Method and system for detecting abnormal data of battery of energy storage power station

Abstract

The invention provides an energy storage power station battery abnormal data detection method and system. The method comprises the following steps: selecting data of a preset time slice from energy storage power station battery operation data to form a plurality of reference data slices; screening out normal reference data of the to-be-tested variable data from the plurality of reference data fragments through clustering; the to-be-measured variable data is any one of monomer temperature, monomer current, monomer voltage or environment temperature; solving probability distribution of normal reference data by using a Gaussian mixture model to form a normal reference probability distribution set; collecting to-be-measured data from the target energy storage power station, and quantifying the difference between the actual distribution of the to-be-measured data and each element in the normal reference probability distribution set by using Jassen-Shannon divergence; and when the difference exceeds a threshold value, determining that the to-be-detected data is abnormal data. According to the method and the system for detecting the abnormal data of the battery in the energy storage power station, the abnormal data of the battery in the energy storage power station can be quickly identified, and potential safety hazards are effectively prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of battery anomaly detection, and in particular to a method and system for detecting abnormal data of batteries in an energy storage power station. Background Art

[0002] In the context of energy storage battery application safety, it is of great significance to use big data science to mine the development law of energy storage battery data. As a popular direction of data science, abnormal data detection technology can be used for data cleaning to reduce the interference of abnormal data on data analysis or for early warning protection to improve the safety of equipment application. Therefore, establishing effective anomaly detection technology oriented to application needs has become one of the key issues in data research. As a common application mode of anomaly detection technology, data cleaning is the basis and prerequisite of data mining. In the anomaly detection research of energy storage batteries, anomaly detection tends to analyze the health status and fault conditions of batteries, focusing specifically on the estimation of battery health characteristics and detection and diagnosis methods. In the characterization of battery health characteristics, research is carried out around battery cell performance, health factor extraction methods, and battery state estimation. These studies provide effective data and feature sources for battery anomaly detection and are the basis for the implementation and application of battery abnormal state detection. When the battery compartment of the energy storage power station performs abnormally, the power station has low operating efficiency and cannot achieve peak load regulation, standby and other functions, affecting the stable operation of the power grid. In addition, the abnormality of the battery compartment will cause the battery to be overcharged or over-discharged, thereby accelerating battery aging and shortening the service life of the equipment. When the battery is abnormal, it may cause safety accidents such as short circuit and fire in the battery.

[0003] At present, the abnormal data detection methods of energy storage power station batteries can be divided into two categories: model-driven methods and data-driven methods. The basic principle of the model-driven method is to compare the simulation data of the energy storage power station with the measured values, and issue a fault alarm when there is a large difference between the two. The key steps include: establishing a battery simulation model and estimating parameters, determining the comparison quantity, and selecting indicators to measure the difference. However, the model-driven method is too dependent on the completeness and accuracy of the data. The key steps of the data-driven method include: determining the feature quantity that is beneficial to the task, selecting a suitable data-driven model, and obtaining labeled fault data. In general, the data-driven method needs to ensure the availability of relevant data sets for the normal operation and fault operation of the battery, and the data-driven method does not have the ability to detect fault types that have not been learned in advance. Summary of the invention

[0004] In order to solve the above problems, the object of the present invention is to provide a method and system for detecting abnormal data of batteries in an energy storage power station.

[0005] A method for detecting abnormal data of a battery in an energy storage power station, comprising:

[0006] Step 1: Obtaining the battery operation data of the energy storage power station; the battery operation data of the energy storage power station includes: single cell temperature, single cell current, single cell voltage and ambient temperature;

[0007] Step 2: Selecting data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments;

[0008] Step 3: Filtering out normal reference data of the variable data to be measured from the multiple reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature;

[0009] Step 4: Use the Gaussian mixture model to solve the probability distribution of normal reference data to form a normal reference probability distribution set;

[0010] Step 5: Collect the data to be tested from the target energy storage power station, and use Jason-Shannon divergence to quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set;

[0011] Step 6: When the difference exceeds the preset threshold, the data to be tested is abnormal data.

[0012] Preferably, the step 4: using a Gaussian mixture model to solve the probability distribution of normal reference data to form a normal reference probability distribution set includes:

[0013]

[0014]

[0015] Among them, π k is the weight of the kth Gaussian component, is the probability density function of the kth Gaussian distribution, Σ k is the covariance matrix of the kth Gaussian component, d is the dimension of the data, μ k is the mean vector of the kth Gaussian component, p j (x) is the probability distribution of the jth normal reference data.

