A power distribution box environmental parameter linkage regulation method and system

Abstract

The application provides a power distribution box environment parameter linkage regulation method and system, comprising: collecting power distribution box environment parameter time series, obtaining parameter reverse measurement value and Pearson correlation coefficient absolute value, and constructing a hybrid similarity matrix; adjusting the hybrid similarity matrix by using a preset importance coefficient, and obtaining an importance adjusted graph Laplacian matrix in combination with a degree matrix; performing characteristic decomposition on the matrix, extracting a second and third small eigenvalue and a corresponding eigenvector; generating a polarization angle vector by using an eigenvalue weighted eigenvector, and obtaining a radial position vector by mapping a second eigenvector component index; generating each environment parameter mapping point in a polar coordinate system according to the two types of vectors, converting into a circular sequence in a Cartesian coordinate system, determining a clustering segmentation boundary according to the Euclidean distance of adjacent mapping points in the circular sequence, classifying the environment parameters into each linkage regulation group, matching and executing a corresponding control strategy, and completing power distribution box environment linkage regulation.

Description

Technical Field

[0001] This application belongs to the field of parameter linkage control, and in particular relates to a method and system for linkage control of environmental parameters in a distribution box. Background Technology

[0002] In real-world operating environments, there are interconnections and interactions within electrical distribution boxes. For example, an increase in temperature may cause changes in relative humidity, or changes in the concentration of certain gases may be accompanied by alterations in heat distribution. Addressing the correlation between parameters by employing a coordinated control strategy to group strongly correlated environmental parameters into the same group, and then matching this group with appropriate integrated control strategies, has become a standard technical approach in the field of electrical distribution box environmental management.

[0003] In the process of coordinated control of environmental parameters in distribution boxes, the grouping of environmental parameters typically relies on basic data similarity calculations or clustering models. However, when constructing parameter association networks, the features of key parameters may become marginalized during the calculation process. Furthermore, spatial partitioning methods struggle to identify high-dimensional implicit structural features when processing parameter association data with topological structures, resulting in unclear clustering boundaries and affecting the rationality of environmental parameter grouping. Therefore, how to measure the mixed correlations between environmental parameter time series and adjust them in conjunction with parameter importance, mine the intrinsic structural features between parameters to determine the grouping boundaries for coordinated control, and achieve a clear division of environmental parameter-dominated linkage relationships and matching of control strategies are technical problems that need to be solved in the field of distribution box environmental management. Summary of the Invention

[0004] This invention proposes a method for coordinated control of environmental parameters in a distribution box, which addresses the problem that existing technologies cannot achieve reasonable allocation of environmental parameters and matching of control strategies, including: Obtain time series data of environmental parameters inside the distribution box, and determine the environmental parameters participating in the linkage group calculation; for any two environmental parameters participating in the calculation, calculate their inverse metric value and the absolute value of the time series Pearson correlation coefficient, and construct a hybrid similarity matrix accordingly; The mixed similarity matrix is ​​adjusted based on the preset importance coefficients of each environmental parameter to obtain the importance-adjusted weight matrix. A degree matrix is ​​constructed based on the importance-adjusted weight matrix, and the importance-adjusted graph Laplacian matrix is ​​calculated from the degree matrix and the importance-adjusted weight matrix. The graph Laplacian matrix is ​​then subjected to eigenvalue decomposition to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvector and third eigenvector. The second and third feature vectors are weighted and combined according to the second and third minor feature values ​​to generate a polarization angle vector. The components of the second feature vector are exponentially mapped to form a radial position vector. Based on the polarization angle vector and the radial position vector, mapping points of all environmental parameters in the polar coordinate system are constructed. All the mapping points are converted into a ring sequence in the Cartesian coordinate system. The Euclidean distance between adjacent mapping points in the ring sequence is used to determine the clustering segmentation boundary. The environmental parameters are divided into linkage control groups through the clustering segmentation boundary. Control strategies are matched for the linkage control groups and executed to realize the linkage control of the distribution box environment.

[0005] Furthermore, the present invention also relates to a power distribution box environmental parameter linkage control system, comprising the following modules: The module is used to acquire time series data of environmental parameters in the distribution box, and determine the environmental parameters to be included in the linkage group calculation. For any two environmental parameters to be included in the calculation, the reverse metric value and the absolute value of the time series Pearson correlation coefficient are calculated, and a hybrid similarity matrix is ​​constructed accordingly. An extraction module is used to adjust the mixed similarity matrix based on preset importance coefficients of various environmental parameters to obtain an importance-adjusted weight matrix, construct a degree matrix based on the importance-adjusted weight matrix, and calculate an importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix; perform eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvector and third eigenvector; The control module is used to weight and combine the second and third feature vectors according to the second and third small feature values ​​to generate a polarization angle vector, and to perform exponential mapping on each component of the second feature vector to form a radial position vector; based on the polarization angle vector and the radial position vector, it constructs mapping points of all environmental parameters in the polar coordinate system, converts all the mapping points into a ring sequence in the Cartesian coordinate system, determines the clustering boundary based on the Euclidean distance between adjacent mapping points in the ring sequence, divides the environmental parameters into linkage control groups through the clustering boundary, matches control strategies for the linkage control groups and executes them to realize linkage control of the distribution box environment.

