Energy storage power station safety control method based on energy storage unit state evaluation

Abstract

The invention relates to an energy storage power station safety control method based on energy storage unit state evaluation, and belongs to the technical field of energy storage power station safety control. The method comprises the steps that the temperature value of a single energy storage unit is collected, position information corresponding to the single energy storage unit is obtained, and therefore three-dimensional information of each energy storage unit is obtained; converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; segmenting data in the temperature distribution matrix, and calculating a diffusion coefficient to obtain the diffusion coefficient; abnormal values are calculated and binarized, then accumulation is carried out, and an accumulated value matrix is constructed and windowing is carried out; using the sum of accumulated values in the windowing as a density statistical value; and then two-dimensional Gaussian fitting is adopted to judge temperature anomaly. Abnormal judgment is performed by using the stability of heat conduction abnormal diffusion value mutation, the abnormal state monitoring of the energy storage unit is completed, and the method is easy to popularize and apply.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy storage power station safety control, and in particular relates to an energy storage power station safety control method based on energy storage unit status evaluation. Background Art

[0002] The existing abnormality detection of energy storage units can be realized based on the abnormal temperature detection collected by the temperature sensor, but when the temperature collected by the temperature sensor is abnormal, it does not mean that the actual temperature of the energy storage unit is abnormal. When the abnormality detection of the energy storage unit is performed based on temperature data, since temperature is conductive, the temperature detection when the temperature sensor fails can be realized based on the conductivity of temperature.

[0003] However, the above technical solution only relies on the change of nearby temperature. However, in actual production, the energy storage unit has a heat dissipation system, which causes the problem of some energy storage units not being stable and having large differences, resulting in the problem of false detection in the existing energy storage unit abnormality monitoring based on temperature sensors.

[0004] That is, if the temperature sensor fails, it may be unreasonable to directly use the temperature value for abnormal detection, because the failure of the temperature sensor does not affect the normal working state of the energy storage battery. Therefore, how to overcome the shortcomings of the existing technology is an urgent problem to be solved in the field of energy storage power station safety control technology. Summary of the invention

[0005] The purpose of the present invention is to solve the deficiencies of the prior art and to provide a method for safety control of an energy storage power station based on energy storage unit status evaluation.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A method for safety control of an energy storage power station based on energy storage unit status assessment comprises the following steps:

[0008] Step (1), collecting the temperature value of a single energy storage unit and obtaining the position information corresponding to the single energy storage unit, thereby obtaining three-dimensional information of each energy storage unit;

[0009] Step (2), converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; then segmenting the data in the temperature distribution matrix and calculating the diffusion coefficient to obtain the diffusion coefficient;

[0010] Step (3), according to the diffusion coefficient of the energy storage unit, the abnormal value of the temperature sensor acquisition value corresponding to it is calculated, and the abnormal values ​​of all energy storage units are binarized, and then the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix; the accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistic value of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function as the final temperature abnormal value; if the final temperature abnormal value is less than or equal to the final temperature abnormal value threshold, it means that the temperature of the current storage unit is abnormal, and control processing is required to eliminate the abnormality.

[0011] Furthermore, preferably, in step (1), the temperature of each energy storage unit is collected by a temperature sensor.

[0012] Furthermore, preferably, in step (1), the position information corresponding to a single energy storage unit is the numbering of the horizontal and vertical coordinates of the energy storage unit; and the three-dimensional information of each energy storage unit is the temperature value of the energy storage unit, the normalized horizontal coordinate numbering value, and the normalized vertical coordinate numbering value.

[0013] Further, preferably, in step (2), the three-dimensional information of each energy storage unit is converted into a temperature distribution matrix, and the specific method is as follows:

[0014] The three-dimensional information of each energy storage unit is converted into a two-dimensional matrix. The values ​​in the matrix are temperature values. The rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​of the collected energy storage unit.

