Network structure adjustment method and device based on environmental data, equipment and medium

Abstract

The invention discloses a network structure adjustment method and device based on environmental data, equipment and a medium. The method comprises the following steps: collecting equipment environment data corresponding to node equipment in a target power system under a preliminary network structure in real time; performing feature extraction on the equipment environment data to obtain a first environment feature, and performing feature dimension reduction on the first environment feature to obtain a second environment feature; network prediction is carried out on the input second environment feature through a network prediction model, a network prediction result is obtained, and the network prediction result comprises prediction network load and / or network stability parameters; and adjusting the preliminary network structure in real time according to a network prediction result, and determining a target network structure. According to the invention, the network structure can be adaptively adjusted in real time according to the equipment environment data collected in real time, the high performance and high stability of the system network can be continuously maintained, and the continuous and efficient operation of the power system can be ensured.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a network structure adjustment method, device, equipment and medium based on environmental data. Background Art

[0002] The network performance and stability of the power system are usually affected by internal and external factors, resulting in fluctuations in network performance and stability, and causing abnormal operation of the power system.

[0003] In related technologies, relevant technical personnel usually regularly analyze the operating parameters of the power system and make regular network structure adjustments based on the analysis results to maintain the high performance and high stability of the network and ensure the efficient operation of the power system. However, there is a problem that the maintenance effect of the high performance and high stability of the network is poor, and the power system is difficult to operate efficiently and sustainably. Summary of the Invention

[0004] The present invention provides a network structure adjustment method, device, equipment and medium based on environmental data to solve the problems of poor maintenance of high performance and high stability of the network and difficulty in sustainable and efficient operation of the power system in existing network structure adjustment methods.

[0005] According to one aspect of the present invention, a method for adjusting a network structure based on environmental data is provided, the method comprising:

[0006] Real-time collection of device environment data corresponding to node devices in the target power system under the preliminary network structure, wherein the device environment data includes environment-related data and / or device-related data;

[0007] Performing feature extraction on the device environment data to obtain a first environment feature, and performing feature dimensionality reduction on the first environment feature to obtain a second environment feature;

[0008] Performing network prediction on the input second environmental feature using a network prediction model to obtain a network prediction result, wherein the network prediction result includes a predicted network load and / or a network stability parameter, the data prediction model being obtained by training a long short-term memory network based on a training sample set;

[0009] The preliminary network structure is adjusted in real time according to the network prediction result to determine a target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth.

[0010] According to another aspect of the present invention, there is provided a network structure adjustment device based on environmental data, the device comprising:

[0011] A data acquisition module, configured to acquire in real time device environment data corresponding to node devices in the target power system under the preliminary network structure, wherein the device environment data includes environment-related data and / or device-related data;

[0012] a feature processing module, configured to extract features from the device environment data to obtain a first environment feature, and perform feature dimensionality reduction on the first environment feature to obtain a second environment feature;

[0013] a network prediction module, configured to perform network prediction on the input second environmental feature using a network prediction model to obtain a network prediction result, wherein the network prediction result includes a predicted network load and / or a network stability parameter, wherein the data prediction model is obtained by training a long short-term memory network based on a training sample set;

[0014] The network adjustment module is used to adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth.

[0015] According to another aspect of the present invention, an electronic device is provided, comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the network structure adjustment method based on environmental data described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the network structure adjustment method based on environmental data according to any embodiment of the present invention when executed.

[0020] The technical solution of the embodiment of the present invention is to collect device environment data corresponding to node devices in the target power system under the preliminary network structure in real time, wherein the device environment data includes environment-related data and / or device-related data; perform feature extraction on the device environment data to obtain a first environment feature, perform feature dimensionality reduction on the first environment feature to obtain a second environment feature; perform network prediction on the input second environment feature through a network prediction model to obtain a network prediction result, wherein the network prediction result includes predicted network load and / or network stability parameters, and the data prediction model is obtained by training a long short-term memory network based on a training sample set; and adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth. The power system network structure can be adjusted in real time and dynamically based on the device environment data, avoiding the limitations of traditional static network structure adjustment methods. By adaptively adjusting the network structure in real time according to the real-time collected device environment data, the high performance and high stability of the system network can be continuously maintained, ensuring the continuous and efficient operation of the power system.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 This is a flow chart of a network structure adjustment method based on environmental data provided in accordance with the first embodiment of the present invention;

