Discharge crushing environmental data monitoring system for battery recovery

Abstract

The invention relates to the technical field of battery recovery monitoring, in particular to a discharge crushing environmental data monitoring system for battery recovery. The system comprises a data acquisition module used for obtaining a sequence corresponding to each dimension based on environment data of different dimensions in the battery crushing equipment; the sample space construction module is used for screening the dimensions to obtain to-be-analyzed dimensions and constructing a sample space; the clustering analysis module is used for clustering the data points in the sample space to obtain clusters; judging whether the data points in each cluster are suspected abnormal data points or not; the abnormal degree acquisition module is used for clustering all the suspected abnormal data points in a time sequence to obtain a time sequence cluster and an abnormal degree; and the monitoring module is used for acquiring the comprehensive abnormal degree and judging whether the discharge breakage of the battery is abnormal or not based on the comprehensive abnormal degree. The condition of the old battery discharge crushing process can be accurately monitored.

Description

Technical Field

[0001] The present invention relates to the technical field of battery recycling monitoring, and in particular to an environmental data monitoring system for discharge and crushing of batteries for recycling. Background Art

[0002] The recycling of used batteries is of great significance for both environmental protection and energy regeneration. Since used batteries may still contain some residual electricity, they usually need to be discharged and crushed to ultimately recover materials and valuable metals. During the battery recycling, discharge and crushing process, it is very necessary to monitor environmental data. On the one hand, this is to prevent the exhaust gas generated during the crushing process from polluting the air environment. On the other hand, thermal reactions may occur during the discharge and crushing process. Lithium batteries, in particular, are at risk of thermal runaway. In addition, fires caused by thermal explosions during the battery crushing process start quickly and polluted gases spread quickly. Therefore, it is necessary to monitor the environmental data within the battery crushing equipment.

[0003] Existing technologies for monitoring environmental data during battery recycling, discharge, and crushing processes typically use threshold monitoring methods. This method has a single monitoring dimension and is insensitive to changes in equipment operating conditions. It is unable to detect the risk of abnormal trends in environmental data and has poor timeliness. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide an environmental data monitoring system for discharge and crushing of battery recycling. The technical solutions adopted are as follows:

[0005] One embodiment of the present invention provides an environmental data monitoring system for battery recycling and discharge crushing, the system comprising:

[0006] The data acquisition module is used to collect environmental data of different dimensions in the battery crushing equipment and pre-process it to obtain the sequence corresponding to each dimension;

[0007] The sample space construction module is used to calculate the sample feature degree of the corresponding dimension of a sequence based on the peak points and trough points in the sequence; filter the dimensions according to the sample feature degree to obtain the dimensions to be analyzed, and construct the sample space;

[0008] The cluster analysis module is used to form a data point from the elements at a time in the sequence corresponding to each dimension to be analyzed, and cluster the data points in the sample space to obtain clusters; based on the distance between a cluster and other clusters and the distance between the cluster and the zero point of the sample space, it is determined whether the data points in the cluster are suspected abnormal data points;

[0009] Anomaly degree acquisition module, used to cluster suspected abnormal data points based on the time interval between each two suspected abnormal data points to obtain time series clusters; and obtain the abnormality degree according to each time series cluster;

[0010] The monitoring module is used to obtain the comprehensive abnormality degree based on the abnormality degree and the data of each dimension to be analyzed in the suspected abnormal data point, and to determine whether the battery discharge and fragmentation are abnormal based on the comprehensive abnormality degree.

[0011] Preferably, the sequences corresponding to each dimension are obtained separately, including:

[0012] The environmental data of each dimension is normalized separately, and the normalized environmental data of one dimension constitutes the sequence corresponding to the dimension.

[0013] Preferably, calculating the sample feature degree of a corresponding dimension of a sequence based on the peak points and trough points in the sequence includes:

[0014] Obtain the peak points and trough points in a sequence and form a fluctuation point sequence in chronological order; obtain the absolute value of the difference between the vertical coordinate values ​​corresponding to each two adjacent fluctuation points in the fluctuation point sequence, and sum them to obtain the fluctuation change characteristic value; multiply the number of elements in the sequence, the fluctuation change characteristic value, and the difference between the maximum and minimum values ​​of the sequence to obtain the sample characteristic degree of the corresponding dimension of the sequence.