[0015] Further, preferably, in step (2), the data in the temperature distribution matrix is ​​then segmented and the diffusion coefficient is calculated to obtain the diffusion coefficient; the specific method is:

[0016] Calculate the abnormal diffusion coefficient of the i-th energy storage unit as D i :

[0017] D i =E i exp(1-F i )

[0018] Among them, E i is the temperature difference of the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit;

[0019] E i The calculation method is:

[0020] Obtain the energy storage units that are currently adjacent to the i-th energy storage unit and are in the same operating state, mark these energy storage units as 1 in the two-dimensional matrix, and mark the remaining energy storage units as 0;

[0021] Extract the center of the area with value 1, use the extracted center as the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit, obtain the temperature value of the center at the current moment, and obtain the three-dimensional data of the center; the center is the center of the common charging and discharging energy storage unit;

[0022] Calculate the absolute value of the difference between the current temperature of the center and the previous temperature of the center, and use the ratio of this absolute value to the current temperature of the center as E. i ;

[0023] F i It is the ratio of the Euclidean distance between the ith energy storage unit and the center of the common charging and discharging energy storage unit to the longest distance. The longest distance refers to the maximum value of the Euclidean distances between all energy storage units and the center of the common operating energy storage unit.

[0024] Furthermore, preferably, the common operation state refers to charging or discharging.

[0025] Further, preferably, in step (3), the abnormal value of the temperature sensor acquisition value corresponding to the energy storage unit is calculated according to the diffusion coefficient of the energy storage unit, and the specific method is:

[0026] Calculate the abnormal value of the temperature sensor data collected by the i-th energy storage unit as H i :

[0027] H i =exp(-L i )*min(D i , D′);

[0028] in,

[0029] D' is the average value of the abnormal diffusion coefficient of the i-th energy storage unit and all the energy storage units adjacent to it; L i is the mutation value of the abnormal diffusion coefficient around the i-th energy storage unit.

[0030] Furthermore, preferably, in step (3), the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix, and the specific method is as follows:

[0031] Obtain the abnormal values ​​of all energy storage units at each moment, and obtain the abnormal value set {H} of all energy storage units at that moment; perform binary classification on {H} at each moment; calculate the mean of the abnormal values ​​of the energy storage units corresponding to the two categories of the binary classification respectively; take the category with the larger mean as the high H category; take the category with the smaller mean as the low H category;

[0032] Then, all energy storage units in the high H category are assigned a value, and the specific value assigned is W=1;

[0033] Assign values ​​to all energy storage units in the low H category, specifically W = 0;

[0034] For each energy storage unit, the W at all times during the data collection process is accumulated to obtain the accumulated value W';

[0035] According to the normalized horizontal coordinate number value, the normalized vertical coordinate number value and the accumulated value W' of each energy storage unit, it is converted into an accumulated value matrix;

[0036] The values ​​in the accumulated value matrix are accumulated values, and the rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​after the energy storage unit is collected and normalized.

[0037] Furthermore, it is preferred that the accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistics of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the variance values ​​corresponding to the two dimensions in the two-dimensional Gaussian function as the final temperature abnormal value; the specific method is:

[0038] The window size is 3*3; the W′ values ​​of all energy storage units in the window are added as the density statistics of the energy storage unit at the center of the window; the Gaussian mixture model is used to fit the multivariate two-dimensional Gaussian function to the density statistics; the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function is obtained as the final temperature anomaly value.

[0039] In the present invention, when the temperature of the current reserve unit is abnormal and needs to be controlled, the present invention does not limit the specific means of processing, as long as the abnormality can be eliminated.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] (1) When the present invention uses the temperature difference to detect the abnormal temperature conduction, the distance data is added to enhance the temperature difference, so that the abnormal heat conduction diffusion value of the energy storage unit close to the center of the common operation energy storage unit is more obvious;

[0042] (2) The present invention considers the abnormal diffusion value of heat conduction of adjacent energy storage units of a single energy storage unit and the abnormal diffusion value of heat conduction during continuous abnormal detection of multiple moments of the unit itself, and uses the stability of the sudden change of the abnormal diffusion value of heat conduction to perform abnormal judgment, thereby completing the monitoring and evaluation of the abnormal state of the energy storage unit, realizing the safe management and control of the energy storage power station, and improving the accuracy of the management center's monitoring and status evaluation of energy storage units with abnormal temperatures;

[0043] (3) The energy storage power station safety control method based on energy storage unit status evaluation of the present invention is simple and reliable, can improve the accuracy of energy storage monitoring, and is easy to promote and apply. DETAILED DESCRIPTION

[0044] The present invention is further described in detail below in conjunction with embodiments.

