the three operating condition change events is analyzed.

[0101] In some embodiments, for the battery cell at the k-th topological position in the battery cell set, the difference in the frequency domain response vector of the battery cell at the k-th topological position in the battery cell set under the first and second operating condition change events is calculated. The difference in the frequency domain response vector of the battery cell at the k-th topological position in the battery cell set under the second and third operating condition change events is also calculated. The mean of the two differences reflects the degree of deviation in the frequency domain response of the battery cell at the k-th topological position in the battery cell set between multiple operating condition change events, and the mean value serves as a second consistency characteristic of the battery cell at the k-th topological position in the battery cell set.

[0102] It should be noted that, unlike existing technologies that rely on labeled data for training graph neural networks, this embodiment extracts the second coordination feature based on the consistency of the frequency domain response across multiple operating condition change events. Healthy battery cells maintain repeatable frequency domain responses under multiple operating condition change events, while the polarization characteristics of degraded battery cells cause deviations in the frequency domain response across these events. The second coordination feature reflects the degree of coordination among battery cells in the frequency domain space; this coordination is based on physical laws rather than data-driven model training.

[0103] S23: Extract the third coordination feature of battery cells in the battery cell set across the event response dataset in the third dimension.

[0104] In some embodiments, S23 includes S231-S233.

[0105] S231: Extract the evolution data of each battery cell in the battery cell set across multiple operating condition change events at each time point from the cross-event response dataset.

[0106] In some embodiments, evolutionary data is extracted from the third modality index direction across the event response dataset. It is understood that the third modality index across the event response dataset corresponds to a sequence of operating condition change events, and the evolutionary data is the evolutionary sequence of each battery cell in the battery cell set across three operating condition change events at each time point.

[0107] Specifically, for a battery cell at the k-th topological position in the battery cell set, at the t-th time point of the time sampling sequence, the response values ​​of this battery cell at the t-th time point of the first operating condition change event, the first modality index is fixed at the k-th position, the second modality index is fixed at the t-th position, and the response values ​​of the third operating condition change event are extracted from the cross-event response dataset. These three response values ​​constitute the evolution data of the battery cell at the k-th topological position in the battery cell set at the t-th time point across multiple operating condition change events.

[0108] S232: Construct a time evolution trajectory from the evolution data of each battery cell in the battery cell set across multiple operating condition change events at each time point.

[0109] Furthermore, for the battery cell at the k-th position in the battery cell set, its evolution data across multiple operating condition change events at various time points is organized chronologically into a time evolution trajectory. The time evolution trajectory is a two-dimensional data structure, with the first dimension corresponding to the time sampling points and the second dimension corresponding to the operating condition change events.

[0110] Understandably, the time evolution trajectory depicts the second-order change characteristics described in step S1. On the time evolution trajectory, the evolution between the first operating condition change event and the second operating condition change event is a first-order change, the evolution between the second operating condition change event and the third operating condition change event is a first-order change, and the relationship between two first-order changes constitutes a second-order change.

[0111] S233: Based on the smoothness of the evolution of the time evolution trajectory of each battery cell in the battery cell set in the time series, the third coordination characteristic of each battery cell is obtained.

[0112] Furthermore, second-order variation features are extracted based on the time evolution trajectory of each battery cell in the battery cell set.

[0113] In some embodiments, for the time evolution trajectory of the battery cell at the k-th topological position in the battery cell set, the difference between the first event and the second event, and the difference between the second event and the third event are calculated at each time point. The difference between these two differences is the second-order difference at that time point. The mean of the second-order differences at all time points is taken. The mean reflects the degree of second-order change of the battery cell at the k-th topological position in the battery cell set on its time evolution trajectory. The mean serves as the third coordination characteristic of the battery cell at the k-th topological position in the battery cell set.