[0016] Preferably, the step 5: collecting the data to be tested from the target energy storage power station, and using Jason-Shannon divergence to quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set, includes:

[0017] Step 5.1: Collecting the data to be measured in a preset time segment from the target energy storage power station;

[0018] Step 5.2: Calculate the probability distribution of the data to be tested;

[0019] Step 5.3: Calculate the JS divergence between each element in the normal reference probability distribution set and the probability distribution of the data to be tested;

[0020] Step 5.4: Calculate the joint anomaly probability based on JS divergence;

[0021] Step 5.5: When the combined abnormal probability is greater than the preset ratio, the data to be tested is determined to be abnormal data.

[0022] Preferably, in step 5.3, the calculation formula of JS divergence is:

[0023]

[0024] Among them, D JS (p(x)∥p j ) is the probability distribution p(x) of the test data and the pth probability distribution of the normal reference j The JS divergence between elements, Represents the test data distribution p(x) and the reference distribution p j The fusion distribution of KL (p(x)∥M j ) is the distribution of the data to be tested p(x) and the fusion distribution M j The KL divergence between .

[0025] Preferably, in step 5.4, the joint abnormality probability calculation formula is:

[0026]

[0027] Among them, ∈ is the threshold, 1{D JS (p(x)∥p j )>∈} is an indicator function, which is 1 when the JS divergence is greater than the threshold, otherwise it is 0. is the number of elements in the normal reference probability distribution, P ab is the joint anomaly probability.

[0028] The present invention also provides an energy storage power station battery abnormal data detection system, comprising:

[0029] An operation data acquisition module is used to acquire the operation data of the energy storage power station battery; the operation data of the energy storage power station battery includes: single cell temperature, single cell current, single cell voltage and ambient temperature;

[0030] A data selection module, used to select data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments;

[0031] A data screening module, used for screening out normal reference data of the variable data to be measured from a plurality of reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature;

[0032] A probability distribution calculation module, used to solve the probability distribution of normal reference data using a Gaussian mixture model to form a normal reference probability distribution set;

[0033] A difference comparison module is used to collect the data to be tested from the target energy storage power station and quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set by using Jason-Shannon divergence;

[0034] The abnormal data detection module is used to mark the data to be tested as abnormal data when the difference exceeds a preset threshold.

[0035] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and wherein the computer program, when executed by the processor, implements the steps in the above-mentioned method for detecting abnormal data of batteries in an energy storage power station.

[0036] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for detecting abnormal data of batteries in an energy storage power station are implemented.

[0037] The beneficial effect of the abnormal data detection method and system for batteries in energy storage power stations provided by the present invention is that: compared with the prior art, the present invention can quickly identify abnormal data of batteries in energy storage power stations by quantifying the difference between the data to be tested and the normal reference probability distribution using Jason-Shannon divergence, thereby effectively preventing potential safety hazards.

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A flow chart of a method for detecting abnormal data of batteries in an energy storage power station provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0041] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0042] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0043] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0044] See also Figure 1 , a method for detecting abnormal data of batteries in energy storage power stations, comprising:

[0045] Step 1: Obtaining the battery operation data of the energy storage power station; the battery operation data of the energy storage power station includes: single cell temperature, single cell current, single cell voltage and ambient temperature;

[0046] Step 2: Selecting data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments;

[0047] Step 3: Filtering out normal reference data of the variable data to be measured from the multiple reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature;

[0048] Step 4: Use the Gaussian mixture model to solve the probability distribution of normal reference data to form a normal reference probability distribution set;

[0049] In step 4, the probability distribution is calculated as:

[0050]

[0051] Among them, π k is the weight of the kth Gaussian component, is the probability density function of the kth Gaussian distribution, Σ k is the covariance matrix of the kth Gaussian component, d is the dimension of the data, μ k is the mean vector of the kth Gaussian component, p j (x) is the probability distribution of the jth normal reference data.

