Insulation anomaly detection method, system and device for energy storage system

Abstract

The application relates to the technical field of electric variable abnormality detection, in particular to an insulation abnormality detection method, system and device for an energy storage system, and specifically comprises the following: the application constructs data vectors at each moment by collecting electric variable data and environmental data of the battery at each moment, analyzes the data vectors collected under each time frame, constructs environmental disturbance adaptive weights of the electric variable, performs clustering through an improved K-means algorithm, and obtains two clustering clusters; determines an abnormal data clustering cluster based on the difference between the insulation resistances corresponding to the elements in the two clustering clusters; determines an insulation abnormality position through insulation positioning technology for all data corresponding to the moment in the abnormal data clustering cluster; the accuracy of the algorithm for real insulation abnormality detection is improved, the false insulation abnormality misjudgment phenomenon caused by environmental disturbance is avoided, and the stability and efficiency of the battery management system for insulation abnormality detection are improved.

Description

Technical Field

[0001] This application relates to the field of electrical variable anomaly detection technology, specifically to insulation anomaly detection methods, systems, and devices for energy storage systems. Background Technology

[0002] With the rapid development of new energy technologies, the number of new energy vehicles on the road has experienced explosive growth. As a core component of new energy vehicles, the energy storage system is responsible for storing and releasing electrical energy, and is a crucial link in ensuring the vehicle's power performance and driving safety. Energy storage systems often use high-voltage, large-capacity battery packs, with voltage levels far exceeding the safe range that the human body can withstand. Insulation failures or leakage accidents can lead to serious safety incidents. Therefore, accurate and efficient insulation testing of energy storage systems, especially power battery packs, is of paramount importance for ensuring user safety and improving product reliability.

[0003] Current technologies typically determine insulation status by detecting electrical variables within the battery pack, such as insulation resistance, voltage, current changes, and battery temperature. However, this method generally fails to adequately consider the influence of the vehicle's environmental conditions. In practical applications, new energy vehicles often operate in various complex and harsh environments, such as high temperature, low temperature, and high humidity. These external environmental factors often cause changes in the physical properties of the insulation materials themselves, leading to dynamic fluctuations in their insulation performance and temporary outliers in the electrical variable data measured by the battery management system. Traditional methods for analyzing electrical variable data to detect insulation anomalies, such as K-means clustering of electrical variable data samples using Euclidean distance, detect all outliers as insulation anomalies, making it difficult to effectively distinguish between dynamic insulation anomalies caused by environmental interference and actual insulation faults. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method, system, and device for detecting insulation anomalies in energy storage systems. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a method for detecting insulation anomalies in an energy storage system, the method comprising the following steps:

[0006] Collect environmental electrical and non-environmental electrical variable data of the vehicle's energy storage battery at various times, as well as environmental data of the vehicle's location;

[0007] Data vectors for each time point are constructed using all environmental electrical variable data, all non-environmental electrical variable data, and all environmental data. Clustering algorithms are used to cluster the data vectors for all time points. During the clustering process, the environmental disturbance adaptive coefficients for each type of environmental electrical variable at each time point are determined based on the differences between each type of environmental data and the corresponding standard environmental data at each time point during the first iteration of clustering.

[0008] The environmental disturbance adaptive coefficient of any environmental electrical variable at each moment of each iteration and the corresponding iteration number are adjusted to adjust the environmental disturbance adaptive coefficient of any environmental electrical variable at each moment of the next iteration.

[0009] Based on the differences in environmental electrical variables and non-environmental electrical variables in the data vectors at any two time points, and combined with the environmental disturbance adaptive coefficients at each iteration, the distance between the data vectors at any two time points at each iteration is calculated; the distance is used as the distance metric in the clustering algorithm, and abnormal data clusters are determined based on the differences in data vectors between different clusters.

[0010] The location of insulation anomalies is determined by using insulation location technology to identify all data vectors and their corresponding times in the abnormal data clusters.

[0011] In one embodiment, the environmental data includes ambient temperature, humidity, and air pressure; the environmental electrical variables include voltage, current, insulation resistance, and capacitance.