[0114] It should be noted that the response of a healthy battery cell remains repeatable under continuous operating conditions, with the first-order change in its time evolution trajectory exhibiting randomness, and the second-order difference fluctuating around zero. The response of a degraded battery cell gradually deviates with changes in operating conditions, with the first-order change in its time evolution trajectory showing a trend, and the second-order difference deviating from zero. The third consistency characteristic identifies the degradation trend through second-order changes, distinguishing it from existing methods that analyze from a single dimension.

[0115] S3: Construct the first coordination feature, the second coordination feature, and the third coordination feature into multidimensional features of the battery cells in the battery cell set.

[0116] It should be noted that existing technologies analyze battery status from a single perspective, making it difficult to comprehensively capture the manifestation of degradation in different physical processes and distinguish between different degradation mechanisms. This embodiment organizes the three-dimensional coordination features into multi-dimensional feature vectors in a multi-dimensional feature space, so that different degradation mechanisms exhibit different spatial distribution patterns in the multi-dimensional feature space.

[0117] Furthermore, in some embodiments, step S3 includes S31-S32.

[0118] S31: In the multidimensional feature space, the first coordination feature of each battery cell in the battery cell set is determined as the component value of the first spatial dimension, the second coordination feature is determined as the component value of the second spatial dimension, and the third coordination feature is determined as the component value of the third spatial dimension.

[0119] In some embodiments, for the battery cell at the k-th topological position in the battery cell set, the first coordination feature of the battery cell at the k-th topological position in the battery cell set obtained in step S214 is used as the component value of the first spatial dimension of the battery cell at the k-th topological position in the multidimensional feature space; the second coordination feature of the battery cell at the k-th topological position in the battery cell set obtained in step S223 is used as the component value of the second spatial dimension of the battery cell at the k-th topological position in the multidimensional feature space; and the third coordination feature of the battery cell at the k-th topological position in the battery cell set obtained in step S233 is used as the component value of the third spatial dimension of the battery cell at the k-th topological position in the multidimensional feature space.

[0120] It should be noted that different degradation mechanisms have different effects on the coordination of the three dimensions. Internal resistance degradation changes the ohmic voltage drop characteristics and ohmic response characteristics of the battery, causing changes in the response difference between the battery cell at the k-th topological position in the battery cell set and its topologically adjacent battery cells. At the same time, it causes changes in the frequency domain response difference of the battery cell at the k-th topological position in the battery cell set between multiple operating condition changes. Therefore, the component values ​​of the battery cell with internal resistance degradation deviate from the normal range in the first and second spatial dimensions, while the component value in the third spatial dimension remains within the normal range.

[0121] Capacity degradation alters the usable capacity characteristics of a battery, causing changes in the response amplitude of the battery cell at the k-th topological position in the battery cell set under continuous operating conditions. These changes in response amplitude lead to variations in the response differences between the battery cell at the k-th topological position and its adjacent topological positions. Furthermore, the changes in response amplitude exhibit a first-order trend in the time evolution trajectory, causing the second-order difference to deviate from zero. Therefore, the component values ​​of a capacity-degraded battery cell deviate from the normal range in the first and third spatial dimensions, while the component value in the second spatial dimension remains within the normal range.

[0122] S32: Based on the component values ​​of the first spatial dimension, the component values ​​of the second spatial dimension, and the component values ​​of the third spatial dimension, determine the spatial position coordinates of each battery cell in the multidimensional feature space.

[0123] Furthermore, for the battery cell at the k-th topological position in the battery cell set, the spatial coordinates of the battery cell at the k-th topological position in the multidimensional feature space are composed of three component values: the first coordination feature, the second coordination feature, and the third coordination feature of the battery cell at the k-th topological position in the battery cell set.