[0045] Those skilled in the art will appreciate that the following examples are only used to illustrate the present invention and should not be considered to limit the scope of the present invention. If no specific techniques or conditions are specified in the examples, the techniques or conditions described in the literature in the art or the product specifications are used. If the manufacturer of the materials or equipment used is not specified, they are all conventional products that can be purchased.

[0046] Example 1

[0047] A method for safety control of an energy storage power station based on energy storage unit status assessment comprises the following steps:

[0048] Step (1), collecting the temperature value of a single energy storage unit and obtaining the position information corresponding to the single energy storage unit, thereby obtaining three-dimensional information of each energy storage unit;

[0049] Step (2), converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; then segmenting the data in the temperature distribution matrix and calculating the diffusion coefficient to obtain the diffusion coefficient;

[0050] Step (3), according to the diffusion coefficient of the energy storage unit, the abnormal value of the temperature sensor acquisition value corresponding to it is calculated, and the abnormal values ​​of all energy storage units are binarized, and then the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix; the accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistic value of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function as the final temperature abnormal value; if the final temperature abnormal value is less than or equal to the final temperature abnormal value threshold, it means that the temperature of the current storage unit is abnormal, and control processing is required to eliminate the abnormality.

[0051] Example 2

[0052] A method for safety control of an energy storage power station based on energy storage unit status assessment comprises the following steps:

[0053] Step (1), collecting the temperature value of a single energy storage unit and obtaining the position information corresponding to the single energy storage unit, thereby obtaining three-dimensional information of each energy storage unit;

[0054] Step (2), converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; then segmenting the data in the temperature distribution matrix and calculating the diffusion coefficient to obtain the diffusion coefficient;

[0055] Step (3), according to the diffusion coefficient of the energy storage unit, the abnormal value of the temperature sensor acquisition value corresponding to it is calculated, and the abnormal values ​​of all energy storage units are binarized, and then the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix; the accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistic value of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function as the final temperature abnormal value; if the final temperature abnormal value is less than or equal to the final temperature abnormal value threshold, it means that the temperature of the current storage unit is abnormal, and control processing is required to eliminate the abnormality.

[0056] In step (1), the temperature of each energy storage unit is collected by a temperature sensor.

[0057] In step (1), the position information corresponding to a single energy storage unit is the number of the horizontal and vertical coordinates of the energy storage unit; the three-dimensional information of each energy storage unit is the temperature value of the energy storage unit, the normalized horizontal coordinate number value, and the normalized vertical coordinate number value. The normalization method used for normalization can be reasonably selected according to needs, such as using the norm function to normalize the horizontal coordinate number value and the vertical coordinate number value, thereby obtaining the normalized horizontal coordinate number value and the vertical coordinate number value.

[0058] In step (2), the three-dimensional information of each energy storage unit is converted into a temperature distribution matrix, and the specific method is as follows:

[0059] The three-dimensional information of each energy storage unit is converted into a two-dimensional matrix. The values ​​in the matrix are temperature values. The rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​of the collected energy storage unit.

[0060] In step (2), the data in the temperature distribution matrix is ​​then segmented and the diffusion coefficient is calculated to obtain the diffusion coefficient; the specific method is:

[0061] Calculate the abnormal diffusion coefficient of the i-th energy storage unit as D i :

[0062] D i =E i *exp(1-F i )

[0063] Among them, E i is the temperature difference of the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit;

[0064] E i The calculation method is:

[0065] Obtain the energy storage units that are currently adjacent to the i-th energy storage unit and are in the same operating state, mark these energy storage units as 1 in the two-dimensional matrix, and mark the remaining energy storage units as 0;

[0066] Extract the center of the area with value 1, use the extracted center as the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit, obtain the temperature value of the center at the current moment, and obtain the three-dimensional data of the center; the center is the center of the common charging and discharging energy storage unit;

[0067] Calculate the absolute value of the difference between the current temperature of the center and the previous temperature of the center, and use the ratio of this absolute value to the current temperature of the center as E. i ;

[0068] F i It is the ratio of the Euclidean distance between the ith energy storage unit and the center of the common charging and discharging energy storage unit to the longest distance. The longest distance refers to the maximum value of the Euclidean distances between all energy storage units and the center of the common operating energy storage unit.