[0015] Preferably, the dimensions to be analyzed are screened according to the degree of sample characteristics, and the sample space is constructed, including:

[0016] Obtain a preset number of dimensions with the highest degree of sample characteristics as the dimensions to be analyzed; and construct a sample space using the dimensions to be analyzed.

[0017] Preferably, judging whether a data point in a cluster is a suspected abnormal data point based on the distance between the cluster and other clusters and the distance between the cluster and the zero point of the sample space includes:

[0018] The distance between the cluster center of a cluster and the zero point of the sample space is obtained and normalized to obtain a first distance; the mean of the distances between the cluster center of the cluster and the cluster centers of other clusters is obtained and normalized to obtain cluster distribution characteristics; the product of the first distance and the cluster distribution characteristics is the suspected abnormality degree of the cluster; based on the suspected abnormality degree of a cluster, it is determined whether the data points in the cluster are suspected abnormal data points.

[0019] Preferably, determining whether a data point in a cluster is a suspected abnormal data point based on the suspected abnormality degree of the cluster includes:

[0020] If the suspected abnormality degree of a cluster is greater than or equal to the first reference threshold, the data points in the cluster are suspected abnormal data points.

[0021] Preferably, obtaining the abnormality degree according to each time series cluster includes:

[0022] Obtain the ratio of the number of suspected abnormal data points to the total number of data points, which is recorded as the abnormal proportion; arrange each time series cluster in time series order to obtain a time series cluster arrangement sequence; in the time series cluster arrangement sequence, calculate the mean of the time interval between the two suspected abnormal data points with the closest time interval in each two adjacent time series clusters, and normalize them to obtain the inter-cluster distribution characteristic value; calculate the mean of the time interval between each suspected abnormal data point and the suspected abnormal data point with the smallest time interval in a time series cluster, and record it as the intra-cluster distribution characteristic value of the time series cluster; use an exponential function with a natural constant as the base to perform negative correlation mapping on the mean of the intra-cluster distribution characteristic values ​​of all time series clusters to obtain the cumulative intra-cluster distribution characteristic value; add the cumulative intra-cluster distribution characteristic value and the inter-cluster distribution characteristic value, and multiply them with the abnormal proportion to obtain the degree of abnormality.

[0023] Preferably, obtaining a comprehensive abnormality degree based on the abnormality degree and the data of each dimension to be analyzed in the suspected abnormal data point includes:

[0024] Calculate the mean of the data corresponding to one of the dimensions to be analyzed among all suspected abnormal data points and normalize it, which is recorded as the mean of the suspected abnormal data of the dimension to be analyzed; calculate the standard deviation of the mean of the suspected abnormal data of all dimensions to be analyzed, and multiply it by the normalized abnormality degree to obtain the comprehensive abnormality degree.

[0025] Preferably, judging whether the battery discharge and fracture are abnormal based on the comprehensive abnormality degree includes:

[0026] When the comprehensive abnormality degree is greater than or equal to the second reference threshold, the battery discharge fragmentation is abnormal.

[0027] The embodiments of the present invention have at least the following beneficial effects: the present application collects environmental data of different dimensions in the battery crushing equipment for analysis, which can increase the amount of information and thus obtain more accurate analysis results; further, the sample feature degree of each dimension is calculated for dimension screening, which achieves the effect of dimensionality reduction while ensuring the importance of the environmental data of the selected dimension for battery crushing and recycling monitoring, thereby ensuring the accuracy of monitoring; then cluster analysis is performed on multiple data points composed of elements at each moment in the sequence corresponding to each dimension to be analyzed, and whether the data points in the cluster are suspected abnormal data points is determined based on the spatial distribution characteristics of the cluster, and then the degree of abnormality is obtained based on the distribution characteristics of the suspected abnormal data points in the time domain, and the comprehensive abnormality degree is obtained by combining the characteristics of the data of each dimension to be analyzed in the suspected abnormal data points, and then monitoring whether the battery discharge crushing occurs abnormally, capturing possible environmental abnormalities in advance, improving timeliness, and obtaining more accurate environmental abnormality monitoring results in the battery crushing equipment during the battery discharge crushing process on the basis of high timeliness. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. 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 any creative work.