[0012] In one embodiment, the process of obtaining the adaptive coefficients of environmental disturbances for each class of environmental electrical variables at each time step during the first iteration of clustering is as follows:

[0013] Based on the differences between various environmental data at each time point and the corresponding standard environmental data, the environmental dynamic coupling coefficient of insulation resistance and the environmental dynamic coupling coefficient of capacitance at each time point are calculated.

[0014] The environmental disturbance adaptive coefficients for each class of environmental electrical variables at each moment during the first iteration of clustering are determined based on the environmental dynamic coupling coefficients of insulation resistance and capacitance.

[0015] In one embodiment, the expressions for the environmental dynamic coupling coefficient of the insulation resistance and the environmental dynamic coupling coefficient of the capacitance at each time point are as follows:

[0016]

[0017] in, The environmental dynamic coupling coefficient of the insulation resistance at time t; ΔH is the absolute value of the difference between the ambient temperature and the standard laboratory temperature at time t; t ΔP is the absolute value of the difference between the ambient humidity at time t and the standard laboratory humidity;t α is the absolute value of the difference between the ambient air pressure and the standard laboratory air pressure at time t; r β r α and γ are both preset positive numbers. r +β r +γ=1;

[0018] ω(C t α is the environmental dynamic coupling coefficient of the capacitor at time t; c β c All are preset positive numbers, α c +β c =1; exp() is an exponential function with the natural constant as the base.

[0019] In one embodiment, the process of obtaining the adaptive coefficients of environmental disturbances for each type of environmental electrical variable at each time step during the first iteration is as follows:

[0020] The environmental dynamic coupling coefficient of the insulation resistance at each moment is used as the environmental disturbance adaptive coefficient of the voltage, current and insulation resistance at each moment during the first iteration, and the environmental dynamic coupling coefficient of the capacitor at each moment is used as the environmental disturbance adaptive coefficient of the capacitor at each moment during the first iteration.

[0021] In one embodiment, the process of obtaining the adaptive coefficient of environmental disturbance for any environmental electrical variable at each time point in the next iteration is as follows:

[0022] Let the adaptive coefficient of the environmental disturbance of the j-th environmental electrical variable at time t in the first iteration be denoted as... Let the adaptive coefficients of the environmental disturbance of the j-th environmental electrical variable at time t in the (k+1)th and kth iterations be denoted as follows:

[0023] when Less than the preset first threshold season Where k is the iteration number in the k-th iteration; K represents the total number of iterations; min() means taking the minimum value;

[0024] when Greater than or equal to the preset first threshold season

[0025] In one embodiment, the expression for the distance between data vectors at any two moments during each iteration is:

[0026] Wherein, D(X) s ,X t X is the data vector at time s during the current iteration. sWith the data vector X at time t t The distance between them; Ba is the set of all key electrical variables; B E Represents the set of all types of environmental electrical variables; Ba\B E Represent sets Ba and B E The difference set; X s (i), X t (i) represent the data values ​​of the i-th non-environmental electrical variable at time s and time t, respectively; X s (j), X t (j) represent the data values ​​of the j-th environmental electrical variable at time s and time t, respectively; ω s (j), ω t (j) represents the environmental disturbance adaptive coefficients of the j-th environmental electrical variable at time s and time t, respectively.

[0027] In one embodiment, the process of obtaining the abnormal data clusters is as follows:

[0028] Calculate the mean value of the insulation resistance in all data vectors within each cluster, and denote it as the first mean value; the cluster with the smallest first mean value is taken as the outlier cluster.

[0029] Secondly, embodiments of this application also provide an insulation anomaly detection system for an energy storage system, wherein the system stores a computer program, and when the computer program is executed by a processor, it implements the steps of the method described in the first aspect above.

[0030] Thirdly, embodiments of this application also provide an insulation anomaly detection device for an energy storage system, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method described in the first aspect above.