[0124] It is understandable that the first, second, and third coordination characteristics of healthy battery cells are all within the normal range, and healthy battery cells cluster together to form characteristic regions in the multidimensional feature space. The first and second spatial dimension component values ​​of battery cells with deteriorated internal resistance deviate from the normal range, and the spatial coordinates of these deteriorated battery cells deviate in the first and second spatial dimension planes. Similarly, the first and third spatial dimension component values ​​of battery cells with deteriorated capacity deviate from the normal range, and the spatial coordinates of these deteriorated battery cells deviate in the first and third spatial dimension planes. The positional distribution pattern in the multidimensional feature space directly reflects the physical characteristics of the degradation mechanism. Internal resistance degradation and capacity degradation deviate along different planar directions in the multidimensional space, and the degradation mechanism can be distinguished by the direction of deviation of the spatial coordinates.

[0125] S4: Perform a mismatch measurement on the multidimensional characteristics of the individual cells in the battery cell set to obtain the mismatch characteristic quantity of the individual cells in the battery cell set.

[0126] It should be noted that step S4 identifies degraded battery cells in the battery cell set through mismatch measurement. It should be noted that, unlike existing technologies that rely on labeled data to train models for identifying deterioration, this embodiment uses positional distribution characteristics in a multi-dimensional feature space for mismatch measurement. Healthy battery cells cluster in the multi-dimensional feature space to form feature regions, while the spatial coordinates of degraded battery cells deviate from these feature regions. Degraded battery cells are identified by analyzing the deviation patterns of their spatial positions.

[0127] Furthermore, in some embodiments, step S4 includes S41-S43.

[0128] S41: In the multidimensional feature space, the spatial location of each battery cell is determined based on the multidimensional features of each battery cell in the battery cell set.

[0129] In some embodiments, the spatial position of each battery cell in the battery cell set is the spatial position coordinate determined in step S32.

[0130] S42: Spatial position deviation pattern recognition is performed on the spatial position of each battery cell in the battery cell set. Spatial position deviation pattern recognition is achieved by analyzing the distribution deviation characteristics of the spatial position of each battery cell in the multi-dimensional feature space.

[0131] Furthermore, spatial position deviation pattern recognition is performed on the spatial positions of each battery cell in the battery cell set. It is understood that spatial position deviation pattern recognition is based on the overall distribution characteristics of each battery cell in the multi-dimensional feature space, rather than analyzing the spatial position of each individual battery cell.

[0132] In some embodiments, the distribution of each battery cell in the battery cell set in a multidimensional feature space is first analyzed to determine the characteristic region formed by the aggregation of healthy battery cells in the battery cell set. The first coordination characteristic, the second coordination characteristic, and the third coordination characteristic of the healthy battery cells are all within the normal range, and the spatial coordinates of the healthy battery cells are aggregated in the multidimensional feature space.

[0133] Furthermore, for the battery cell at the k-th topological position in the battery cell set, we analyze the deviation of its spatial coordinates from the cluster of healthy battery cells. This deviation reflects the degree of difference between the battery cell at the k-th topological position and the group of healthy battery cells in the multidimensional feature space.

[0134] It should be noted that spatial position deviation pattern recognition analyzes the distribution deviation characteristics of each battery cell's spatial position in a multi-dimensional feature space. These deviation characteristics include not only the distance relationship between the spatial position coordinates and the healthy region, but also the directional characteristics of the spatial position coordinates deviating from the healthy region. As described in step S3, battery cells with deteriorated internal resistance deviate within the plane formed by the first and second spatial dimensions, while battery cells with deteriorated capacity deviate within the plane formed by the first and third spatial dimensions. The directional characteristics of deviation differ for different degradation mechanisms. Spatial position deviation pattern recognition identifies deteriorated battery cells and distinguishes degradation mechanisms by analyzing the distance and direction of the deviation.

[0135] S43: Obtain the mismatch characteristics of each battery cell from the spatial position deviation pattern recognition results.

[0136] Furthermore, the mismatch characteristics of each battery cell are obtained from the spatial position deviation pattern recognition results.

[0137] In some embodiments, for a battery cell at the k-th topological position in a battery cell set, a mismatch feature is obtained based on the deviation relationship between the spatial coordinates of the battery cell at the k-th topological position and the cluster region of healthy battery cells. The mismatch feature quantifies the degree of deviation of the battery cell at the k-th topological position in the multidimensional feature space.