[0029] Figure 1 This is a system block diagram of an environmental data monitoring system for discharge and crushing of battery recycling provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of an environmental data monitoring system for battery recycling and discharging according to the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0031] Unless defined otherwise, 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 invention belongs.

[0032] The following describes in detail a specific scheme of an environmental data monitoring system for discharge and crushing of battery recycling provided by the present invention with reference to the accompanying drawings.

[0033] Example: The primary application scenario of this invention is to monitor environmental data within the battery crushing equipment during battery recycling, discharge, and crushing. Based on this data, determine whether any abnormalities in the environmental data could cause abnormal gas concentrations or fire risks. Because the primary temperature changes and gas release occur during the crushing process, these monitoring steps are primarily performed within the battery crushing equipment.

[0034] See also Figure 1 , which shows a system block diagram of an environmental data monitoring system for discharge and crushing of battery recycling provided by an embodiment of the present invention. The system includes the following modules:

[0035] The data acquisition module is used to collect environmental data of different dimensions in the battery crushing equipment and perform preprocessing to obtain the sequences corresponding to each dimension.

[0036] The discharge and crushing of old batteries is carried out in the battery crushing equipment. Therefore, it is necessary to install temperature sensors, humidity sensors, and various gas concentration sensors in the battery crushing equipment to collect environmental data in different dimensions. They monitor the temperature and humidity of the equipment environment during the battery crushing process, as well as the gas concentrations of various waste gases that may be generated by battery crushing, such as combustible gases H2 and CH4, and toxic gases HF, CO, and VOC.

[0037] To enable subsequent correlation analysis of multidimensional data, all sensors are set to have the same monitoring frequency of 10Hz. The collected environmental data is transmitted to the data processing terminal. Furthermore, the environmental data 10 seconds before the real-time monitoring point is obtained to monitor the situation at the real-time monitoring point. The implementer can adjust the length of the environmental data collection period based on actual conditions. Because the magnitudes of different dimensions are different, the collected values ​​corresponding to the environmental data of each dimension need to be normalized. The specific normalization model is:

[0038] x' represents the normalized data value, x max is the maximum value in the current dimension environment data, that is, the monitoring threshold, x min =x represents the minimum value of the environmental data in the current dimension, that is, the initial monitoring parameters of the equipment before crushing begins, such as exhaust gas concentration of 0 and temperature and humidity representing the actual air temperature and humidity. x represents the data value before normalization. The environmental data of each dimension is normalized separately, and the normalized environmental data of a dimension constitutes the sequence corresponding to that dimension.

[0039] The sample space construction module is used to calculate the sample feature degree of the corresponding dimension of a sequence based on the peak points and trough points in the sequence; filter the dimensions according to the sample feature degree to obtain the dimensions to be analyzed, and construct the sample space.

[0040] It is not just the extreme abnormal changes in a single data that will lead to corresponding environmental anomalies. For example, if the temperature rises to the threshold, it indicates that there is an abnormal fire risk, and if the exhaust gas concentration rises to the threshold, it indicates that there is an abnormal gas pollution. However, before the environmental data in each dimension rises to the monitoring threshold, if there are simultaneous changes in environmental data in multiple dimensions, it also indicates the hidden dangers of environmental anomalies. For example, if the gas concentration changes at the same time as the temperature changes, an explosion may occur under the combined action of exhaust gas and high temperature, and there is a possibility of further deterioration and fire. Similarly, if the temperature also rises at the same time as the gas concentration changes, the rising temperature may cause the battery to break and release more exhaust gas. At the same time, in terms of humidity, some battery materials (such as lithium metal and electrolyte) will react violently with water, releasing hydrogen or heat. Therefore, when there are abnormal changes in humidity, the possibility of abnormal environmental monitoring values ​​will also increase. Therefore, the common abnormal changes in multi-dimensional environmental data indicate a higher risk of environmental anomalies.