[0031] The embodiments of this application have at least the following beneficial effects:

[0032] This application constructs data vectors for each time frame by collecting electrical variable data and environmental data from the battery at various times. It analyzes the data vectors collected in each time frame, constructs adaptive weights for environmental disturbances of electrical variables, and uses an improved clustering algorithm for clustering. This allows for the differentiation and identification of insulation anomalies caused by faults and dynamic insulation anomalies caused by environmental interference. Compared to clustering algorithms that directly analyze outliers based on multiple electrical variable parameters, the adaptive weights for environmental disturbances consider the influence of external environmental data on changes in insulation electrical variable parameters. This avoids the situation where short-term abnormal fluctuations in electrical variables caused by changes in the external environment are falsely detected as insulation anomalies, improving the accuracy of the algorithm in detecting true insulation anomalies and avoiding false detections of insulation anomalies caused by environmental disturbances. This enhances the stability and efficiency of the battery management system in detecting insulation anomalies. Attached Figure Description

[0033] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 A flowchart illustrating the steps of an insulation anomaly detection method for an energy storage system provided in one embodiment of this application;

[0035] Figure 2 This is a flowchart illustrating the steps of an insulation anomaly detection method for energy storage systems. Detailed Implementation

[0036] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the insulation anomaly detection method, system, and apparatus for energy storage systems proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0038] The following description, in conjunction with the accompanying drawings, details the specific scheme of the insulation anomaly detection method, system, and device for energy storage systems provided in this application.

[0039] Please see Figure 1The diagram illustrates a flowchart of an insulation anomaly detection method for an energy storage system according to an embodiment of this application. The method includes the following steps:

[0040] Step S1: Collect environmental electrical variable data and non-environmental electrical variable data of the vehicle's energy storage battery at each time point, as well as environmental data of the vehicle's location.

[0041] (1) Key electrical variable data of the energy storage battery are collected in real time through the battery management system equipped in the new energy vehicle, including voltage, current, insulation resistance, capacitance value and battery temperature; environmental data of the vehicle's current location are collected in real time based on the on-board sensors, including ambient temperature, humidity and air pressure. All electrical variable data and environmental data are accompanied by accurate timestamp information to ensure data synchronization.

[0042] Data packets are created based on timestamp information, and key electrical variable data and environmental data collected at the same time are combined and stored.

[0043] The collected raw data is smoothed by normalization to bring it to the range of [0,1], thereby improving data quality and facilitating subsequent data analysis.

[0044] The data is processed to handle missing values. The data packets in the continuous time frame are checked for missing data. If missing data is found, the average value of the data in the next n frames is used as the interpolation to supplement the data in that frame. The value of n can be set by the implementer according to the implementation scenario. This application does not make any special limitation. In this embodiment, n is 2.

[0045] (2) Different electrical variables in the electrical variable data collected in this application are affected differently by the environment. Among them, voltage, current, insulation resistance and capacitance are greatly affected by environmental factors, while battery temperature is less affected by environmental factors. Therefore, in the embodiments of this application, voltage, current, insulation resistance and capacitance are regarded as environmental electrical variables, and battery temperature is regarded as a non-environmental electrical variable.

[0046] Step S2: Construct data vectors for each time moment using all environmental electrical variable data, all non-environmental electrical variable data, and all environmental data. Cluster the data vectors for all time moments using a clustering algorithm. During the clustering process, based on the differences between each type of environmental data and the corresponding standard environmental data at each time moment, determine the environmental disturbance adaptive coefficients for each type of environmental electrical variable at each time moment during the first iteration of clustering.

[0047] Based on the preprocessed key electrical variable data and environmental data at each time point, taking the data at time t as an example, the set of all key electrical variable data collected at time t is denoted as the key electrical variable data set Ba for that time point. t , Let the set of all environmental data collected at time t be denoted as the environmental data set En at that time. t , Among them, U t I t , C t , These represent the battery pack voltage, current, insulation resistance, capacitance, and battery temperature data at time t, respectively. H t P t Let X represent the ambient temperature, humidity, and air pressure data at time t, respectively. The vector consisting of all key electrical variables and all environmental data at time t is taken as the data vector X for that time. t This allows us to obtain the data vector within the data packets for each time frame.