[0069] The common operating state refers to the same charging or discharging.

[0070] In step (3), the abnormal value of the temperature sensor acquisition value corresponding to the energy storage unit is calculated according to the diffusion coefficient of the energy storage unit, and the specific method is:

[0071] Calculate the abnormal value of the temperature sensor data collected by the i-th energy storage unit as H i :

[0072] H i =exp(-L i )*min(D i , D′);

[0073] in,

[0074] D' is the average value of the abnormal diffusion coefficient of the i-th energy storage unit and all the energy storage units adjacent to it; L iis the mutation value of the abnormal diffusion coefficient around the i-th energy storage unit.

[0075] In step (3), the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix. The specific method is:

[0076] Obtain the abnormal values ​​of all energy storage units at each moment, and obtain the abnormal value set {H} of all energy storage units at that moment; perform binary classification on {H} at each moment; calculate the mean of the abnormal values ​​of the energy storage units corresponding to the two categories of the binary classification respectively; take the category with the larger mean as the high H category; take the category with the smaller mean as the low H category;

[0077] Then, all energy storage units in the high H category are assigned a value, and the specific value assigned is W=1;

[0078] Assign values ​​to all energy storage units in the low H category, specifically W = 0;

[0079] For each energy storage unit, the W at all times during the data collection process is accumulated to obtain the accumulated value W';

[0080] According to the normalized horizontal coordinate number value, the normalized vertical coordinate number value and the accumulated value W' of each energy storage unit, it is converted into an accumulated value matrix;

[0081] The values ​​in the accumulated value matrix are accumulated values, and the rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​after the energy storage unit is collected and normalized.

[0082] The accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistics of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the variance values ​​corresponding to the two dimensions in the two-dimensional Gaussian function as the final temperature anomaly value; the specific method is as follows:

[0083] The window size is 3*3; the W′ values ​​of all energy storage units in the window are added as the density statistics of the energy storage unit at the center of the window; the Gaussian mixture model is used to fit the multivariate two-dimensional Gaussian function to the density statistics; the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function is obtained as the final temperature anomaly value.

[0084] Example 3

[0085] A method for safety control of an energy storage power station based on energy storage unit status assessment comprises the following steps:

[0086] Step (1), collecting the temperature value of a single energy storage unit and obtaining the position information corresponding to the single energy storage unit, thereby obtaining three-dimensional information of each energy storage unit;

[0087] Among them, a temperature sensor is installed inside a single energy storage unit. The temperature sensor sends the collected temperature data to the data processing center through wired and / or wireless means. The data processing center performs two-dimensional coordinate numbering on the position coordinate information of the single energy storage unit. Since the energy storage units are arranged in a regular manner, the two-digit numbers are sequentially increased from 1 in the horizontal and vertical coordinate dimensions to construct a two-dimensional coordinate system of the energy storage unit. Then, the two-dimensional coordinate information of the final entire energy storage unit in the energy storage station is normalized separately, and the three-dimensional information A, B, C corresponding to each energy storage unit can be obtained; the three-dimensional information corresponding to the i-th energy storage unit is A i ,B i ,C i ; Among them, the temperature value of the i-th energy storage unit is A i , the normalized horizontal axis is B i , the normalized vertical coordinate is C i .

[0088] Step (2), converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; then segmenting the data in the temperature distribution matrix and calculating the diffusion coefficient to obtain the diffusion coefficient;

[0089] The three-dimensional information of each energy storage unit is converted into a two-dimensional matrix. The values ​​in the matrix are temperature values. The rows and columns of the matrix correspond to the normalized horizontal and vertical coordinates of the energy storage unit. With the use of the energy storage battery, the temperature distribution gradually diffuses due to the thermal conduction of the temperature. When the heat generation and heat dissipation reach a balance, the temperature will no longer rise and the temperature distribution will be relatively stable.

[0090] When the temperature distribution is stable, the abnormality detection of the energy storage unit can be achieved through the temperature change in continuous time and the local temperature equilibrium state.

[0091] However, when the energy storage unit is just started and the temperature is in a diffusion state, the abnormal detection of the energy storage unit at this time needs to consider the influence of the normal diffusion of the temperature. If the abnormal detection of the energy storage unit can be completed at this time, it is helpful to sense the abnormality of the energy storage unit in advance, reduce the abnormal risk of the energy storage unit, and improve the safety of the energy storage station.