[0041] Therefore, we integrate multidimensional environmental data to construct a multidimensional sample space. Because data points tend to be closer together when the dimensionality is too high, the distance relationship between data points is weakened, leading to inaccurate feature analysis results. Therefore, dimensionality reduction is necessary to construct a multidimensional sample space. Sequences corresponding to dimensions with smaller fluctuation frequencies and amplitudes indicate that less feature information can be extracted. However, datasets with larger fluctuation frequencies and amplitudes indicate that more feature information can be extracted, making them more suitable as sample spaces for subsequent analysis. On the other hand, the greater the difference between the maximum and minimum values ​​within a dimensional dataset, the more likely the data in that dimension is to exhibit large-scale abnormal variations, making it more suitable as an indicator of the sample space for abnormal feature analysis.

[0042] Furthermore, the sample feature degree of the corresponding dimension of a sequence is calculated based on the peak points and trough points in a sequence. Specifically, the peak points and trough points in a sequence are obtained and compose a fluctuation point sequence in chronological order; the absolute value of the difference between the vertical coordinate values ​​corresponding to every two adjacent fluctuation points in the fluctuation point sequence is obtained, and the sum is performed to obtain the fluctuation change characteristic value; the number of elements in the sequence, the fluctuation change characteristic value and the difference between the maximum value and the minimum value of the sequence are multiplied to obtain the sample feature degree of the corresponding dimension of the sequence.

[0043] The specific calculation model is:

[0044] Among them, R q Indicates the sample feature degree of the qth dimension; n pIt represents the number of peak points and trough points in the sequence of the qth dimension, that is, the number of points in the fluctuation point sequence. The more peak points and trough points there are, the greater the frequency of environmental data fluctuation in this dimension and the stronger the degree of sample characteristics. The peak points and trough points are obtained using the AMPD peak search algorithm. i is the amplitude of the ith fluctuation point in the fluctuation point sequence, that is, the vertical span from this fluctuation point to the next fluctuation point, expressed by the absolute value of the difference between the vertical coordinates of two adjacent fluctuation points in the fluctuation point sequence. Indicates the characteristic value of fluctuation change, which integrates the amplitude of all fluctuations. The larger the cumulative sum of the amplitudes, the larger the overall level of the fluctuation amplitude of the current dimension and the stronger the degree of sample characteristics; x' max is the maximum value of the normalized environmental data in the sequence corresponding to the qth dimension, x' min is the minimum value of the normalized environmental data, x' max -x' min The larger the difference, the greater the difference between the maximum and minimum values ​​in the sequence corresponding to the current dimension, the more likely there is a large-span abnormal change, and the stronger the sample characteristics.

[0045] Furthermore, a preset number of dimensions with the highest sample characteristic degree are obtained as dimensions to be analyzed. Preferably, in the embodiment of the present invention, a reference value of the preset number is set to 3, which can be determined by the implementer according to actual conditions. The sample space is constructed using the preset number of dimensions to be analyzed.

[0046] The clustering analysis module is used to form a data point from the elements at a moment in the sequence corresponding to each dimension to be analyzed, and cluster the data points in the sample space to obtain clusters; based on the distance between a cluster and other clusters and the distance between the cluster and the zero point of the sample space, it is determined whether the data points in the cluster are suspected abnormal data points.

[0047] The above constructs a sample space for analysis. Furthermore, a data point in the sample space is composed of the elements at the corresponding moment in the sequence corresponding to each dimension to be analyzed. Once a multidimensional sample space is obtained, multidimensional environmental monitoring values ​​can be jointly analyzed based on this multidimensional sample space, thereby obtaining environmental anomaly characteristics that may indicate risks. Furthermore, the data points in the sample space are clustered using the kmeans clustering algorithm. Clustering is based on Euclidean clustering between two data points. The k value is obtained using the elbow method, and the initial cluster centers are randomly selected to obtain kmeans clustering results, thereby obtaining different clusters.