[0006] This invention constructs a hybrid similarity matrix by fusing inverse metrics and time-series Pearson correlation coefficients, enabling the measurement of the correlation between multi-source environmental parameters in a distribution box. The importance coefficient-adjusted Laplace matrix balances the weights of different parameters in real-world scenarios. By extracting the second and third smallest eigenvalues ​​and their corresponding eigenvectors, polarization angle vectors and radial position vectors are constructed, mapping the environmental parameter features to a polar coordinate system and converting them into a circular sequence in a Cartesian coordinate system, thus transforming the data features into an easily distinguishable geometric spatial structure. This mechanism determines two dominant segmentation boundaries based on the intervals of the circular sequences after spectral feature mapping, dividing the environmental parameters into two dominant linkage control groups. Within each linkage control group, temperature and humidity, gas safety, and electrical safety parameters are further identified, and corresponding control commands are matched. This allows the two grouping results to adapt to multi-category control requirements, thereby improving the accuracy of linkage control strategy matching. Attached Figure Description

[0007] Figure 1 A schematic diagram of topological and spectral feature decomposition of node similarity for environmental parameters; Figure 2 This is a schematic diagram of polar coordinate mapping clustering for environmental parameters. Detailed Implementation

[0008] In this application, the terms "exemplarily" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplarily" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0009] In the first embodiment, the present invention proposes a method for linkage control of environmental parameters of a power distribution box, comprising: S1. Obtain time series data of environmental parameters inside the distribution box, and determine the environmental parameters participating in the linkage group calculation. For any two environmental parameters participating in the calculation, calculate their inverse metric value and the absolute value of the time series Pearson correlation coefficient, and construct a hybrid similarity matrix accordingly.

[0010] Inside the distribution box, environmental data within a set time period is collected synchronously according to a set sampling frequency, forming time series data corresponding to each environmental parameter. The environmental parameters include physical environmental parameters and electrical safety parameters. The physical environmental parameters include temperature, humidity, smoke concentration, and harmful gas concentration. The electrical safety parameters include voltage and leakage current. The electrical safety parameters are subclasses of the environmental parameters and participate in subsequent hybrid similarity matrix construction and linkage grouping calculation. The number N of environmental parameters participating in the linkage grouping calculation is an integer greater than or equal to 3. The collected time series data is normalized to reduce the influence of dimensions.

[0011] For any two normalized environmental parameter time series, the numerical mean of each environmental parameter within a preset time period is calculated. The absolute value of the difference between the two numerical means is taken, and the product exponent is obtained by multiplying the absolute value of the difference by the inverse of a preset scale decay constant β. A reverse metric is then calculated using the natural constant as the base and the product exponent as the exponent. The reverse metric characterizes the degree of reverse similarity between the means of the two environmental parameters; a larger value indicates that the means of the two environmental parameters are closer within the preset time period.

[0012] Calculate the covariance of the two environmental parameter time series, and calculate the standard deviation of the two environmental parameter time series respectively. Divide the covariance by the product of the two standard deviations to obtain the Pearson correlation coefficient of the time series. Take the absolute value of the Pearson correlation coefficient of the time series to obtain the absolute value of the Pearson correlation coefficient of the time series.

[0013] Set a first weight coefficient and a second weight coefficient, both of which are greater than or equal to 0, and the sum of the first weight coefficient and the second weight coefficient is 1. Multiply the absolute value of the time series Pearson correlation coefficient by the first weight coefficient, then multiply the inverse metric value by the second weight coefficient, and add the resulting product to the first product to obtain the mixed similarity between the two environmental parameters. Iterate through all environmental parameters participating in the linkage group calculation to calculate the mixed similarity between each pair. The mixed similarity is used to construct a mixed similarity matrix with the number of rows and columns equal to the number of environmental parameters participating in the linkage group calculation. The diagonal elements of the matrix are set to 0, indicating that the similarity of the environmental parameters themselves does not participate in the graph edge weight calculation.

[0014] In an optional embodiment, the step of calculating the inverse metric and the absolute value of the time-series Pearson correlation coefficient for any two environmental parameters involved in the calculation, and constructing a mixed similarity matrix accordingly, includes: For any two environmental parameters a and b, extract the normalized time series data. and ; calculate and The absolute value of the covariance is calculated separately. variance and The variances are multiplied, and the square root of the product is obtained. The absolute value of the covariance is then divided by the square root of the product to obtain the absolute value of the Pearson correlation coefficient of the time series. ; Calculate the mean values ​​of environmental parameter a and environmental parameter b within a preset time period after normalization, and obtain the results. and ; then calculate minus The absolute value of the difference is multiplied by the inverse of the preset scale attenuation constant β to obtain the product exponent. The inverse metric is then calculated using the natural constant as the base and the product exponent as the exponent. ; The preset first weighting coefficient The absolute value of the Pearson correlation coefficient with the time series Multiply to obtain the first product, then use the preset second weighting coefficients. With the reverse metric value Multiplying them yields a second product. Adding the first product to the second product gives the mixture similarity between environmental parameters a and b. The first weighting coefficient With the second weighting coefficient Both are greater than or equal to 0, and their sum is 1, which is determined by all mixed similarities. This constitutes the hybrid similarity matrix.

[0015] Taking the ambient temperature parameter 'a' and relative humidity parameter 'b' inside the distribution box as examples, assuming a sampling period of 1 minute, 60 sampling points from the past 60 minutes are extracted and normalized using the Min-Max method to obtain a time series vector of length 60. and ,in , Calculate the absolute value of the covariance between the two. For example, the calculated result is 0.012; calculate separately. variance and variance The square root of the product of the two is 0.0173. The absolute value of the Pearson correlation coefficient for the time series is calculated by dividing 0.012 by 0.0173. .