[0092] The diffusion coefficient of each temperature is obtained and used to correct the subsequent temperature anomaly detection results to obtain accurate temperature detection results.

[0093] The abnormal diffusion coefficient of the i-th energy storage unit is D i , the calculation formula is as follows:

[0094] D i =E i *exp(1-Fi )

[0095] Among them, E i is the temperature difference of the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit, E i The larger the value, the greater the temperature change at the center of the energy storage unit being charged and discharged. The temperature should be in the rising or falling stage. When the temperature rises, the energy storage unit is enabled, and when the temperature drops, the energy storage unit is not used.

[0096] E i The calculation method of is as follows: obtain the energy storage units that are currently adjacent to the i-th energy storage unit and are in the same operating state (the operating state is the same as charging or discharging), mark these energy storage units as 1 in the two-dimensional matrix, and mark the remaining energy storage units as 0;

[0097] Extract the center of the area with value 1, use the extracted center as the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit, obtain the temperature value of the center at the current moment, and obtain the three-dimensional data of the center; the center is the center of the common charging and discharging energy storage unit;

[0098] Calculate the absolute value of the difference between the current temperature of the center and the previous temperature of the center, and use the ratio of this absolute value to the current temperature of the center as E. i ;

[0099] The temperature value of the center at the previous moment is the temperature value of the center extracted at the previous moment according to the same method as above, that is, the center at the previous moment is re-extracted.

[0100] Due to the different distances from the battery boundaries in the operating state, the contact with the heat sink is different, and the heat conduction effect is unstable relative to the current center temperature.

[0101] The reason why the maximum temperature is not directly used as the center value is that although the same batch of batteries are working together, the states of different batteries are different, which will lead to instability, and due to the conductivity of temperature, there will be more temperature values ​​equal to the maximum value, which may lead to inaccurate judgment. Therefore, the present invention selects the center of the energy storage batteries as the temperature judgment standard, because the center is often the middle of the energy storage battery distribution and the worst heat dissipation position of the operating energy storage batteries. Therefore, the center is preferably used as the monitoring position point.

[0102] E iis the temperature difference of the center of the energy storage unit that is charged and discharged together with the i-th energy storage unit. However, if the distance between the i-th energy storage unit and the temperature center of the energy storage unit that is operated together is farther, the temperature change of the center is smaller, that is, it is an energy storage unit closer to the heat dissipation point, and thus it may have a larger temperature change. At this time, even if the temperature detection data has a larger E i value, which is also a relatively normal value; if the distance is closer, the distance affected by the heat dissipation factor is smaller, then E i Should have a smaller value; if the distance is closer, E i The larger the value is, the more likely the current energy storage unit is to have an abnormality, and then the abnormal diffusion coefficient D of the i-th energy storage unit is obtained. i =E i *exp(1-F i );

[0103] Among them, F i It is the ratio of the Euclidean distance between the ith energy storage unit and the center of the common charging and discharging energy storage unit to the longest distance. The longest distance refers to the maximum value of the Euclidean distances between all energy storage units and the center of the common operating energy storage unit.

[0104] The Euclidean distance is calculated by using the normalized horizontal and vertical coordinates of the temperature center of the common operating energy storage unit and the horizontal and vertical coordinates of the i-th energy storage unit to calculate the Euclidean distance between two points.

[0105] When E i The larger the F i The smaller the time, the i The larger the value, the more obvious the temperature anomaly diffusion is. i The larger the F i The larger the D i Towards the middle, the temperature anomaly diffusion is relatively normal at this time; E i The smaller the F i The larger the D i The smaller the value, the more normal the temperature anomaly diffusion is; i The smaller the F i The smaller the D i Towards the middle, the diffusion of temperature anomalies is relatively normal.

[0106] Step (3), performing local anomaly division according to the diffusion coefficient of a single energy storage unit nearby, and obtaining an abnormal value of a single temperature sensor.