[0048] In the kmeans clustering results, if there are clusters farther from the zero point of the sample space, the higher the corresponding temperature, humidity, and exhaust gas concentration levels, the more likely they are to be a cluster of data points with abnormal characteristics. On the other hand, if a cluster is farther from other clusters, it indicates a higher degree of outlier status, and the more the temperature, humidity, and exhaust gas concentration represented within it deviate from normal levels, the more likely it is an abnormal cluster. The zero point of the sample space is the origin of the sample space. For example, in three-dimensional space, the zero point of the sample space is the origin of the three-dimensional rectangular coordinate system.

[0049] Therefore, the distance between a cluster and other clusters and the distance between the cluster and the zero point of the sample space are used to determine whether the data points within the cluster are suspected anomalies. Specifically, the distance between the cluster center of a cluster and the zero point of the sample space is obtained and normalized to obtain a first distance. The mean of the distances between the cluster center of the cluster and the cluster centers of all other clusters is obtained and normalized to obtain the cluster distribution feature. The product of the first distance and the cluster distribution feature is the suspected anomaly level of the cluster.

[0050] The specific calculation model for the suspected abnormality degree is:

[0051] Among them, D c represents the suspected abnormality degree of the cth cluster, a0 represents the zero point of the sample space, ac represents the cluster center of the cth cluster, ||a0-a c || is the Euclidean distance from the cluster center of the cth cluster to the zero point of the sample space, orm(||a0-a c ||) is the first distance, which represents the distance between the cth cluster and the zero point. The larger the distance, the higher the suspected abnormality of the data points in the cluster. m is the number of clusters, n m -1 is the number of clusters after removing the cth cluster, a i represents the cluster center of the i-th cluster, ||a i -a c || is the Euclidean distance from the cluster center of the i-th cluster to the cluster center of the c-th cluster, It represents the cluster distribution characteristics, that is, the normalized value of the mean distance between all other clusters and the cth cluster. The larger the value, the higher the degree of outlier of the cluster. The more the temperature, humidity, and exhaust gas concentration represented in the cluster deviate from the normal level, the more likely it is an abnormal cluster. norm represents the normalization operation.

[0052] Furthermore, based on the suspected abnormality degree of each cluster, a first reference threshold is set to determine whether the data points within each cluster are suspected abnormal data points. Preferably, in the embodiment of the present application, the value of the first reference threshold is 0.25. Specifically, the implementer can adjust the first reference threshold based on scenario experiments to obtain a first reference threshold that conforms to the actual application scenario. If the suspected abnormality degree of a cluster is greater than or equal to the first reference threshold, the data points within the cluster are suspected abnormal data points.

[0053] In this way, all suspected abnormal data points can be obtained.

[0054] The abnormality degree acquisition module is used to cluster the suspected abnormal data points based on the time interval between each two suspected abnormal data points to obtain time series clusters; and obtain the abnormality degree according to each time series cluster.

[0055] After the cluster analysis, we have obtained the suspected abnormal data points. Further, we need to analyze the suspected abnormal data points in combination with the time series characteristics.

[0056] Too many batteries that are broken and clustered together at similar times, or incomplete battery discharge, will generate higher amounts of heat and gas, making it more likely to cause fire risks and abnormal exhaust emissions. Therefore, relying solely on the aforementioned cluster spatial distribution characteristics for anomaly analysis has limitations. Here, the distribution characteristics of suspected abnormal data points in the time domain are combined to further determine the degree of abnormality in the dataset. When suspected abnormal data points are more clustered in the time domain, it indicates that there are a large number of broken batteries at similar times or changes in heat or gas aggregation caused by incomplete battery discharge, corresponding to a higher degree of abnormal risk. When suspected abnormal data points are relatively discrete and uniform in the time domain distribution, the risk of causing an explosion, fire, or an abnormal surge in gas is relatively low.