[0016] To calculate the reverse metric, first calculate the mean of the normalized temperature and humidity sequence over a preset 60-minute period. Assume the calculated mean temperature... average humidity Take the absolute value of the difference between the two, which is 0.16. Set the scale decay constant. The product is 3.0. Multiplying its opposite, -3.0, by 0.16 yields a product exponent of -0.48. Using the natural constant e as the base and -0.48 as the exponent, the inverse measure is calculated. Set the first weighting coefficient. Second weighting coefficient The sum of the two is 1. Then, 0.6 and... Multiplying them gives the first product, 0.4162. Then multiply 0.4 by... Multiplying them yields a second product of 0.2475. Adding the two together gives the mixture similarity between temperature and humidity. Perform the above calculation on all parameter pairs involved in the linkage grouping calculation, and calculate all the mixed similarities. This forms a complete N×N dimensional hybrid similarity matrix.

[0017] S2, adjust the hybrid similarity matrix based on the preset importance coefficients of each environmental parameter to obtain the importance-adjusted weight matrix, construct the degree matrix according to the importance-adjusted weight matrix, and calculate the importance-adjusted graph Laplacian matrix from the degree matrix and the importance-adjusted weight matrix; perform eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue and their corresponding second eigenvector and third eigenvector.

[0018] An importance coefficient greater than 0 is pre-assigned based on the degree of influence of each environmental parameter in the distribution box on safe operation.

[0019] For the element in the i-th row and j-th column of the mixed similarity matrix, multiply it by the product of the importance coefficients corresponding to the i-th and j-th environmental parameters. After traversing and replacing all elements, the importance-adjusted edge weight matrix is ​​obtained.

[0020] Calculate the sum of the elements in each row of the importance-adjusted edge weight matrix, use the sum as the diagonal elements, and set all off-diagonal elements to zero to construct the corresponding degree matrix.

[0021] Subtracting the importance-adjusted edge weight matrix from the degree matrix yields the importance-adjusted graph Laplacian matrix.

[0022] The Jacobi method is used to solve for all eigenvalues ​​and corresponding eigenvectors of the importance-adjusted graph Laplacian matrix. All eigenvalues ​​are sorted in ascending order from smallest to largest. The eigenvalue in the second position is extracted as the second smallest eigenvalue, and the eigenvector corresponding to the second smallest eigenvalue is extracted as the second eigenvector. The eigenvalue in the third position is extracted as the third smallest eigenvalue, and the eigenvector corresponding to the third smallest eigenvalue is extracted as the third eigenvector.

[0023] In an optional embodiment, adjusting the mixed similarity matrix based on preset importance coefficients of each environmental parameter to obtain an importance-adjusted weight matrix, constructing a degree matrix based on the importance-adjusted weight matrix, and calculating the importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix include: Obtain the preset importance coefficient of environmental parameter a And the preset importance coefficient of environmental parameter b ; The preset importance coefficient With the preset importance coefficient and the mixed similarity before adjustment After adjusting the weights by all importance We obtain the importance-adjusted edge weight matrix W; For the i-th row of the importance-adjusted weight matrix W, calculate the sum of all importance-adjusted weights in the i-th row to obtain the degree of the i-th environmental parameter node. Construct a diagonal matrix of the degrees of all environmental parameter nodes as the degree matrix D, where the number of environmental parameters N participating in the linkage group calculation is an integer greater than or equal to 3, and i is an integer greater than zero and less than or equal to the number of environmental parameters N participating in the linkage group calculation. Subtracting the importance-adjusted edge weight matrix W from the degree matrix D yields the importance-adjusted graph Laplacian matrix L.

[0024] Four core linkage control parameters (N=4) are selected from the collected environmental parameters: temperature, humidity, voltage, and leakage current, with preset importance coefficients of 1.2, 0.8, 1.0, and 1.5, respectively. When processing temperature, parameter a... , and humidity, parameter b, When considering the relationship, extract the unadjusted mixed similarity calculated above. ,Will , and Multiply by: 1.2 × 0.8 × 0.6637, to obtain the edge weights after importance adjustment between temperature and humidity nodes. Since the preset importance coefficient is used to amplify or weaken the influence of different environmental parameter nodes on the linkage grouping, the weights after the importance adjustment are used to construct the graph Laplacian matrix, and are not required to be limited to the interval between 0 and 1. This operation is repeated for all parameter pairs to obtain a symmetric 4×4-dimensional importance adjustment weight matrix W, with the weights on the main diagonal set to 0.

[0025] After adjusting the importance of the generated weighted matrix W, the sum of the elements in each row is calculated to obtain the degree of each environmental parameter node. Taking the first row containing temperature as an example, assuming the weights of temperature, humidity, voltage, and leakage current in the first row of the matrix are [0, 0.6372, 0.4210, 0.8520], the sum of these four values ​​is calculated as 1.9102, which is taken as the degree of the first environmental parameter node. Similarly, the degrees of the other three rows are calculated, for example, 1.5201, 1.8450, and 2.1334 respectively. These degrees are then placed on the main diagonal to construct a 4×4 degree matrix D. Subtracting the weighted matrix W from the degree matrix D yields a 4×4 importance-adjusted graph Laplacian matrix L containing network topology and node importance features. After constructing the weighted parametric topology network, eigenvalue decomposition is performed on the graph Laplacian matrix. The principle of this process is as follows: Figure 2 As shown, the numerical values ​​of the connections between nodes represent the importance of the parameters. After adjusting the edge weights, the eigenvalues ​​and eigenvectors are obtained.

[0026] In an optional embodiment, the step of performing eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvectors and third eigenvectors includes: Solving the characteristic equation of the graph Laplacian matrix L yields N eigenvalues ​​and their corresponding eigenvectors, where N is an integer greater than or equal to 3; The N feature values ​​are arranged in ascending order from smallest to largest, while maintaining the correspondence between feature vectors and feature values; Remove the eigenvector corresponding to the smallest eigenvalue, extract the second-ranked eigenvalue as the second smallest eigenvalue, and extract the eigenvector corresponding to the second smallest eigenvalue as the second eigenvector. The third-ranked feature value is extracted as the third smallest feature value, and the feature vector corresponding to the third smallest feature value is extracted as the third feature vector.