[0107] After obtaining D of the i-th energy storage unit iFinally, if the temperature sensor is abnormal, there will only be a single point with a large abnormal diffusion coefficient; if the energy storage unit is indeed abnormal, due to thermal conductivity, a high abnormal diffusion coefficient will also appear around the single energy storage unit, but because the outermost energy storage unit of the energy storage unit working together has a better contact with the heat dissipation effect, the heat conduction path, whether it is air cooling or liquid cooling, may cause changes in the heat dissipation path during the startup or shutdown stage, and at this time there will also be a large abnormal diffusion coefficient value, and therefore it is necessary to reduce the temperature abnormality monitoring interference caused by heat dissipation when realizing temperature monitoring.

[0108] Because the operating power and heat dissipation power are different each time, it is difficult for the energy storage units that work together to have temperature anomalies at the same time and continuously. The greater the position in the energy storage units that work together and are in contact with the heat dissipation, the temperature change at that position should be relatively similar even if there is a change. If there is a large difference in the change, it also indicates that there is a temperature anomaly in the energy storage unit.

[0109] The abnormal value of the temperature sensor value corresponding to the i-th energy storage unit is H i :

[0110] H i =exp(-L i )*min(D i , D′)

[0111] Among them, L i is the mutation value of the abnormal diffusion coefficient around the i-th energy storage unit; L i The acquisition method is to obtain the ratio of the average value D′ of the abnormal diffusion coefficient D value of the energy storage unit within the domain (usually 8 adjacent ones), and calculate the absolute value of the difference between the ratio and 1. Where L i The larger it is, the less similar the abnormal diffusion coefficient D values ​​of the surrounding energy storage units are, which means that the mutation of the abnormal diffusion coefficient D of the energy storage units within the field is higher.

[0112] If the temperature sensor fails at this time, L i The larger the value of is, the D value of the energy storage unit around it is relatively uniform, and the D' value is low, and the D' value is in the normal range; if the temperature of the i-th energy storage unit is abnormal, its L i The value is relatively low. Since heat conduction diffuses all around, the D value of the surrounding energy storage units is relatively uniform, but the D' value of the surrounding energy storage units will be relatively large. However, if the D value itself is small (a normal energy storage unit), it is just close to the heat dissipation change area or close to the mutation point. In order to prevent the small D value from causing greater interference to the normal energy storage unit, min(D i , D′). Among them, exp(-Li ) indicates L i Map negative correlations and prevent them from being zero.

[0113] H i The larger the value, the more likely it is that the temperature is abnormal. The abnormal value H of each energy storage unit at each moment in the data collection process is obtained by the above method. i ;

[0114] Obtain the abnormal values ​​of all energy storage units in the collected data at each moment, and obtain the abnormal value set {H} of all energy storage units at that moment;

[0115] Perform binary classification on {H} at each moment, preferably using k-means algorithm for binary classification; calculate the mean of the outliers of the two categories of the binary classification; take the category with the larger mean as the high H category; take the category with the smaller mean as the low H category;

[0116] Then, all energy storage units in the high H category are assigned a value, and the specific value assigned is W=1;

[0117] Assign values ​​to all energy storage units in the low H category, specifically W = 0;

[0118] For each energy storage unit, add up W at all times during the data collection process to get W';

[0119] For example, there are 10 moments in the data collection process. Then, for each energy storage unit, W'=W1+…+W 10 ; W1 represents the W value of the energy storage unit at the first moment; W 10 Indicates the W value of the energy storage unit at the 10th moment.

[0120] The high H classification results in 10 consecutive rounds of collected data are accumulated, where 10 is a hyperparameter that can be adjusted according to the specific implementation scenario. The accumulated result is the accumulation of the high H classification result values ​​of the energy storage unit in 10 rounds W'=W1+…+W 10 , so the maximum value of W' does not exceed 10, and W1 represents the value of the energy storage unit with the high H classification result in the first round of collected data;

[0121] According to the normalized horizontal coordinate number value, the normalized vertical coordinate number value and the accumulated value W' of each energy storage unit, it is converted into an accumulated value matrix; the values ​​in the matrix are accumulated, and the rows and columns of the matrix correspond to the normalized horizontal and vertical coordinates of the energy storage unit;

[0122] If W' of all energy storage units is random, their density distribution is relatively uniform; if there is an energy storage unit with a concentrated density distribution and a Gaussian data distribution, it means that there is an energy storage unit with abnormal temperature.