[0138] Understandably, the mismatch characteristics of healthy battery cells in a battery cell set are within the normal range, while the mismatch characteristics of degraded battery cells deviate from the normal range. The mismatch characteristics quantify the overall performance of each battery cell in the battery cell set across three coordination dimensions into a single numerical value.

[0139] S5: Based on the series electrical connection relationship, identify the deterioration source battery cells according to the distribution of the mismatch characteristics of battery cells in the topological space.

[0140] It should be noted that step S5 identifies the deterioration source battery cell based on the distribution of mismatch characteristics in the topological space. It should be explained that, unlike existing technologies that independently analyze the mismatch degree of each battery cell and select the cell with the largest mismatch degree, this embodiment identifies the deterioration source from the distribution pattern of mismatch characteristics in the topological space. In the series connection configuration, the deterioration source battery cell not only has a high mismatch degree itself, but also affects the response coordination of its topologically adjacent battery cells, causing the mismatch characteristics of the topologically adjacent battery cells to be affected, forming a mismatch characteristic distribution pattern centered on the deterioration source. The deterioration source battery cell exhibits local extrema of mismatch characteristics in the topological position sequence; identifying the extrema positions from the distribution of mismatch characteristics in the topological space allows for accurate location of the deterioration source battery cell.

[0141] Furthermore, in some embodiments, step S5 includes S51-S54.

[0142] S51: Determine the topological location index of each battery cell in the battery cell set from the series electrical connection relationship.

[0143] In some embodiments, the topological location index of each battery cell in the battery cell set is obtained from the series electrical connection relationship. It is understood that in step S1, the first modal index across the event response dataset is arranged strictly according to the topological location order of each battery cell in the battery cell set; the first modal index is the topological location index.

[0144] S52: Arrange the mismatch characteristics of each battery cell in the battery cell set according to the topological position index, and construct the topological distribution sequence of mismatch characteristics.

[0145] Furthermore, the mismatch characteristics of each battery cell in the battery cell set are arranged according to the topological position index to construct a topological distribution sequence of mismatch characteristics.

[0146] In some embodiments, the mismatch characteristics of each battery cell in the battery cell set are arranged sequentially according to the ascending order of their topological location indices. The topological distribution sequence of the mismatch characteristics is a one-dimensional sequence, where the sequence index corresponds to the topological location index, and the sequence value corresponds to the mismatch characteristic of the battery cell at that topological location.

[0147] S53: Identify the extreme values ​​of the topological distribution sequence of mismatched features in the topological location direction, and identify the topological location index where the mismatched features exhibit local extreme values.

[0148] Furthermore, extreme values ​​are identified in the topological location direction of the topological distribution sequence of mismatch feature quantities.

[0149] In some embodiments, for each topological location in the topological distribution sequence of mismatch features, the relationship between the mismatch feature of that topological location and the mismatch feature of its adjacent topological locations is analyzed. If the mismatch feature of that topological location is greater than the mismatch feature of its adjacent topological locations, then that topological location exhibits a local extremum of the mismatch feature.

[0150] It should be noted that the degraded source battery cell disrupts the response coordination with its topologically adjacent battery cells in the series topology. According to step S2, the first coordination characteristic of the degraded source battery cell deviates from the normal range, and the disruption of the response coordination of the degraded source battery cell to its topologically adjacent battery cells also affects the first coordination characteristic of the topologically adjacent battery cells. According to step S3, the first coordination characteristic is a component value of the first spatial dimension in the multidimensional feature space, and the first spatial dimension component values ​​of the degraded source battery cell and its topologically adjacent battery cells deviate. According to step S4, the positional deviation in the multidimensional feature space leads to the deviation of the mismatch characteristic quantity. Therefore, the degree of deviation of the mismatch characteristic quantity of the degraded source battery cell is the greatest, and the mismatch characteristic quantity of its topologically adjacent battery cells is also affected, forming a mismatch characteristic quantity distribution pattern centered on the degraded source. The topological position of the degraded source battery cell shows a local extreme value of the mismatch characteristic quantity.