[0052] Step 5: Collect the data to be tested from the target energy storage power station, and use Jason-Shannon divergence to quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set;

[0053] Furthermore, step 5 includes:

[0054] Step 5.1: Collecting the data to be measured in a preset time segment from the target energy storage power station;

[0055] Step 5.2: Calculate the probability distribution of the data to be tested;

[0056] Step 5.3: Calculate the JS divergence between each element in the normal reference probability distribution set and the probability distribution of the data to be tested;

[0057] In step 5.3, the calculation formula of JS divergence is:

[0058]

[0059] Among them, D JS (p(x)∥p j ) is the probability distribution p(x) of the test data and the pth probability distribution of the normal reference j The JS divergence between elements,

[0060] Step 5.4: Calculate the joint anomaly probability based on JS divergence;

[0061] In step 5.4, the formula for calculating the joint anomaly probability is:

[0062]

[0063] Among them, ∈ is the threshold, 1{D JS (p(x)∥p j )>∈} is an indicator function, which is 1 when the JS divergence is greater than the threshold, otherwise it is 0. is the number of elements in the normal reference probability distribution, P ab is the joint anomaly probability.

[0064] Step 5.5: When the combined abnormal probability is greater than the preset ratio, the data to be tested is determined to be abnormal data.

[0065] Step 6: When the difference exceeds the preset threshold, the data to be tested is abnormal data.

[0066] The present invention further describes the abnormal data detection method of the energy storage power station battery in combination with specific embodiments of the present invention.

[0067] The present invention adopts Gaussian mixture model (GMM) to characterize uncertainty, solve the probability distribution of normal data, and use Jensen-Shannon (JS) divergence to characterize the difference between the actual distribution of the test data to be tested and the elements in the normal reference distribution set, so as to realize the abnormal data detection of the battery of the energy storage power station. First, select normal data, sample one hour of fragment data through the sliding window algorithm, construct multiple reference data fragments, use the data other than the test data in the current test data, such as the test data is the monomer temperature, use other data as covariates for clustering, filter out the monomer temperature reference under normal circumstances, select appropriate normal data, and divide the reference group. In order to improve the model, analyze the prerequisites for feature extraction, filter the data that meets the feature extraction conditions, so as to improve the accuracy of subsequent feature calculation. Then, use the Gaussian mixture model to solve the probability distribution of normal data and form a normal reference distribution set. Use Jason-Shannon divergence to characterize the difference between the actual distribution of the test data to be tested and the elements in the normal reference distribution set. If most JS divergences are much greater than 0, it is considered that the data is abnormal.

[0068] The specific steps are as follows:

[0069] (1) Obtaining battery operation data of energy storage power station

[0070] (2) Construct reference data segments. Select normal monomer current, monomer voltage, monomer temperature, ambient temperature and other data, sample a period of time segment data through the sliding window algorithm, and construct multiple reference data segments X i . Set the window size to W (e.g. 5 minutes), and the data sequence to: X = {x1, x2, .., x t}Where T is the total length of the data, the time segment extracted by the sliding window is:

[0071] X i ={x i ,x i+1 ,..,x i+w-1}where i=1,2,…,T-W+1

[0072] Among them, x i Including monomer temperature Tcell , single cell current I cell , single cell voltage V cell , ambient temperature T env And other variables.

[0073] (3) Construct normal reference data for the variable data to be measured through clustering. Use the other variable data in the reference data except the variable to be measured as covariates for clustering, screen out the monomer temperature reference under normal conditions, and select N groups of appropriate normal data. K-Means clustering is used as an example for explanation.

[0074] First, construct a reference clustering data set. Suppose the variable to be measured is the monomer temperature T cell , then take other covariates to form the reference clustering data set

[0075]

[0076] Among them, DX i represents the covariate data within the window and does not include the variable to be tested.

[0077] Secondly, the K-Means clustering algorithm is used to cluster the reference data set DX i Clustering is performed to find out the temperature of the normal monomer. K-Means clustering algorithm is used to cluster the reference data set DX i Perform clustering to find out the temperature T cell Related data fragments. The basic steps of the K-Means clustering algorithm are as follows:

[0078] 1. Initialization: The user selects K initial cluster centers μ1, μ2, ..., μ k

[0079] 2. Assign data: Assign each data point DX in the data set i Assign to the cluster center μ that is closest to it k , so that each data point belongs to a cluster.

[0080]

[0081] Among them, c i It is the data point DX i The cluster to which it belongs, μ k is the cluster center of cluster k.

[0082] 3. Update center: Update the cluster center according to the member data points of each cluster,

[0083]

[0084] Among them C k are all the data points in cluster k.