[0123] With the energy storage unit as the center, the sum of the W' values ​​within the 3*3 window range is counted; windowing is to increase the density value when the data is statistically dense, to prevent poor density statistics due to excessive sparseness, and because its stability itself has a certain variability, windowing can make the result relatively stable.

[0124] Specifically, the W' values ​​of all energy storage units in the window are added together as the statistical result of the energy storage unit density at the center of the window; that is, the eight neighborhoods plus themselves add up to a total of 9 W' (the eight neighborhoods are a circle around them);

[0125] For the density statistics, a Gaussian mixture model is used to fit a multivariate two-dimensional Gaussian function; the Gaussian mixture model is an existing model and will not be described in detail in this invention.

[0126] Two-dimensional Gaussian fitting is used because the coordinates are two-dimensional data;

[0127] The number of Gaussians used in the Gaussian mixture model is determined based on the accumulated values ​​in the matrix and the horizontal and vertical coordinate number values ​​of the normalized energy storage units corresponding to the rows and columns of the matrix. The accumulated results can be classified by the mean shift algorithm, and the number of classification results of the mean shift algorithm is used as the number of Gaussians.

[0128] The mean shift is selected because it is a density-based clustering. When using the mean shift algorithm for classification, the preferred cluster radius is 0.06; it is set to such a small value of 0.06 because the horizontal and vertical coordinates have been normalized. Those skilled in the art should know that the mean shift algorithm can be replaced by other density clustering algorithms or other methods for determining the number of Gaussians in a Gaussian mixture model.

[0129] All two-dimensional Gaussian function fitting results F(x,y) are obtained, where F is the Gaussian function fitting function, x, y are two variance values, and the maximum value of the variance values ​​corresponding to the two dimensions in the two-dimensional Gaussian function is obtained as the final temperature anomaly value = max(x,y); the reason for using the Gaussian mixture model for fitting is that the data distribution of the accumulated high H classification results in the cumulative result fits the coordinate axis direction of the Gaussian distribution uncertainly. The variance value is used as the final temperature anomaly value because the smaller the variance, the greater the concentration of the data distribution.

[0130] The present invention uses the variance value as a temperature anomaly judgment standard. A threshold value of the final temperature anomaly value is set. When the final temperature anomaly value is less than or equal to the final temperature anomaly value threshold, it indicates that the temperature of the current reserve unit is abnormal and needs to be controlled, thereby completing the safety management of the energy storage station.

[0131] The present invention does not specifically limit the threshold of the final abnormal temperature value, which can be set according to actual conditions.

[0132] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for safety control of energy storage power station based on energy storage unit status assessment, characterized in that: The steps include: Step (1), collecting the temperature value of a single energy storage unit and obtaining the position information corresponding to the single energy storage unit, thereby obtaining three-dimensional information of each energy storage unit; Step (2), converting the three-dimensional information of each energy storage unit into a temperature distribution matrix; then segmenting the data in the temperature distribution matrix and calculating the diffusion coefficient to obtain the diffusion coefficient; Step (3), according to the diffusion coefficient of the energy storage unit, the abnormal value of the temperature sensor acquisition value corresponding to it is calculated, and the abnormal values ​​of all energy storage units are binarized, and then the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix; the accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistic value of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function as the final temperature abnormal value; if the final temperature abnormal value is less than or equal to the final temperature abnormal value threshold, it means that the temperature of the current storage unit is abnormal, and control processing is required to eliminate the abnormality.

2. The energy storage power station safety control method based on energy storage unit status assessment according to claim 1 is characterized in that: In step (1), the temperature of each energy storage unit is collected by a temperature sensor.

3. The energy storage power station safety control method based on energy storage unit status assessment according to claim 1 is characterized in that: In step (1), the position information corresponding to a single energy storage unit is the number of the horizontal and vertical coordinates of the energy storage unit; the three-dimensional information of each energy storage unit is the temperature value of the energy storage unit, the normalized horizontal coordinate number value, and the normalized vertical coordinate number value.

4. The energy storage power station safety control method based on energy storage unit status assessment according to claim 1 is characterized in that: In step (2), the three-dimensional information of each energy storage unit is converted into a temperature distribution matrix, and the specific method is as follows: The three-dimensional information of each energy storage unit is converted into a two-dimensional matrix. The values ​​in the matrix are temperature values. The rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​of the collected energy storage unit.