[0151] S54: Identify the cell in which the topological location index of the mismatch feature quantity shows a local extremum as the deterioration source cell.

[0152] Furthermore, the battery cells where the topological location index of the mismatch feature quantity exhibits local extrema are identified as the deterioration source battery cells.

[0153] In some embodiments, the corresponding battery cell is determined from the topological location index of the local extremum identified in step S53, and the battery cell is the deterioration source battery cell.

[0154] Furthermore, after identifying the deteriorated battery cell, the method further includes:

[0155] The multiple operating condition change events are divided into multiple time windows according to their chronological order. In some embodiments, each time window contains three consecutive operating condition change events that occur in chronological order.

[0156] Understandably, as described in step S1, the three consecutive operating condition change events are used to construct a cross-event response dataset and extract second-order change features. Multiple operating condition change events during the continuous operation of the energy storage power station are divided chronologically: the first to third operating condition change events constitute the first time window, the fourth to sixth operating condition change events constitute the second time window, the seventh to ninth operating condition change events constitute the third time window, and so on. The time windows do not overlap, and each time window independently contains three consecutive operating condition change events.

[0157] For each of the multiple time windows, the process of constructing a cross-event response dataset, extracting the first coordinated feature, extracting the second coordinated feature, extracting the third coordinated feature, constructing multi-dimensional features, and obtaining the mismatch feature quantity is executed to obtain the mismatch feature quantity of the degraded source battery cell in each time window.

[0158] The mismatch characteristics of the deteriorated battery cells in each time window are arranged in chronological order to construct a temporal evolution sequence of mismatch characteristics.

[0159] Furthermore, after constructing the temporal evolution sequence of the mismatch features, the method further includes:

[0160] A degradation process prediction model is established based on the temporal evolution sequence of the mismatch features.

[0161] In some embodiments, the degradation process prediction model is established by fitting a functional relationship between the mismatch feature and the time window in the time-series evolution sequence of the mismatch feature. The functional relationship reflects the evolution of the mismatch feature over time. Optionally, the functional relationship can be a linear function, a polynomial function, an exponential function, or a logarithmic function.

[0162] It should be noted that, unlike existing technologies that rely on a single operating parameter to predict battery life, this embodiment establishes a degradation process prediction model based on the temporal evolution of mismatch characteristics. The mismatch characteristics integrate the coordination features of the degradation source battery cell in three dimensions: topological space, frequency domain space, and temporal evolution. As described in steps S2 to S4, the mismatch characteristics reflect the comprehensive degradation degree of the degradation source battery cell in multiple physical processes.

[0163] The degradation process prediction model is used to predict the mismatch characteristics of future time windows. Specifically, the index of the future time window is substituted into the functional relationship to calculate the predicted mismatch characteristics corresponding to that time window. For example, for the nth future time window, the time window index n is substituted into the functional relationship to obtain the predicted mismatch characteristics of the nth time window.

[0164] A failure threshold for mismatch features is set, and time windows that reach the failure threshold are identified from the predicted mismatch features. Specifically, for each future time window, the predicted mismatch feature for that time window is compared with the failure threshold. When the predicted mismatch feature is not less than the failure threshold, that time window is identified as the time window that reaches the failure threshold.

[0165] In some embodiments, the mismatch characteristic quantity failure threshold is determined based on the statistical characteristics of the mismatch characteristics of healthy battery cells in the battery cell set. As described in step S4, healthy battery cells cluster in a multidimensional feature space to form feature regions, and the mismatch characteristics of healthy battery cells are within the normal range. Specifically, the mean and standard deviation of the mismatch characteristics of healthy battery cells in the battery cell set are calculated, and the mismatch characteristic quantity failure threshold is set to the mean plus a certain multiple of the standard deviation. For example, the mismatch characteristic quantity failure threshold can be set to the mean of the mismatch characteristics of healthy battery cells plus 3 times the standard deviation. When the predicted mismatch characteristic quantity of the deteriorating source battery cell exceeds this threshold, the deteriorating source battery cell is considered to have reached a failure state.