[0027] Since the second and third smallest eigenvalues ​​need to be extracted later, the number N of environmental parameters participating in the linkage group calculation is set to an integer greater than or equal to 3. When the number of collected environmental parameters exceeds 3, all collected environmental parameters can be selected as the environmental parameters participating in the linkage group calculation, or at least 3 core linkage control parameters can be selected as the environmental parameters participating in the linkage group calculation. Solving the characteristic equation of the 4×4 dimensional Laplacian matrix L generated above yields N=4 eigenvalues. For example, the result obtained is: , , , At the same time, four 4-dimensional feature vectors corresponding to these four feature values ​​are obtained. , , , The four extracted feature values ​​are sorted in ascending order according to their numerical values, while maintaining the correspondence between the feature vector and the feature value. The reordered feature value sequence is: [0.00, 0.68, 1.42, 2.85].

[0028] After sorting in ascending order, remove the eigenvector corresponding to the smallest eigenvalue that is first in the sequence. Extract the eigenvalue 0.68, which is second in the sequence, as the second smallest eigenvalue, and extract its corresponding eigenvector. As the second feature vector, the third feature value (1.42) in the sequence is extracted as the third smallest feature value, and its corresponding feature vector is extracted. As the third eigenvector. After obtaining the second and third eigenvectors, the Euclidean norm of the corresponding eigenvectors is calculated first, and each eigenvector is divided by its own Euclidean norm to normalize both the second and third eigenvectors to unit vectors; if the eigenvectors output by the eigenvalue decomposition algorithm are already unit vectors, then the vectors are kept unchanged. To avoid instability in subsequent exponential mapping results due to the non-uniqueness of the positive and negative directions of the eigenvectors, after extracting the second and third eigenvectors, sign unification processing is performed on the second and third eigenvectors respectively; specifically, the first component with an absolute value greater than a preset sign determination threshold is found in the corresponding eigenvector in component order. If the component is negative, the entire eigenvector is multiplied by -1; if the component is positive, the eigenvector is kept unchanged.

[0029] S3, the second feature vector and the third feature vector are weighted and combined according to the second and third small feature values ​​to generate a polarization angle vector. The components of the second feature vector are exponentially mapped to form a radial position vector. Based on the polarization angle vector and the radial position vector, the mapping points of all environmental parameters in the polar coordinate system are constructed. All the mapping points are converted into a ring sequence in the Cartesian coordinate system. The clustering segmentation boundary is determined based on the Euclidean distance between adjacent mapping points in the ring sequence. The environmental parameters are divided into linkage control groups through the clustering segmentation boundary. The control strategy is matched to the linkage control group and executed to realize the linkage control of the distribution box environment.

[0030] The ratio of the second smallest eigenvalue to the third smallest eigenvalue is calculated as the first combination coefficient r, and the difference between r and the first combination coefficient r is used as the second combination coefficient. The first combination coefficient r is multiplied by the third eigenvector to obtain the first vector, the second combination coefficient is multiplied by the second eigenvector to obtain the second vector, and the first vector and the second vector are added to obtain the polarization angle vector.

[0031] Extract each component from the second eigenvector after Euclidean norm normalization and sign uniformity processing, calculate the exponent of each component with the natural constant as the base, and form a radial position vector from the results of all exponent calculations in the original component order.

[0032] Using the components of the polarization angle vector corresponding to the environmental parameter as polar angles and the corresponding components of the radial position vector as polar radii, a mapping point corresponding to each environmental parameter is constructed in the polar coordinate system. All environmental parameters are mapped to the polar coordinate plane to obtain the distribution points of each parameter, and the parameters are grouped according to the spacing between these points.

[0033] The polar angles are transformed to the range of zero to twice pi. All mapped points are then sorted in ascending order according to their polar angle values. The x-coordinate is calculated by multiplying the polar radius by the cosine of the polar angle, and the y-coordinate is calculated by multiplying the polar radius by the sine of the polar angle. This transforms the sorted polar coordinate system mapped points into a continuous circular sequence in Cartesian coordinates. For polar angle components in the polarization angle vector that are less than 0, 2π is added to these components to make them fall within the 0 to 2π range; for polar angle components already within the 0 to 2π range, these components remain unchanged. Subsequent sorting, adjacent distance calculations, and circular truncation are all based on the transformed polar angles. Because the sine and cosine functions are periodic, calculating Cartesian coordinates using the transformed polar angles is equivalent to calculating coordinate positions using the original negative angles, but the sorting order should be determined by the transformed polar angles.

[0034] The Cartesian Euclidean distance between any two adjacent mapping points in the circular sequence is calculated sequentially. The closed-loop Euclidean distance is then calculated between the last and first mapping points as adjacent nodes. The mapping point pairs with the largest and second largest Euclidean distances are identified. A circular truncation is performed at the locations of these two mapping point pairs, dividing the sequence into two segments. The environmental parameter set corresponding to each segment serves as a dominant linkage control group. These two dominant linkage control groups characterize the most significant parameter correlation boundaries in the circular sequence. It is not limited to each dominant linkage control group containing only the same parameter category; when different categories of environmental parameters are contained within the same dominant linkage control group, subsequent control strategy matching steps further generate corresponding control commands according to the parameter subcategories within the group.