[0048] Environmental factors such as temperature, humidity, and air pressure can affect electrical variables such as insulation resistance and capacitance. For example, increased temperature reduces insulation resistance, increased humidity causes moisture absorption on the insulation surface, forming conductive paths, and changes in air pressure affect the discharge characteristics of the insulation gap, leading to dynamic fluctuations in electrical variables such as resistance and capacitance under different environmental conditions. For instance, when a vehicle is driven in rainy or humid environments, the insulation resistance may temporarily decrease, prompting the system to issue an abnormal warning. However, once the environment returns to normal, the insulation resistance returns to the normal range. This change in physical characteristics is often short-lived and reversible, but it manifests as abnormal fluctuations in electrical variables in the insulation detection system. These false anomalies caused by environmental factors increase the false alarm rate, leading to unnecessary repairs and increased maintenance costs.

[0049] During vehicle operation, vehicles are often in complex environments. Traditional insulation anomaly detection methods only analyze electrical variables and do not consider the dynamic coupling relationship between environmental factors and electrical variables. This makes it impossible to distinguish between normal electrical variable fluctuations caused by environmental fluctuations and abnormal changes caused by real faults. As a result, false insulation anomaly warnings are easily generated when the environment changes drastically, affecting the accuracy and robustness of insulation anomaly detection.

[0050] To address the aforementioned issues, this application employs the K-means algorithm and constructs an improved K-means algorithm with adaptive environmental disturbance coefficients. This algorithm distinguishes between insulation anomalies caused by faults and dynamic insulation anomalies caused by environmental interference, thereby improving the robustness and accuracy of the system in detecting insulation anomalies.

[0051] Specifically, in K-means clustering, clustering is performed based on the electrical variable data characteristics of the samples. Due to environmental factors, the electrical variable characteristics fluctuate significantly. Traditional distance measurement formulas based on Euclidean distance result in fragmented clustering, making it difficult to accurately distinguish between genuine insulation anomalies and pseudo-insulation anomalies caused by interference. Therefore, this application introduces an environmental disturbance adaptive coefficient for the electrical variable data characteristics, which fluctuate significantly with external environmental factors. This improves the distance measurement formula during clustering, adaptively adjusting the contribution of electrical variables to distance measurement clustering based on environmental fluctuations. This adaptively reduces the interference of outlier electrical variables on the clustering results when the environment changes drastically, and reduces false insulation anomaly warnings caused by environmental fluctuations.

[0052] To construct the environmental disturbance adaptive coefficient, we first analyze the effects of different environmental factors on insulation resistance and capacitance, and then construct the environmental dynamic coupling coefficients of insulation resistance and capacitance at each time point.

[0053] Specifically, (1) the construction of the environmental dynamic coupling coefficient of insulation resistance is based on the fact that environmental conditions such as temperature, humidity, and air pressure have a significant physical influence on insulation resistance. When in a high-temperature environment, the molecular thermal motion of the insulating material intensifies, reducing its resistivity and causing the insulation resistance to decrease; while too low a temperature may cause the material to become brittle, leading to a decrease in the insulation performance of the material. Humidity has a more obvious effect on insulation resistance. In a high-humidity environment, due to the discharge state of the battery, the temperature of the battery pack rises, and condensation is easily generated on the surface of the insulating material, forming a conductive water film, which significantly reduces the surface insulation resistance and may even cause an increase in leakage current. Changes in air pressure indirectly change the breakdown voltage and insulation strength of the insulation system by affecting air density and air gap discharge characteristics. Especially in high-altitude or low-pressure environments, the insulation gap is more likely to break down. The coupling effect of these environmental factors causes the insulation resistance to exhibit dynamic fluctuation characteristics in actual operation, resulting in system misjudgment and thus generating false insulation abnormality warnings.

[0054] Based on the above analysis, the environmental dynamic coupling coefficient of the insulation resistance at each moment is calculated, and the expression is as follows:

[0055]

[0056] in, The environmental dynamic coupling coefficient of the insulation resistance at time t; ΔH represents the absolute value of the difference between the ambient temperature and the standard laboratory temperature at time t; t ΔP represents the absolute value of the difference between the ambient humidity at time t and the standard laboratory humidity; t α represents the absolute value of the difference between the ambient air pressure and the standard laboratory air pressure at time t; r β r α and γ are both preset positive numbers.r β r α and γ represent coefficients indicating the degree to which changes in ambient temperature, humidity, and air pressure affect the insulation resistance value, respectively. r +β r +γ=1; exp() is an exponential function with the natural constant as the base.