[0085] 4. Iteration: Repeat steps 2 and 3 until the cluster center is stable or the maximum number of iterations is reached.

[0086] Then, use the covariate data to obtain N groups of normal reference data for the variable data to be tested. Take the covariate data of the W-length window of the data to be tested as input, use K-Means clustering reasoning, and obtain the cluster μ corresponding to the data k All data points C k If C k If the number of data points in the equation is less than N, the covariate data is considered abnormal, and the abnormality judgment of the variable to be measured is terminated.

[0087] Finally, N groups of data are randomly selected to form N groups of normal reference data. reference groups, where K is a parameter set by the user.

[0088] (4) Use the Gaussian mixture model to solve the probability distribution of the normal data of the measured variables and form a normal reference distribution set.

[0089] First, calculate The joint probability density function of the Gaussian mixture model of each reference group in the reference groups. Let the data sets of these reference groups be X ref :

[0090]

[0091] For each data point x ref_i , its probability distribution is the weighted sum of all Gaussian components. Specifically expressed as:

[0092]

[0093] Among them, π k is the weight of the kth Gaussian component, satisfying is the probability density function of the kth Gaussian distribution.

[0094]

[0095] Among them, μ k is the mean vector of the kth Gaussian component, Σ k is the covariance matrix of the kth Gaussian component, and d is the dimension of the data.

[0096] Then, the maximum expectation (EM) algorithm is used to estimate the probability distribution parameters of the data: Use the EM (Expectation Maximization) algorithm to estimate the parameters of the model, that is, to estimate the mean μ of each Gaussian component k , covariance matrix Σ k and weight π kThe reference group data can be modeled by a Gaussian mixture model, and the probability distribution set P of the normal reference group data can be obtained.

[0097]

[0098] (5) Use the Jensen-Shannon (JS) divergence to characterize the actual distribution of the test data variable and the elements in the normal reference distribution set P If most JS divergences are much greater than 0, the data is considered abnormal.

[0099] First, take the window data of the variable to be measured W and calculate its probability distribution p(x). For each reference distribution Calculate the JS divergence D of the test data distribution p(x) JS (p(x)∥p j ):

[0100]

[0101] in,

[0102] Then, calculate the values ​​of all JS divergences and set a threshold ∈. Calculate the joint anomaly probability P ab , if P ab If the ratio is greater than a certain ratio, the data to be tested is considered abnormal data. ab The calculation method is:

[0103]

[0104] Among them, 1{D JS (p(x)∥p j )>∈} is an indicator function, which is 1 when the JS divergence is greater than the threshold, and 0 otherwise.

[0105] (7) Provide fixed-point positioning warning for batteries with abnormal data.

[0106] Compared with the prior art, the advantages of this application are as follows:

[0107] (1) Simple implementation: GMM is relatively simple to implement and does not require complex calculations, making it suitable for rapid deployment and application.

[0108] (2) Strong practical application capabilities: GMM can effectively process data sets containing missing values, which is consistent with the actual power station operation data. It can also characterize the distribution of nonlinear system data, which is consistent with the actual working characteristics of batteries.

[0109] (3) Strong robustness: Compared with some other models, GMM is more robust to outliers. The accuracy of the model will not be significantly affected by a few outliers. It can quickly identify abnormal data of batteries in energy storage power stations and effectively prevent potential safety hazards.

[0110] The present invention also provides an energy storage power station battery abnormal data detection system, comprising:

[0111] An operation data acquisition module is used to acquire the operation data of the energy storage power station battery; the operation data of the energy storage power station battery includes: single cell temperature, single cell current, single cell voltage and ambient temperature;

[0112] A data selection module, used to select data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments;

[0113] A data screening module, used for screening out normal reference data of the variable data to be measured from a plurality of reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature;

[0114] A probability distribution calculation module, used to solve the probability distribution of normal reference data using a Gaussian mixture model to form a normal reference probability distribution set;

[0115] A difference comparison module is used to collect the data to be tested from the target energy storage power station and quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set by using Jason-Shannon divergence;

[0116] The abnormal data detection module is used to mark the data to be tested as abnormal data when the difference exceeds a preset threshold.

[0117] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and characterized in that when the computer program is executed by the processor, the steps in the above-mentioned method for detecting abnormal data of batteries in an energy storage power station are implemented. Compared with the prior art, the beneficial effects of a computer-readable storage medium provided by the present invention are the same as the beneficial effects of the method for detecting abnormal data of batteries in an energy storage power station described in the above-mentioned technical solution, and are not elaborated herein.