[0035] For each linkage control group, a preset control strategy knowledge base is invoked to identify the types of environmental parameters included. These subclasses include temperature and humidity parameters, gas safety parameters, and electrical safety parameters. A candidate set of strategies for the group is then generated based on the identification results. If multiple parameter subclasses are matched within the same linkage control group, threshold comparisons are performed on each subclass, and corresponding control command sequences are generated. When a linkage control group contains both temperature and humidity parameters, and the collected values ​​both exceed the set safety thresholds, a comprehensive temperature and humidity control command is sent to the main control module of the distribution box, triggering the cooling fan and dehumidification module. The dehumidification module performs condensation dehumidification or adsorption dehumidification, without activating the heating function. When a linkage control group contains smoke concentration and harmful gas concentration parameters, the highest-level linkage alarm command is triggered, and the power supply to non-critical circuits inside the distribution box is cut off. When a linkage control group contains voltage or leakage current parameters, and the corresponding real-time collected values ​​exceed the abnormal alarm threshold, an electrical safety control command is generated, adjusting the power supply status of the target branch and sending an abnormal status signal to the control center. For mixed parameter combinations that do not match the preset combination rules, single parameter threshold judgment is performed according to the subclass to which each parameter belongs, and the judgment result is written into the strategy candidate set within the group; if multiple candidate control commands exist at the same time, they are issued in the order of priority of electrical safety, gas safety, and temperature and humidity, and control commands with the same priority are issued in parallel.

[0036] In an optional embodiment, the step of weighting and combining the second and third eigenvectors based on the second and third minor eigenvalues ​​to generate a polarization angle vector, and then performing an exponential mapping on each component of the second eigenvector to form a radial position vector, includes: The ratio of the second smallest eigenvalue to the third smallest eigenvalue is calculated as the first combination coefficient r, and the difference between r and the first combination coefficient r is used as the second combination coefficient. Multiply the first combination coefficient r by the third feature vector to obtain a first vector, multiply the second combination coefficient by the second feature vector to obtain a second vector, and add the first vector and the second vector to obtain the polarization angle vector; Using the natural constant as the base and the i-th component of the second eigenvector after Euclidean norm normalization and sign uniformity as the exponent, the i-th radial position is obtained by exponential calculation, and the radial position vector is composed of all radial positions.

[0037] Based on the extracted second smallest eigenvalue (0.68) and third smallest eigenvalue (1.42), the ratio of the two is calculated: r = 0.4789. This ratio is used as the first combination coefficient. Subtracting 0.4789 from 1 yields 0.5211, which is used as the second combination coefficient. The first combination coefficient 0.4789 is then compared with the third eigenvector. Multiplying each component separately yields the first vector. Combine the second combination coefficient 0.5211 with the second eigenvector. Multiplying each term by the others yields the second vector. The polarization angle vector is calculated by adding the first and second vectors element-wise. .

[0038] In the radial position mapping calculation, using the natural constant e as the base, an exponential mapping calculation is performed on the first component 0.55 of the second eigenvector after Euclidean norm normalization and sign uniformity processing to obtain the first radial position. Mapping calculations are performed on the second component 0.32 to obtain the second radial position. Similarly, the third and fourth radial positions were calculated as follows: and These four calculated values ​​are concatenated sequentially to form a radial position vector. .

[0039] In an optional embodiment, the step of converting all the mapping points into a circular sequence in Cartesian coordinates, determining the clustering boundary based on the Euclidean distance between adjacent mapping points in the circular sequence, and classifying the environmental parameters into linkage control groups through the clustering boundary includes: Sort all the mapping points in ascending order according to their polar angles, and extract the polar radius and polar angle of each sorted mapping point. Multiplying the polar radius of the k-th mapping point by the cosine of the polar angle gives the x-coordinate of the k-th mapping point in the Cartesian coordinate system, and multiplying the polar radius of the k-th mapping point by the sine of the polar angle gives the y-coordinate of the k-th mapping point in the Cartesian coordinate system. Calculate the Euclidean distance between all adjacent mapping points and the Euclidean distance between the first and last mapping points respectively; Find the two mapping point pairs that have the largest and second largest Euclidean distance values, and perform a circular cut between the two mapping point pairs that have the largest and second largest Euclidean distance values ​​to divide the sorted sequence into two segments. The set of environmental parameters corresponding to each segment is used as a linkage control group, where k is an integer greater than zero and less than or equal to the number N of environmental parameters participating in the linkage group calculation.

[0040] The mapping points of the four environmental parameters in the polar coordinate system are sorted according to their polar angles. The polarization angle vectors are arranged in the parameter order of temperature, humidity, voltage, and leakage current. First, the polar angles less than 0 are converted to the range of 0 to 2π, resulting in polar angles of 0.3441 rad, 6.1387 rad, 0.0277 rad, and 6.1956 rad for temperature, humidity, voltage, and leakage current, respectively. After sorting the converted polar angles in ascending order, the parameter sequence is voltage, temperature, humidity, and leakage current, and their corresponding polar radius arrays [0.619, 1.733, 1.377, 0.549] are extracted. Taking the first mapping point after sorting, with voltage, radius 0.619, and polar angle 0.0277 rad as an example, calculate its x-coordinate x=0.619 and y-coordinate y=0.017 in the Cartesian coordinate system; similarly, calculate the (x,y) coordinates of the other three consecutive mapping points as temperature (1.631, 0.584), humidity (1.362, -0.198), and leakage current (0.547, -0.048).