[0057] Preferably, in the embodiments of this application, α r β r α and γ are set to 0.6, 0.3, and 0.1, respectively. As another embodiment of this application, the implementer can set α according to the actual situation. r β r The value of γ is not specifically limited in this application.

[0058] (2) Environmental temperature and humidity also significantly affect the electrical characteristics of capacitors. Changes in ambient temperature alter the dielectric constant of the capacitor's dielectric material. Increased temperature typically increases the dielectric constant, leading to increased capacitance. However, excessively high temperatures can also increase dielectric loss and even accelerate aging, reducing the capacitor's insulation performance. Humidity affects the capacitor's dielectric and insulation layers through hygroscopic absorption. High humidity environments easily cause the insulating material to absorb moisture, forming weak conductive channels, leading to an increase in the dielectric constant and leakage current. This causes short-term fluctuations in capacitance and reduces insulation reliability. The combined effect of temperature and humidity results in dynamic changes in capacitor characteristics under different environmental conditions. If these changes are not differentiated, they can easily lead to misjudgments in the detection system, resulting in false insulation anomaly warnings.

[0059] Based on the above analysis, the environmental dynamic coupling coefficient of the capacitor at each moment is calculated, and the expression is as follows:

[0060]

[0061] Wherein, ω(C t α is the environmental dynamic coupling coefficient of the capacitor at time t; c β c All are preset positive numbers, α c β c α represents the coefficients α and α', respectively, indicating the degree to which changes in ambient temperature and humidity affect the capacitance value. c +β c =1; exp() is an exponential function with the natural constant as the base.

[0062] Preferably, in the embodiments of this application, α c β c The values ​​are set to 0.6 and 0.4 respectively. As another embodiment of this application, the implementer can set α according to the actual situation. c β cThe value of is not specifically limited in this application.

[0063] Then, the environmental disturbance adaptive coefficients for each class's environmental electrical variables at each moment during the first iteration of clustering are determined using the environmental dynamic coupling coefficients calculated above. Among these, the insulation resistance R... iso The environmental disturbance adaptive coefficients of the capacitor C are constructed based on the above-mentioned dynamic coupling coefficient settings. In this application, the environmental dynamic coupling coefficient of the insulation resistance at each time moment is used as the environmental disturbance adaptive coefficient of the insulation resistance at each time moment during the first iteration, and the environmental dynamic coupling coefficient of the capacitor at each time moment is used as the environmental disturbance adaptive coefficient of the capacitor at each time moment during the first iteration.

[0064] Furthermore, in the dynamic coupling coefficient analysis of the above environmental factors and electrical variables, changes in environmental factors can cause significant fluctuations in the values ​​of insulation resistance and capacitance relative to the standard state. When these electrical variable characteristics are measured at a distance, the numerical anomalies caused by the environment lead to non-clustering and are easily misjudged as an abnormal insulation state.

[0065] Insulation anomalies are typically caused by a series of changes in electrical variables resulting from a decrease in insulation resistance. Decreased insulation resistance leads to increased voltage, exceeding safe voltage values, and also increases leakage current. This series of changes triggers an insulation anomaly warning from the battery management system. Extreme environmental conditions, such as temperature, humidity, and air pressure, can cause short-term changes in insulation resistance, inevitably leading to transient anomalies in battery voltage and current. Without an adaptive environmental disturbance coefficient, these transient anomalies will be clustered into insulation anomaly clusters during clustering, resulting in false alarms.

[0066] Therefore, in this application, the environmental dynamic coupling coefficient of the insulation resistance at each moment is used as the environmental disturbance adaptive coefficient of the voltage and current at each moment during the first iteration.

[0067] Step S3: Adjust the environmental disturbance adaptive coefficient of any environmental electrical variable at each moment of the next iteration based on the magnitude of the environmental disturbance adaptive coefficient of any environmental electrical variable at each moment of each iteration and the corresponding iteration number.