[0166] The time window when the failure threshold of the mismatch characteristic quantity is reached is determined as the predicted failure time node of the deterioration source battery cell.

[0167] Example 2 is an embodiment of the present invention. This embodiment provides a state monitoring system for an energy storage power station, including: a data acquisition unit, used to acquire electrical parameter response data of a set of battery cells connected in series under multiple operating condition change events, and to construct a cross-event response dataset from the electrical parameter response data under multiple operating condition change events;

[0168] The coordination feature extraction unit is used to extract the first coordination feature of the battery cells in the battery cell set based on the series electrical connection relationship in the first dimension, and to extract the second coordination feature and the third coordination feature of the battery cells in the battery cell set in the cross-event response dataset in the second and third dimensions, respectively.

[0169] A multi-dimensional feature construction unit is used to construct the first coordination feature, the second coordination feature, and the third coordination feature into multi-dimensional features of the battery cells in the battery cell set.

[0170] The mismatch measurement unit is used to measure the mismatch of the multidimensional features of the battery cells in the battery cell set and obtain the mismatch feature quantity of the battery cells in the battery cell set.

[0171] The degradation source identification unit is used to identify degradation source battery cells based on the distribution of mismatch characteristics of battery cells in the topological space according to the series electrical connection relationship.

[0172] This embodiment also provides an electronic device applicable to a state monitoring method for an energy storage power station, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the state monitoring method for an energy storage power station as proposed in the above embodiment.

[0173] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements a state monitoring method for an energy storage power station as proposed in the above embodiments.

[0174] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for monitoring the status of an energy storage power station proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0175] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0176] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for monitoring the condition of an energy storage power station, characterized in that, include: Acquire electrical parameter response data of a battery cell assembly connected in series under multiple operating condition change events, and construct a cross-event response dataset from the electrical parameter response data under multiple operating condition change events. Based on the series electrical connection relationship, the first coordination feature of the battery cells in the battery cell set is extracted from the cross-event response dataset in the first dimension. The second coordination feature and the third coordination feature of the battery cells in the battery cell set are extracted from the cross-event response dataset in the second and third dimensions, respectively. The first coordination feature, the second coordination feature, and the third coordination feature are constructed as multidimensional features of the battery cells in the battery cell set; Mismatch measurement is performed on the multidimensional features of the battery cells in the battery cell set to obtain the mismatch feature quantity of the battery cells in the battery cell set; Based on the series electrical connection relationship, the deterioration source battery cells are identified according to the distribution of the mismatch characteristics of the battery cells in the topological space. The extraction of the first coordination feature of battery cells in the battery cell set based on the series electrical connection relationship in the first dimension of the cross-event response dataset includes: Construct a topological adjacency matrix based on series electrical connections; For each battery cell in the battery cell set, obtain the topological adjacency position index of each battery cell from the topological adjacency matrix; The response data of each battery cell and the topological adjacency position of each battery cell are extracted from the cross-event response dataset, and a topological space vector field structure is constructed in the topological space. The coordination features of the topological space vector field structure are extracted to obtain the first coordination features of each battery cell; The step of measuring the mismatch of multidimensional features of battery cells in the battery cell set to obtain the mismatch feature quantity of battery cells in the battery cell set includes: In the multidimensional feature space, the spatial position of each battery cell is determined based on the multidimensional features of each battery cell in the battery cell set. Spatial position deviation pattern recognition is performed on the spatial position of each battery cell in the battery cell set. Spatial position deviation pattern recognition is achieved by analyzing the distribution deviation characteristics of the spatial position of each battery cell in a multi-dimensional feature space. The mismatch characteristics of each battery cell are obtained from the spatial position deviation pattern recognition results; The method of identifying deterioration source battery cells based on the distribution of mismatch characteristics of battery cells in the topological space according to the series electrical connection relationship includes: The topological location index of each battery cell in the battery cell set is determined from the series electrical connection relationship; The mismatch feature quantities of each battery cell in the battery cell set are arranged according to the topological position index to construct a topological distribution sequence of mismatch feature quantities; The topological distribution sequence of the mismatch feature quantity is subjected to extreme value identification in the topological location direction, and the topological location index where the mismatch feature quantity exhibits local extreme values ​​is identified. The cell in which the topological location index of the mismatch feature quantity shows a local extremum is identified as the cell source of degradation. Based on the frequency domain response data of each battery cell in the battery cell set under multiple operating condition changes, the second coordination characteristics of each battery cell are obtained. Based on the smoothness of the evolution trajectory of each battery cell in the battery cell set over time, the third coordination characteristic of each battery cell is obtained.