[0041] Calculate the Euclidean distances between adjacent mapping points and between the first and last mapping points: Following a circular arrangement of voltage, temperature, humidity, and leakage current, calculate the distance between voltage and temperature. =1.159, the distance between temperature and humidity. =0.828, the distance between humidity and leakage current. =0.828, and the first and last distances between leakage current and voltage. =0.096. The mapping point pair with the largest Euclidean distance value is found to be voltage and temperature, with a distance of 1.159. When multiple identical second-largest Euclidean distance values ​​exist, the second-largest mapping point pair is determined according to the descending order of the starting index of the mapping point pair in the circular sorting. Thus, the mapping point pair with the second-largest Euclidean distance value is determined to be humidity and leakage current, with a distance of 0.828. A circular cutoff is performed between these two mapping point pairs with the largest and second-largest distances, splitting the cyclic sequence into two segments: one containing {temperature, humidity}, and the other containing {leakage current, voltage}. The environmental parameter set corresponding to each segment is treated as a separate linkage control group.

[0042] In an optional embodiment, the step of matching a control strategy to the linkage control group and executing it to achieve linkage control of the distribution box environment includes: For each of the divided linkage control groups, extract the set of environmental parameters contained therein; The current values ​​of temperature and humidity parameters in each linkage control group are obtained and compared with the preset safe operating threshold. If the values ​​exceed the safe operating threshold, a first control command sequence is generated to activate the cooling fan and dehumidification module of the power distribution box. The current values ​​of gas safety parameters in each linkage control group are obtained and compared with the preset gas safety threshold. If the gas safety threshold is exceeded, a gas safety alarm command sequence is generated, the alarm module is activated and the power supply of non-critical circuits inside the distribution box is cut off. The current values ​​of electrical parameters in each linkage control group are obtained and compared with the preset abnormal alarm threshold. If the abnormal alarm threshold is exceeded, a second control command sequence is generated to adjust the power supply status of the target branch of the distribution box and send an abnormal status signal to the control center. The generated control command sequence is sent to the actuator node of the distribution box and executed synchronously.

[0043] Based on the aforementioned four-parameter ring truncation results, the first linkage control group containing temperature and humidity, and the second linkage control group containing voltage and leakage current are obtained from the divided linkage control groups. If, in a complete calculation scenario containing smoke concentration and harmful gas concentration, the first linkage control group simultaneously contains temperature, humidity, smoke concentration, and harmful gas concentration, and the second linkage control group contains voltage and leakage current, then the temperature and humidity parameter set and the gas safety parameter set are first identified in the first linkage control group, and the electrical safety parameter set is identified in the second linkage control group.

[0044] For the temperature and humidity parameters in the first linkage control group, the current cabinet temperature is obtained as 42℃ and the relative humidity as 85%RH, and compared with the preset safe operating thresholds, such as a temperature limit of 35℃ and a humidity limit of 75%RH. If the comparison finds that both environmental parameters exceed the safe operating thresholds, a first control command sequence is generated, such as commands [0xAA0x010x010xFF] and [0xAA0x020x010x01], which activate the cooling fan and dehumidification module of the distribution box. For the gas safety parameters in the first linkage control group, the current values ​​of smoke concentration and harmful gas concentration are obtained and compared with preset gas safety thresholds. If any parameter exceeds the corresponding gas safety threshold, a gas safety alarm command sequence is generated, which activates the alarm module and cuts off the power supply to non-critical circuits inside the distribution box.

[0045] For the electrical safety parameter set in the second linkage control group, the current values ​​of the electrical parameters are obtained. It is detected that the voltage remains within the normal range, but the leakage current of the target branch is as high as 35mA, exceeding the 30mA abnormal alarm threshold. After confirming the over-limit state, a second control command sequence is generated, such as command 0x010x050x000x100xFF0x00 and an alarm status data packet. The circuit breaker power supply to the abnormal target branch is adjusted and cut off, and an abnormal alarm status signal with equipment identification is sent to the control center. If the actual linkage control group includes non-preset mixed combinations such as temperature and voltage, humidity and smoke concentration, single parameter threshold judgment or combined threshold judgment is performed according to the parameter subclasses. Control commands are output according to the priority order of electrical safety, gas safety, and temperature and humidity to ensure that mixed combinations can also complete strategy matching. The generated control command sequence is sent to the actuator nodes of the distribution box and executed synchronously.

[0046] In the second embodiment, the present invention also proposes a power distribution box environmental parameter linkage control system, comprising the following modules: The module is used to acquire time series data of environmental parameters in the distribution box, and determine the environmental parameters to be included in the linkage group calculation. For any two environmental parameters to be included in the calculation, the reverse metric value and the absolute value of the time series Pearson correlation coefficient are calculated, and a hybrid similarity matrix is ​​constructed accordingly. An extraction module is used to adjust the mixed similarity matrix based on preset importance coefficients of various environmental parameters to obtain an importance-adjusted weight matrix, construct a degree matrix based on the importance-adjusted weight matrix, and calculate an importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix; perform eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvector and third eigenvector; The control module is used to weight and combine the second and third feature vectors according to the second and third small feature values ​​to generate a polarization angle vector, and to perform exponential mapping on each component of the second feature vector to form a radial position vector; based on the polarization angle vector and the radial position vector, it constructs mapping points of all environmental parameters in the polar coordinate system, converts all the mapping points into a ring sequence in the Cartesian coordinate system, determines the clustering boundary based on the Euclidean distance between adjacent mapping points in the ring sequence, divides the environmental parameters into linkage control groups through the clustering boundary, matches control strategies for the linkage control groups and executes them to realize linkage control of the distribution box environment.