5. The energy storage power station safety control method based on energy storage unit status assessment according to claim 1 is characterized in that: In step (2), the data in the temperature distribution matrix is ​​then segmented and the diffusion coefficient is calculated to obtain the diffusion coefficient; the specific method is: Calculate the abnormal diffusion coefficient of the i-th energy storage unit as D i : D i =E i *exp(1-F i ) Among them, E i is the temperature difference of the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit; E i The calculation method is: Obtain the energy storage units that are currently adjacent to the i-th energy storage unit and are in the same operating state, mark these energy storage units as 1 in the two-dimensional matrix, and mark the remaining energy storage units as 0; Extract the center of the area with value 1, use the extracted center as the center of the common charging and discharging energy storage unit corresponding to the i-th energy storage unit, obtain the temperature value of the center at the current moment, and obtain the three-dimensional data of the center; the center is the center of the common charging and discharging energy storage unit; Calculate the absolute value of the difference between the current temperature of the center and the previous temperature of the center, and use the ratio of this absolute value to the current temperature of the center as E. i ; F i It is the ratio of the Euclidean distance between the ith energy storage unit and the center of the common charging and discharging energy storage unit to the longest distance. The longest distance refers to the maximum value of the Euclidean distances between all energy storage units and the center of the common operating energy storage unit.

6. The energy storage power station safety control method based on energy storage unit status assessment according to claim 5 is characterized in that: The common operating state refers to the same charging or discharging.

7. The energy storage power station safety control method based on energy storage unit status evaluation according to claim 5 is characterized in that: In step (3), the abnormal value of the temperature sensor acquisition value corresponding to the energy storage unit is calculated according to the diffusion coefficient of the energy storage unit, and the specific method is: Calculate the abnormal value of the temperature sensor data collected by the i-th energy storage unit as H i : H i =exp(-L i )*min(D i ,D′); in, D' is the average value of the abnormal diffusion coefficient of the i-th energy storage unit and all the energy storage units adjacent to it; L i is the mutation value of the abnormal diffusion coefficient around the i-th energy storage unit.

8. The energy storage power station safety control method based on energy storage unit status evaluation according to claim 7 is characterized in that: In step (3), the binarization results of each energy storage unit in the acquisition process are accumulated, and then the position information corresponding to each energy storage unit is used to obtain the accumulated value matrix. The specific method is: Obtain the abnormal values ​​of all energy storage units at each moment, and obtain the abnormal value set {H} of all energy storage units at that moment; perform binary classification on {H} at each moment; calculate the mean of the abnormal values ​​of the energy storage units corresponding to the two categories of the binary classification respectively; take the category with the larger mean as the high H category; take the category with the smaller mean as the low H category; Then, all energy storage units in the high H category are assigned a value, and the specific value assigned is W=1; Assign values ​​to all energy storage units in the low H category, specifically W = 0; For each energy storage unit, the W at all times during the data collection process is accumulated to obtain the accumulated value W'; According to the normalized horizontal coordinate number value, the normalized vertical coordinate number value and the accumulated value W' of each energy storage unit, it is converted into an accumulated value matrix; The values ​​in the accumulated value matrix are accumulated values, and the rows and columns of the matrix correspond to the normalized horizontal coordinate number values ​​and the normalized vertical coordinate number values ​​after the energy storage unit is collected and normalized.

9. The energy storage power station safety control method based on energy storage unit status assessment according to claim 8 is characterized in that: The accumulated value matrix is ​​windowed, and the sum of the accumulated values ​​in the window is used as the density statistics of the energy storage unit at the center of the window; then a two-dimensional Gaussian fitting is used to obtain the maximum value of the variance values ​​corresponding to the two dimensions in the two-dimensional Gaussian function as the final temperature anomaly value; the specific method is as follows: The window size is 3*3; the W' values ​​of all energy storage units in the window are added as the density statistics of the energy storage unit at the center of the window; the Gaussian mixture model is used to fit the multivariate two-dimensional Gaussian function to the density statistics; the maximum value of the corresponding variance values ​​of the two dimensions in the two-dimensional Gaussian function is obtained as the final temperature anomaly value.

Images