2. The energy storage power station status monitoring method as described in claim 1, characterized in that, The step of constructing the first coordination feature, the second coordination feature, and the third coordination feature into multidimensional features of the battery cells in the battery cell set includes: In the multidimensional feature space, the first coordination feature of each battery cell in the battery cell set is determined as the component value of the first spatial dimension, the second coordination feature is determined as the component value of the second spatial dimension, and the third coordination feature is determined as the component value of the third spatial dimension. Based on the component values ​​of the first spatial dimension, the component values ​​of the second spatial dimension, and the component values ​​of the third spatial dimension, the spatial position coordinates of each battery cell are determined in the multidimensional feature space.

3. The energy storage power station status monitoring method as described in claim 2, characterized in that, After identifying the deteriorated battery cell, the process also includes: The multiple operating condition change events are divided into multiple time windows according to their chronological order. For each of the multiple time windows, the following processes are executed: cross-event response dataset construction, first coordination feature extraction, second coordination feature extraction, third coordination feature extraction, multi-dimensional feature construction, and mismatch feature quantity acquisition, to obtain the mismatch feature quantity of the degraded source battery cell in each time window. The mismatch characteristics of the degraded battery cells in the multiple time windows are arranged in chronological order to construct a temporal evolution sequence of mismatch characteristics.

4. The energy storage power station status monitoring method as described in claim 3, characterized in that, After constructing the temporal evolution sequence of the mismatch feature, the method further includes: A degradation process prediction model is established based on the temporal evolution sequence of the mismatch features; The degradation process prediction model is used to predict the mismatch characteristics of future time windows. Set a failure threshold for mismatch features, and identify the time window that reaches the failure threshold from the predicted mismatch features; The time window when the failure threshold of the mismatch characteristic quantity is reached is determined as the predicted failure time node of the deterioration source battery cell.

5. A condition monitoring system for an energy storage power station, employing the condition monitoring method for an energy storage power station as described in any one of claims 1 to 4, characterized in that, include: The data acquisition unit is used to acquire the electrical parameter response data of a set of battery cells connected in series under multiple operating condition change events, and to construct the electrical parameter response data under multiple operating condition change events into a cross-event response dataset. The coordination feature extraction unit is used to extract the first coordination feature of the battery cells in the battery cell set based on the series electrical connection relationship in the first dimension of the cross-event response dataset, and to extract the second coordination feature and the third coordination feature of the battery cells in the battery cell set in the second dimension and the third dimension, respectively. A multi-dimensional feature construction unit is used to construct the first coordination feature, the second coordination feature, and the third coordination feature into multi-dimensional features of the battery cells in the battery cell set; The mismatch measurement unit is used to measure the mismatch of the multidimensional features of the battery cells in the battery cell set and obtain the mismatch feature quantity of the battery cells in the battery cell set. The degradation source identification unit is used to identify the degradation source battery cells based on the series electrical connection relationship and the distribution of the mismatch characteristics of the battery cells in the battery cell set in the topological space.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the energy storage power station status monitoring method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the energy storage power station status monitoring method according to any one of claims 1 to 4.

Images