[0057] Then, all suspected abnormal data points are clustered according to the time interval between each two suspected abnormal data points to obtain time series clusters. That is, the suspected abnormal data points are clustered using the time interval between the suspected abnormal data points instead of the Euclidean distance, and the abnormality of the data set is judged by evaluating the clustering results. On the basis that the more suspected abnormal data points there are, the higher the risk of environmental abnormality. When the suspected abnormal data points in the clustering results are densely distributed in the time domain of the time series clusters, and there is a certain interval between the time series clusters, it means that the suspected abnormal data points have strong clustering in the time domain distribution, and the environmental abnormality risk is high; when the suspected abnormal data points in the clustering results are discretely distributed in the time domain of the time series clusters, and the interval distance between the time series clusters is small, it means that the suspected abnormal data points have weak clustering in the time domain distribution, are not easily clustered into obvious clusters, and the corresponding environmental abnormality risk is low.

[0058] Furthermore, the degree of anomaly is obtained based on each time series cluster. Specifically, the ratio of the number of suspected anomaly data points to the total number of data points is obtained, which is recorded as the anomaly ratio; each time series cluster is arranged in chronological order to obtain a time series cluster arrangement sequence; in the time series cluster arrangement sequence, the mean time interval between the two suspected anomaly data points with the closest time interval in each two adjacent time series clusters is calculated, and normalized to obtain the inter-cluster distribution characteristic value; the mean time interval between each suspected anomaly data point and the suspected anomaly data point with the smallest time interval in a time series cluster is calculated, and recorded as the intra-cluster distribution characteristic value of the time series cluster; an exponential function with a natural constant as the base is used to negatively correlate the mean of the intra-cluster distribution characteristic values ​​of all time series clusters to obtain the cumulative intra-cluster distribution characteristic value; the cumulative intra-cluster distribution characteristic value is added to the inter-cluster distribution characteristic value, and then multiplied by the anomaly ratio to obtain the degree of anomaly.

[0059] The specific calculation model of the abnormality degree is:

[0060] Among them, E represents the degree of abnormality, N is the total number of data points, and N t is the number of suspected abnormal data points, is the ratio of suspected abnormal data points, that is, the abnormal ratio. The larger the ratio, the more suspected abnormal data points there are, the greater the risk of environmental abnormality, and the greater the degree of abnormality of the data set. z is the number of time series clusters after one-dimensional clustering, n z -1 is the number of time series clustering intervals, Δh i The time interval between the two suspected abnormal data points with the closest time interval in the i-th time series cluster and the next time series cluster after the i-th time series cluster in the time series cluster arrangement sequence is: Represents the mean of the time intervals between time series clusters. Tanh is a positive proportional normalization used to unify the magnitude. is the inter-cluster distribution characteristic value. The larger the value, the more suspected abnormal data points are distributed in regional clusters. The stronger the clustering of abnormal clusters, the stronger the abnormality of the data set. i represents the intra-cluster distribution characteristic value of the i-th time series cluster, n i is the number of suspected abnormal data points in the i-th time series cluster, s j_near It represents the time interval between the jth suspected abnormal data point in the i-th time series cluster and the nearest suspected abnormal data point (the suspected abnormal data point with the closest time interval). represents the cumulative intra-cluster distribution eigenvalue; This represents the mean time interval between each suspected anomaly data point and its nearest point within all time series clusters. A smaller value indicates that the suspected anomaly data points are more tightly clustered within the clustered distribution. The stronger the clustering of the anomaly clusters, the more abnormal the dataset, and the greater the risk of environmental anomaly. exp is an exponential function with the natural constant e as its base. This function can be used to determine the degree of anomaly in the currently collected data.

[0061] The monitoring module is used to obtain the comprehensive abnormality degree based on the abnormality degree and the data of each dimension to be analyzed in the suspected abnormal data point, and to determine whether the battery discharge and fragmentation are abnormal based on the comprehensive abnormality degree.

[0062] Since single extreme outliers can also cause data points in the sample space to be outliers, but single extreme outliers may be caused by sensor errors or equipment environmental parameter control such as temperature switching, and changes in single environmental parameters do not have certain explosion and fire risk characteristics, we comprehensively analyze the data value distribution of all dimensions to be analyzed and correct the abnormality level of the final data set. The standard deviation of the mean of the data corresponding to each dimension to be analyzed at the suspected abnormal data point is obtained. When the standard deviation is smaller, the data outliers in the corresponding multidimensional sample space are more likely to be composed of outliers from multiple dimensions, and the abnormality level of the corresponding multidimensional data set is relatively high. If the standard deviation of the suspected abnormal data values ​​of all dimensions is relatively large, it means that the data outliers in the multidimensional sample space are more likely to be composed of extreme outliers from a single dimension, and the abnormality level of the corresponding multidimensional data set is relatively low.