[0047] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for linkage control of environmental parameters in a power distribution box, characterized in that, Includes the following steps: Obtain time series data of environmental parameters inside the distribution box, and determine the environmental parameters participating in the linkage group calculation; for any two environmental parameters participating in the calculation, calculate their inverse metric value and the absolute value of the time series Pearson correlation coefficient, and construct a hybrid similarity matrix accordingly; The mixed similarity matrix is ​​adjusted based on the preset importance coefficients of each environmental parameter to obtain the importance-adjusted weight matrix. A degree matrix is ​​constructed based on the importance-adjusted weight matrix, and the importance-adjusted graph Laplacian matrix is ​​calculated from the degree matrix and the importance-adjusted weight matrix. The graph Laplacian matrix is ​​then subjected to eigenvalue decomposition to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvector and third eigenvector. The second and third feature vectors are weighted and combined according to the second and third minor feature values ​​to generate a polarization angle vector. The components of the second feature vector are exponentially mapped to form a radial position vector. Based on the polarization angle vector and the radial position vector, mapping points of all environmental parameters in the polar coordinate system are constructed. All the mapping points are converted into a ring sequence in the Cartesian coordinate system. The Euclidean distance between adjacent mapping points in the ring sequence is used to determine the clustering segmentation boundary. The environmental parameters are divided into linkage control groups through the clustering segmentation boundary. Control strategies are matched for the linkage control groups and executed to realize the linkage control of the distribution box environment.

2. The method according to claim 1, characterized in that, For any two environmental parameters involved in the calculation, their inverse metric values ​​and absolute values ​​of the time-series Pearson correlation coefficients are calculated, and a mixed similarity matrix is ​​constructed based on these values, including: For any two environmental parameters a and b, extract the normalized time series data. and ; calculate and The absolute value of the covariance is calculated separately. variance and The variances are multiplied, and the square root of the product is obtained. The absolute value of the covariance is then divided by the square root of the product to obtain the absolute value of the Pearson correlation coefficient of the time series. ; Calculate the mean values ​​of environmental parameter a and environmental parameter b within a preset time period after normalization, and obtain the results. and ; then calculate minus The absolute value of the difference is multiplied by the inverse of the preset scale attenuation constant β to obtain the product exponent. The inverse metric is then calculated using the natural constant as the base and the product exponent as the exponent. ; The preset first weighting coefficient The absolute value of the Pearson correlation coefficient with the time series Multiply to obtain the first product, then use the preset second weighting coefficients. With the reverse metric value Multiplying them yields a second product. Adding the first product to the second product gives the mixture similarity between environmental parameters a and b. The first weighting coefficient With the second weighting coefficient Both are greater than or equal to 0, and their sum is 1, which is determined by all mixed similarities. This constitutes the hybrid similarity matrix.

3. The method according to claim 1, characterized in that, The process involves adjusting the mixed similarity matrix based on preset importance coefficients for each environmental parameter to obtain an importance-adjusted weight matrix, constructing a degree matrix based on the importance-adjusted weight matrix, and calculating the importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix. Obtain the preset importance coefficient of environmental parameter a And the preset importance coefficient of environmental parameter b ; The preset importance coefficient With the preset importance coefficient and the mixed similarity before adjustment After adjusting the weights by all importance We obtain the importance-adjusted edge weight matrix W; For the i-th row of the importance-adjusted weight matrix W, calculate the sum of all importance-adjusted weights in the i-th row to obtain the degree of the i-th environmental parameter node. Construct a diagonal matrix of the degrees of all environmental parameter nodes as the degree matrix D, where the number of environmental parameters N participating in the linkage group calculation is an integer greater than or equal to 3, and i is an integer greater than zero and less than or equal to the number of environmental parameters N participating in the linkage group calculation. Subtracting the importance-adjusted edge weight matrix W from the degree matrix D yields the importance-adjusted graph Laplacian matrix L.

4. The method according to claim 2, characterized in that, The step of performing eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second and third eigenvectors includes: Solving the characteristic equation of the graph Laplacian matrix L yields N eigenvalues ​​and their corresponding eigenvectors, where N is an integer greater than or equal to 3; The N feature values ​​are arranged in ascending order from smallest to largest, while maintaining the correspondence between feature vectors and feature values; Remove the eigenvector corresponding to the smallest eigenvalue, extract the second-ranked eigenvalue as the second smallest eigenvalue, and extract the eigenvector corresponding to the second smallest eigenvalue as the second eigenvector. The third-ranked feature value is extracted as the third smallest feature value, and the feature vector corresponding to the third smallest feature value is extracted as the third feature vector.

5. The method according to claim 1, characterized in that, The step of weighting and combining the second and third eigenvectors based on the second and third smallest eigenvalues ​​to generate a polarization angle vector, and then performing an exponential mapping on each component of the second eigenvector to form a radial position vector, includes: The ratio of the second smallest eigenvalue to the third smallest eigenvalue is calculated as the first combination coefficient r, and the difference between r and the first combination coefficient r is used as the second combination coefficient. Multiply the first combination coefficient r by the third feature vector to obtain a first vector, multiply the second combination coefficient by the second feature vector to obtain a second vector, and add the first vector and the second vector to obtain the polarization angle vector; Using the natural constant as the base and the i-th component of the second eigenvector after Euclidean norm normalization and sign uniformity as the exponent, the i-th radial position is obtained by exponential calculation, and the radial position vector is composed of all radial positions.