[0118] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for detecting abnormal data of batteries in an energy storage power station are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the method for detecting abnormal data of batteries in an energy storage power station described in the above-mentioned technical solution, which will not be elaborated here.

[0119] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for detecting abnormal data of batteries in energy storage power stations, characterized in that: include: Step 1: Obtain the battery operation data of the energy storage power station; The energy storage power station battery operation data includes: single cell temperature, single cell current, single cell voltage and ambient temperature; Step 2: Selecting data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments; Step 3: Filtering out normal reference data of the variable data to be measured from the multiple reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature; Step 4: Use the Gaussian mixture model to solve the probability distribution of normal reference data to form a normal reference probability distribution set; Step 5: Collect the data to be tested from the target energy storage power station, and use Jason-Shannon divergence to quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set; Step 6: When the difference exceeds the preset threshold, the data to be tested is abnormal data.

2. A method for detecting abnormal data of batteries in an energy storage power station according to claim 1, characterized in that: The step 4: using a Gaussian mixture model to solve the probability distribution of normal reference data to form a normal reference probability distribution set, includes: Where x represents a data point in the normal reference data, π k is the weight of the kth Gaussian component, is the probability density function of the kth Gaussian distribution, Σ k is the covariance matrix of the kth Gaussian component, d is the dimension of the data, μ k is the mean vector of the kth Gaussian component, p j (x) is the probability distribution of the jth normal reference data.

3. The method for detecting abnormal data of batteries in an energy storage power station according to claim 2, characterized in that: The step 5: collecting the data to be tested from the target energy storage power station, and using Jason-Shannon divergence to quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set, including: Step 5.1: Collecting the data to be measured in a preset time segment from the target energy storage power station; Step 5.2: Calculate the probability distribution of the data to be tested; Step 5.3: Calculate the JS divergence between each element in the normal reference probability distribution set and the probability distribution of the data to be tested; Step 5.4: Calculate the joint anomaly probability based on JS divergence; Step 5.5: When the combined abnormal probability is greater than the preset ratio, the data to be tested is determined to be abnormal data.

4. The method for detecting abnormal data of batteries in an energy storage power station according to claim 3, characterized in that: In step 5.3, the calculation formula of JS divergence is: Among them, D JS (p(x)∥p j ) is the probability distribution p(x) of the test data and the pth probability distribution of the normal reference j The JS divergence between elements, Represents the probability distribution of the data to be tested p(x) and p j The fusion distribution of KL (p(x)∥M j ) is the probability distribution p(x) of the data to be tested and the fusion distribution M j The KL divergence between KL (p j ∥M j ) is the reference distribution p j and fusion distribution M j The KL divergence between .

5. The method for detecting abnormal data of batteries in an energy storage power station according to claim 4, characterized in that: In step 5.4, the formula for calculating the joint anomaly probability is: Among them, ∈ is the threshold, 1{D JS (p(x)∥p j )>∈} is an indicator function, which is 1 when the JS divergence is greater than the threshold, otherwise it is 0. is the number of elements in the normal reference probability distribution, P ab is the joint anomaly probability.

6. An energy storage power station battery abnormal data detection system, characterized in that: include: An operation data acquisition module is used to obtain the battery operation data of the energy storage power station; The energy storage power station battery operation data includes: single cell temperature, single cell current, single cell voltage and ambient temperature; A data selection module, used to select data of a preset time segment from the battery operation data of the energy storage power station to form a plurality of reference data segments; A data screening module, used for screening out normal reference data of the variable data to be measured from a plurality of reference data segments by clustering; the variable data to be measured is any one of the cell temperature, cell current, cell voltage or ambient temperature; A probability distribution calculation module, used to solve the probability distribution of normal reference data using a Gaussian mixture model to form a normal reference probability distribution set; A difference comparison module is used to collect the data to be tested from the target energy storage power station and quantify the difference between the actual distribution of the data to be tested and each element in the normal reference probability distribution set by using Jason-Shannon divergence; The abnormal data detection module is used to mark the data to be tested as abnormal data when the difference exceeds a preset threshold.

7. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method for detecting abnormal data of batteries in an energy storage power station as described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method for detecting abnormal data of batteries in an energy storage power station as described in any one of claims 1 to 5 are implemented.

Images