[0063] The comprehensive anomaly degree is thus obtained based on the degree of anomaly and the data for each dimension to be analyzed in the suspected anomaly data point. Specifically, the mean of the data corresponding to one dimension to be analyzed for all suspected anomaly data points is calculated and normalized, and recorded as the mean of the suspected anomaly data for that dimension to be analyzed. The standard deviation of the mean of the suspected anomaly data for all dimensions to be analyzed is calculated and multiplied by the normalized degree of anomaly to obtain the comprehensive anomaly degree.

[0064] The specific calculation model of the comprehensive abnormality degree is:

[0065] Among them, E' represents the comprehensive abnormality degree, Eg represents the normalized value of the abnormality degree, It represents the result of normalizing the mean of the data corresponding to the first dimension to be analyzed among all suspected abnormal data points. Similarly, and They represent the normalized results of the mean values ​​of the second and third dimensions to be analyzed of all suspected abnormal data points, that is, the mean of the suspected abnormal data. This mean reflects the deviation level of the data value in the current dimension relative to the normal value. It is the standard deviation of the mean of the suspected abnormal data of the three dimensions to be analyzed. The smaller the standard deviation, the stronger the abnormality of all dimensions to be analyzed, and the more likely the monitoring environment abnormality is. It excludes the situation where the abnormality is large due to a single extreme abnormality in a certain dimension. The abnormality degree obtained is caused by multi-dimensional abnormalities, which indicates a higher environmental risk abnormality.

[0066] In this way, the comprehensive abnormality level of the currently collected data can be obtained. The greater the comprehensive abnormality level, the greater the possibility of abnormality. Finally, based on the comprehensive abnormality level, it is judged whether the battery discharge crushing is abnormal. Specifically, a second reference threshold is set. Preferably, the value of the second reference threshold in the embodiment of the present invention is 0.4. The accurate empirical threshold of the specific scenario can be adjusted on the basis of the second reference threshold through experimental debugging in the corresponding scenario. When the comprehensive abnormality level is greater than or equal to the second reference threshold, it is considered that there is a certain environmental abnormality risk in the current real-time monitoring data, and the battery discharge crushing is abnormal. At this time, the battery crushing process is stopped and a real-time alarm is issued to remind the staff of the relevant monitored environmental hazards, and the battery at the feed port is adjusted.

[0067] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0068] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0069] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A battery recycling and discharging crushing environmental data monitoring system, characterized in that: The system includes: The data acquisition module is used to collect environmental data of different dimensions in the battery crushing equipment and pre-process it to obtain the sequence corresponding to each dimension; The sample space construction module is used to calculate the sample feature degree of the corresponding dimension of a sequence based on the peak points and trough points in the sequence; filter the dimensions according to the sample feature degree to obtain the dimensions to be analyzed, and construct the sample space; The cluster analysis module is used to form a data point from the elements at a time in the sequence corresponding to each dimension to be analyzed, and cluster the data points in the sample space to obtain clusters; based on the distance between a cluster and other clusters and the distance between the cluster and the zero point of the sample space, it is determined whether the data points in the cluster are suspected abnormal data points; Anomaly degree acquisition module, used to cluster suspected abnormal data points based on the time interval between each two suspected abnormal data points to obtain time series clusters; and obtain the abnormality degree according to each time series cluster; The monitoring module is used to obtain the comprehensive abnormality degree based on the abnormality degree and the data of each dimension to be analyzed in the suspected abnormal data point, and to determine whether the battery discharge and fragmentation are abnormal based on the comprehensive abnormality degree.

2. The environmental data monitoring system for battery recycling and discharge crushing according to claim 1 is characterized in that: The sequence corresponding to each dimension is obtained respectively, including: The environmental data of each dimension is normalized separately, and the normalized environmental data of one dimension constitutes the sequence corresponding to the dimension.