6. The method according to claim 1, characterized in that, The process involves converting all the mapping points into a circular sequence in Cartesian coordinates, determining the clustering boundary based on the Euclidean distance between adjacent mapping points in the circular sequence, and classifying environmental parameters into coordinated control groups using the clustering boundary. Sort all the mapping points in ascending order according to their polar angles, and extract the polar radius and polar angle of each sorted mapping point. Multiplying the polar radius of the k-th mapping point by the cosine of the polar angle gives the x-coordinate of the k-th mapping point in the Cartesian coordinate system, and multiplying the polar radius of the k-th mapping point by the sine of the polar angle gives the y-coordinate of the k-th mapping point in the Cartesian coordinate system. Calculate the Euclidean distance between all adjacent mapping points and the Euclidean distance between the first and last mapping points respectively; Find the two mapping point pairs that have the largest and second largest Euclidean distance values, and perform a circular cut between the two mapping point pairs that have the largest and second largest Euclidean distance values ​​to divide the sorted sequence into two segments. The set of environmental parameters corresponding to each segment is used as a linkage control group, where k is an integer greater than zero and less than or equal to the number N of environmental parameters participating in the linkage group calculation.

7. The method according to claim 1, characterized in that, The process of matching and executing control strategies for the linkage control group to achieve linkage control of the distribution box environment includes: For each of the divided linkage control groups, extract the set of environmental parameters contained therein; The current values ​​of temperature and humidity parameters in each linkage control group are obtained and compared with the preset safe operating threshold. If the value exceeds the safe operating threshold, a first control command sequence is generated to activate the cooling fan and dehumidification module of the power distribution box. The current values ​​of gas safety parameters in each linkage control group are obtained and compared with the preset gas safety threshold. If the values ​​exceed the gas safety threshold, a gas safety alarm command sequence is generated, the alarm module is activated, and the power supply to non-critical circuits inside the distribution box is cut off. The current values ​​of electrical parameters in each linkage control group are obtained and compared with the preset abnormal alarm threshold. If the abnormal alarm threshold is exceeded, a second control command sequence is generated to adjust the power supply status of the target branch of the distribution box and send an abnormal status signal to the control center. The generated control command sequence is sent to the actuator node of the distribution box and executed synchronously.

8. A power distribution box environmental parameter linkage control system, characterized in that, Includes the following modules: The module is used to acquire time series data of environmental parameters in the distribution box, and determine the environmental parameters to be included in the linkage group calculation. For any two environmental parameters to be included in the calculation, the reverse metric value and the absolute value of the time series Pearson correlation coefficient are calculated, and a hybrid similarity matrix is ​​constructed accordingly. An extraction module is used to adjust the mixed similarity matrix based on preset importance coefficients of various environmental parameters to obtain an importance-adjusted weight matrix, construct a degree matrix based on the importance-adjusted weight matrix, and calculate an importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix; perform eigenvalue decomposition on the graph Laplacian matrix to extract the second smallest eigenvalue, the third smallest eigenvalue, and their corresponding second eigenvector and third eigenvector; The control module is used to weight and combine the second and third feature vectors according to the second and third small feature values ​​to generate a polarization angle vector, and to perform exponential mapping on each component of the second feature vector to form a radial position vector; based on the polarization angle vector and the radial position vector, it constructs mapping points of all environmental parameters in the polar coordinate system, converts all the mapping points into a ring sequence in the Cartesian coordinate system, determines the clustering boundary based on the Euclidean distance between adjacent mapping points in the ring sequence, divides the environmental parameters into linkage control groups through the clustering boundary, matches control strategies for the linkage control groups and executes them to realize linkage control of the distribution box environment.

9. The system according to claim 8, characterized in that, For any two environmental parameters involved in the calculation, their inverse metric values ​​and absolute values ​​of the time-series Pearson correlation coefficients are calculated, and a mixed similarity matrix is ​​constructed based on these values, including: For any two environmental parameters a and b, extract the normalized time series data. and ; calculate and The absolute value of the covariance is calculated separately. variance and The variances are multiplied, and the square root of the product is obtained. The absolute value of the covariance is then divided by the square root of the product to obtain the absolute value of the Pearson correlation coefficient of the time series. ; Calculate the mean values ​​of environmental parameter a and environmental parameter b within a preset time period after normalization, and obtain the results. and ; then calculate minus The absolute value of the difference is multiplied by the inverse of the preset scale attenuation constant β to obtain the product exponent. The inverse metric is then calculated using the natural constant as the base and the product exponent as the exponent. ; The preset first weighting coefficient The absolute value of the Pearson correlation coefficient with the time series Multiply to obtain the first product, then use the preset second weighting coefficients. With the reverse metric value Multiplying them yields a second product. Adding the first product to the second product gives the mixture similarity between environmental parameters a and b. The first weighting coefficient With the second weighting coefficient Both are greater than or equal to 0, and their sum is 1, which is determined by all mixed similarities. This constitutes the hybrid similarity matrix.

10. The system according to claim 8, characterized in that, The process involves adjusting the mixed similarity matrix based on preset importance coefficients for each environmental parameter to obtain an importance-adjusted weight matrix, constructing a degree matrix based on the importance-adjusted weight matrix, and calculating the importance-adjusted graph Laplacian matrix using the degree matrix and the importance-adjusted weight matrix. Obtain the preset importance coefficient of environmental parameter a And the preset importance coefficient of environmental parameter b ; The preset importance coefficient With the preset importance coefficient and the mixed similarity before adjustment After adjusting the weights by all importance We obtain the importance-adjusted edge weight matrix W; For the i-th row of the importance-adjusted weight matrix W, calculate the sum of all importance-adjusted weights in the i-th row to obtain the degree of the i-th environmental parameter node. Construct a diagonal matrix of the degrees of all environmental parameter nodes as the degree matrix D, where the number of environmental parameters N participating in the linkage group calculation is an integer greater than or equal to 3, and i is an integer greater than zero and less than or equal to the number of environmental parameters N participating in the linkage group calculation. Subtracting the importance-adjusted edge weight matrix W from the degree matrix D yields the importance-adjusted graph Laplacian matrix L.

Images