[0068] Environmental disturbance adaptive coefficients for various environmental electrical variables can compensate for data features of electrical variables that are directly or indirectly affected by environmental factors in distance metrics. When environmental disturbances are large, such as drastic changes in temperature, humidity, or air pressure relative to normal conditions, the environmental disturbance adaptive coefficients will decrease accordingly, reducing the impact of electrical variables on sample clustering, increasing the system's tolerance to abnormal fluctuations in electrical variables, and reducing the risk of misjudgment. Conversely, when environmental disturbances are small, the weighting coefficients will increase, enhancing the effect of electrical variables on sample clustering under normal conditions, thereby helping to accurately identify real insulation faults and improving the accuracy and robustness of insulation anomaly detection.

[0069] In particular, when the environmental factors of both sets of data vectors change significantly relative to normal environmental factors, the environmental disturbance adaptive coefficients of both sets of data vectors will be very small. This makes it difficult to separate the two sets of data vectors. To avoid this situation, the environmental disturbance adaptive coefficient is set to be less than a preset first threshold during the first iteration. The environmental disturbance adaptive coefficient increases with each iteration, and its dynamic update formula is as follows:

[0070] when Less than the preset first threshold season

[0071] when Greater than or equal to the preset first threshold season

[0072] in, For the first iteration, the environmental disturbance adaptive coefficient of the j-th environmental electrical variable at time t; , respectively, are the adaptive coefficients of the j-th environmental electrical variable at time t during the (k+1)-th and k-th iterations; k is the iteration number during the k-th iteration; K represents the total number of iterations; min() indicates taking the minimum value.

[0073] This design of the dynamic environmental perturbation coefficient ensures that when the environmental perturbation coefficient is very small, it can increase with the number of iterations, preventing data vectors from being difficult to distinguish during clustering due to a very small environmental perturbation coefficient, thus avoiding inaccurate sample clustering; the dynamic increase of the environmental perturbation coefficient is limited to no more than This further constrains the impact of electrical variables on clustering under extreme environmental factors, preventing excessively large weights from being misdetected as insulation anomalies caused by faults. In this embodiment, [the following is omitted as it is not part of the main text]. The value is set to 0.25. As another embodiment of this application, the implementer may set it according to actual circumstances. The value of .

[0074] Step S4: Based on the differences in environmental electrical variables and non-environmental electrical variables in the data vectors at any two time points, and combined with the environmental disturbance adaptive coefficients at each iteration, calculate the distance between the data vectors at any two time points in each iteration; use the distance as the metric distance in the clustering algorithm; and determine the abnormal data clusters based on the differences in data vectors between different clusters.

[0075] Using the environmental disturbance adaptive coefficients obtained above, the distance between data vectors at any two moments in each iteration is calculated, as expressed by:

[0076]

[0077] Wherein, D(X) s ,X t X is the data vector at time s during the current iteration. s With the data vector X at time t t The distance between them; Ba is the set of all key electrical variables; B E B represents the set of all types of environmental electrical variables. E ={U,I,R iso ,C};Ba\B E Represent sets Ba and B E The difference set, i.e., the set of all types of non-environmental electrical variables, includes battery temperature in this application; X s (i), X t (i) represent the data values ​​of the i-th non-environmental electrical variable at time s and time t, respectively; X s (j), X t (j) represent the data values ​​of the j-th environmental electrical variable at time s and time t, respectively; ω s (j), ω t (j) represents the adaptive coefficients of the j-th environmental disturbance of the environmental electrical variable at time s and time t during the current iteration process.

[0078] The K-means algorithm takes data vectors from all time points as input, sets the number of clusters to 2, randomly selects two points as cluster centers, sets the maximum number of iterations to 50, and uses the distance calculated above as the metric distance between any two data vectors at any two time points during each iteration in the clustering process. The output is two clusters. The clustering process of the K-means algorithm is existing technology, and its specific process will not be described in detail.

[0079] The mean value of insulation resistance within all data vectors in each cluster is calculated and denoted as the first mean value. The cluster with the smallest first mean value is selected as the outlier cluster. This completes the insulation anomaly detection of automotive battery electrical variable data collected under different time periods and environments using the K-means algorithm improved based on environmental disturbance adaptive coefficients.