3. The environmental data monitoring system for battery recycling and discharging according to claim 1 is characterized in that: The step of calculating the sample feature degree of a corresponding dimension of a sequence based on peak points and trough points in the sequence includes: Obtain the peak points and trough points in a sequence and form a fluctuation point sequence in chronological order; obtain the absolute value of the difference between the vertical coordinate values ​​corresponding to each two adjacent fluctuation points in the fluctuation point sequence, and sum them to obtain the fluctuation change characteristic value; multiply the number of elements in the sequence, the fluctuation change characteristic value, and the difference between the maximum and minimum values ​​of the sequence to obtain the sample characteristic degree of the corresponding dimension of the sequence.

4. The environmental data monitoring system for battery recycling and discharging according to claim 1 is characterized in that: The steps of screening the dimensions according to the degree of sample characteristics to obtain the dimensions to be analyzed and constructing the sample space include: Obtain a preset number of dimensions with the highest degree of sample characteristics as the dimensions to be analyzed; and construct a sample space using the dimensions to be analyzed.

5. The environmental data monitoring system for discharge and crushing of battery recycling according to claim 1 is characterized in that: The step of judging whether a data point in a cluster is a suspected abnormal data point based on the distance between the cluster and other clusters and the distance between the cluster and the zero point of the sample space includes: The distance between the cluster center of a cluster and the zero point of the sample space is obtained and normalized to obtain a first distance; the mean of the distances between the cluster center of the cluster and the cluster centers of other clusters is obtained and normalized to obtain cluster distribution characteristics; the product of the first distance and the cluster distribution characteristics is the suspected abnormality degree of the cluster; based on the suspected abnormality degree of a cluster, it is determined whether the data points in the cluster are suspected abnormal data points.

6. The environmental data monitoring system for discharge and crushing of battery recycling according to claim 5, characterized in that: The step of determining whether a data point in a cluster is a suspected abnormal data point based on the suspected abnormality degree of the cluster includes: If the suspected abnormality degree of a cluster is greater than or equal to the first reference threshold, the data points in the cluster are suspected abnormal data points.

7. The environmental data monitoring system for battery recycling and discharging according to claim 1 is characterized in that: Obtaining the degree of abnormality based on each time series cluster includes: Obtain the ratio of the number of suspected abnormal data points to the total number of data points, which is recorded as the abnormal proportion; arrange each time series cluster in time series order to obtain a time series cluster arrangement sequence; in the time series cluster arrangement sequence, calculate the mean of the time interval between the two suspected abnormal data points with the closest time interval in each two adjacent time series clusters, and normalize them to obtain the inter-cluster distribution characteristic value; calculate the mean of the time interval between each suspected abnormal data point and the suspected abnormal data point with the smallest time interval in a time series cluster, and record it as the intra-cluster distribution characteristic value of the time series cluster; use an exponential function with a natural constant as the base to perform negative correlation mapping on the mean of the intra-cluster distribution characteristic values ​​of all time series clusters to obtain the cumulative intra-cluster distribution characteristic value; add the cumulative intra-cluster distribution characteristic value and the inter-cluster distribution characteristic value, and multiply them with the abnormal proportion to obtain the degree of abnormality.

8. The environmental data monitoring system for battery recycling and discharging according to claim 1 is characterized in that: The method of obtaining a comprehensive abnormality degree based on the abnormality degree and the data of each dimension to be analyzed in the suspected abnormal data point includes: Calculate the mean of the data corresponding to one of the dimensions to be analyzed among all suspected abnormal data points and normalize it, which is recorded as the mean of the suspected abnormal data of the dimension to be analyzed; calculate the standard deviation of the mean of the suspected abnormal data of all dimensions to be analyzed, and multiply it by the normalized abnormality degree to obtain the comprehensive abnormality degree.

9. The environmental data monitoring system for discharge and crushing of battery recycling according to claim 1, characterized in that: The method of judging whether battery discharge and fracture are abnormal based on the comprehensive abnormality degree includes: When the comprehensive abnormality degree is greater than or equal to the second reference threshold, the battery discharge fragmentation is abnormal.

Images