[0080] Step S5: Determine the location of insulation anomalies by using insulation location technology for all data vectors and their corresponding times in the abnormal data cluster.

[0081] Key parameters are extracted from the detected insulation anomaly data, including the timestamps of all data vectors in the anomaly data clusters and all electrical variable data in the data vectors. Using the timestamp information, the anomaly events are correlated with the vehicle's operating status and the electrical parameters of each individual battery cell and module recorded by the battery management system.

[0082] Subsequently, an insulation location algorithm was used to locate and analyze the abnormal positions. This application employs a segmented insulation impedance detection method to perform segmented testing on each module and its circuit of the battery pack, measuring the insulation resistance and capacitance values ​​of each circuit segment or battery module to the vehicle body, and identifying areas with abnormal abrupt changes in electrical parameters. The segmented insulation impedance detection method is existing technology, and its specific process will not be elaborated further.

[0083] Finally, an insulation anomaly location report is generated, indicating the faulty module and abnormal parameters, which can quickly and accurately locate the fault source, improve maintenance efficiency, and reduce the safety risks caused by insulation faults in energy storage battery packs.

[0084] A flowchart of the insulation anomaly detection method for energy storage systems is shown below. Figure 2 As shown.

[0085] Based on the same inventive concept as the above methods, embodiments of this application also provide an insulation anomaly detection system for an energy storage system, wherein the system stores a computer program, and when the computer program is executed by a processor, it implements the steps of any one of the above-described insulation anomaly detection methods for an energy storage system.

[0086] Based on the same inventive concept as the above methods, embodiments of this application also provide an insulation anomaly detection device for an energy storage system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described insulation anomaly detection methods for an energy storage system.

[0087] In summary, the embodiments of this application provide an insulation anomaly detection method for energy storage systems. By collecting electrical variable data and environmental data from the battery at various times, a data vector is constructed for each time frame. The data vectors collected in each time frame are analyzed to construct an adaptive weighting for environmental disturbances in the electrical variables. An improved K-means algorithm is used for clustering, enabling the differentiation and identification of insulation anomalies caused by faults and dynamic insulation anomalies caused by environmental interference. Compared to the K-means algorithm, which directly clusters and analyzes outliers based on multiple electrical variable parameters, the adaptive weighting for environmental disturbances considers the influence of external environmental data on changes in insulation electrical variable parameters. This avoids the situation where short-term abnormal fluctuations in electrical variables caused by changes in the external environment are mistakenly detected as insulation anomalies, improving the accuracy of the algorithm in detecting true insulation anomalies and avoiding false detections of insulation anomalies caused by environmental disturbances. This enhances the stability and efficiency of the battery management system in detecting insulation anomalies.

[0088] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this application. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0089] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0090] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for insulation anomaly detection for an energy storage system, characterized by, The method comprises the following steps: Collecting environmental electric variable data and non-environmental electric variable data of the vehicle energy storage battery at each time, and environment data of the vehicle; Using all the environmental electric variable data, all the non-environmental electric variable data and all the environment data at each time to construct a data vector at each time; and performing clustering on the data vectors at all times by a clustering algorithm, wherein, in the clustering process, based on the difference between each type of environment data at each time and the corresponding standard environment data, the environmental disturbance adaptive coefficient of each type of environmental electric variable at each time in the first iteration of clustering is determined; Based on the size of the environmental disturbance adaptive coefficient of any environmental electric variable at each time in each iteration and the corresponding iteration number, the environmental disturbance adaptive coefficient of the any environmental electric variable at each time in the next iteration is adjusted; Based on the difference between the environmental electric variables and the difference between the non-environmental electric variables in the data vectors of any two times, and combined with the environmental disturbance adaptive coefficient in each iteration, the distance between the data vectors of any two times in each iteration is calculated; the distance is used as a measurement distance in the clustering algorithm, and the abnormal data cluster is determined based on the difference between the data vectors in different clusters. The environmental data includes environmental temperature, humidity and air pressure; the environmental electric variables include voltage, current, insulation resistance and capacitance.

2. The insulation abnormality detection method for an energy storage system according to claim 1, characterized by, The process of obtaining the environmental disturbance adaptive coefficient of each type of environmental electric variable at each time in the first iteration of clustering is as follows:

3. The insulation abnormality detection method for an energy storage system according to claim 2, characterized by, Based on the difference between each type of environment data at each time and the corresponding standard environment data, the environmental dynamic coupling coefficient of the insulation resistance and the environmental dynamic coupling coefficient of the capacitance at each time are calculated; Based on the environmental dynamic coupling coefficients of the insulation resistance and the capacitance, the environmental disturbance adaptive coefficient of each type of environmental electric variable at each time in the first iteration of clustering is determined. The expression of the environmental dynamic coupling coefficient of the insulation resistance and the environmental dynamic coupling coefficient of the capacitance at each time is as follows:

4. The insulation abnormality detection method for an energy storage system according to claim 3, characterized by, The process of obtaining the environmental disturbance adaptive coefficient of each type of environmental electric variable at each time in the first iteration is as follows: Wherein, is the environmental dynamic coupling coefficient of the insulation resistance at time t; is the absolute value of the difference between the environmental temperature at time t and the standard laboratory temperature; AH t is the absolute value of the difference between the environmental humidity at time t and the standard laboratory humidity; AP t is the absolute value of the difference between the environmental air pressure at time t and the standard laboratory air pressure; a r , β r , γ are all preset positive numbers, a r + β r + γ = 1; ω(C t ) is the environmental dynamic coupling coefficient of the capacitor at time t; α c , β c are both preset positive numbers, α c + β c = 1; exp() is an exponential function with the natural constant as the base number.

5. The insulation abnormality detection method for an energy storage system according to claim 3, characterized by, The environmental dynamic coupling coefficient of the insulation resistance at each time is used as the environmental disturbance adaptive coefficient of the voltage, current and insulation resistance at each time in the first iteration, and the environmental dynamic coupling coefficient of the capacitance at each time is used as the environmental disturbance adaptive coefficient of the capacitance at each time in the first iteration. The process of obtaining the environmental disturbance adaptive coefficient of the any environmental electric variable at each time in the next iteration is as follows:

6. The insulation abnormality detection method for an energy storage system according to claim 1, characterized by, The expression of the distance between the data vectors of any two times in each iteration is as follows: The environmental disturbance adaptive coefficient of the jth environmental electrical variable at time t in the first iteration is denoted as The environmental disturbance adaptive coefficients of the jth environmental electrical variable at time t in the kth and k+1th iterations are denoted as When less than a preset first threshold , let wherein k is the iteration number at the kth iteration; K represents the total iteration number; min() represents taking the minimum value; When greater than or equal to a preset first threshold value then 7. The insulation abnormality detection method for an energy storage system according to claim 1, characterized by, The process of obtaining the abnormal data cluster is as follows: where D(X s , X t ) is the distance between the data vector X s at time s and the data vector X t at time t in the current iteration process; Ba is the set of all kinds of key electrical variables; B E represents the set of all kinds of environmental electrical variables; Ba\B E represents the difference set of the set Ba and the set B E ; X s (i), X t (i) represent the data values of the ith non-environmental electrical variable at time s and time t, respectively; X s (j), X t (j) represent the data values of the jth environmental electrical variable at time s and time t, respectively; ω s (j), ω t (j) represent the environmental disturbance adaptive coefficients of the jth environmental electrical variable at time s and time t in the current iteration process, respectively.

8. The insulation abnormality detection method for an energy storage system according to claim 2, characterized by, The mean value of the insulation resistance in all data vectors in each cluster is calculated, denoted as the first mean value; and the cluster with the smallest first mean value is taken as the abnormal data cluster. The computer program is executed by the processor to realize the steps of the insulation abnormality detection method for the energy storage system according to any one of claims 1-8.

9. An insulation anomaly detection system for an energy storage system in which a computer program is stored, characterized by The processor executes the computer program to realize the steps of the insulation abnormality detection method for the energy storage system according to any one of claims 1-8.

10. An insulation anomaly detection device for an energy storage system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, ​

Images