[0024] Figure 2 This is a flow chart of a network structure adjustment method based on environmental data provided in accordance with a second embodiment of the present invention;

[0025] Figure 3 This is a flow chart of a network structure adjustment method based on environmental data provided in accordance with a third embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of the structure of a network structure adjustment device based on environmental data provided in accordance with a fourth embodiment of the present invention;

[0027] Figure 5The present invention is a schematic structural diagram of an electronic device for implementing the network structure adjustment method based on environmental data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] Example 1

[0031] Figure 1 A flowchart of a network structure adjustment method based on environmental data is provided for the first embodiment of the present invention. This embodiment is applicable to situations where the network structure of a power system is adjusted to improve the overall network performance of the power system. The method can be executed by a network structure adjustment device based on environmental data. The network structure adjustment device based on environmental data can be implemented in the form of hardware and / or software. The network structure adjustment device based on environmental data can be configured in a computer. Figure 1 As shown, the method includes:

[0032] S110 , collecting device environment data corresponding to node devices in the target power system under the preliminary network structure in real time.

[0033] The preliminary network structure may be understood as a network structure to be implemented, and may include a preliminary network path and / or a preliminary network bandwidth.

[0034] The target power system can be understood as a power system to be subjected to real-time network structure adjustment. In the implementation of the present invention, the target power system can be an embedded power system.

[0035] The node device may be a data transmission device in the target power system, etc.

[0036] The device environment data can be understood as device data and environment data related to the node device. Optionally, the device environment data includes environment-related data and / or device-related data. The environment-related data can be understood as environment data related to the node device. Optionally, the environment-related data includes temperature data and / or humidity data. The device-related data includes device data related to the power-saving device. Optionally, the device-related data includes at least one of device load data, device current data, and device voltage data.

[0037] S120 , performing feature extraction on the device environment data to obtain a first environment feature, and performing feature dimensionality reduction on the first environment feature to obtain a second environment feature.

[0038] The first environmental feature may be understood as a feature related to the device environment data.

[0039] The second environmental feature can be understood as a low-latitude feature related to the first environmental feature.

[0040] Optionally, extracting features from the device environment data to obtain a first environment feature includes:

[0041] Performing feature extraction on the device environment data to determine a plurality of device environment features, and determining, for each of the device environment features, a feature change rate corresponding to the device environment feature;

[0042] A target feature vector is constructed according to the device environment feature and the feature change rate, and the target feature vector is used as the first environment feature of the device environment data.

[0043] In the embodiment of the present invention, the first environmental feature may be a feature vector.

[0044] The device environment characteristics may include environment-related characteristics and / or device-related characteristics. Optionally, the environment-related characteristics may include temperature characteristics and / or humidity characteristics. The device-related characteristics may include at least one of device load characteristics, device current characteristics, and device voltage characteristics.

[0045] The characteristic change rate can be understood as the change rate of the device environment characteristic. In the embodiment of the present invention, the calculation method of the characteristic change rate is not specifically limited here.

[0046] The target feature vector may correspond to the first environmental feature.

[0047] Specifically, the following is an exemplary explanation of extracting features from the device environment data to obtain the first environment feature:

[0048] 1. Perform feature extraction on the device environment data to determine multiple device environment features.

[0049] 2. Calculate the characteristic change rate of each device environment characteristic to obtain a first environment characteristic. In an embodiment of the present invention, the node device may include multiple related devices. Exemplarily, the first environment characteristic may be as follows:

[0050] x i =[T i ,H i ,L i ,I i ,V i ,ΔT i ,ΔH i ,ΔL i ,ΔI i ,ΔV i ] T ;

[0051] Among them, x i represents the target feature vector of the i-th related device, T i represents the temperature characteristics of the i-th related equipment; H i represents the humidity characteristics of the i-th related equipment; L i I represents the equipment load characteristics of the i-th related equipment; i Represents the current characteristics of the i-th related device; V i represents the voltage characteristic of the i-th related device; ΔT i Indicates the temperature change rate of the i-th related device; ΔH i Indicates the humidity change rate of the i-th related equipment; ΔL i Indicates the equipment load change rate of the i-th related equipment; ΔI i Indicates the current change rate of the i-th related device; ΔV i Represents the voltage change rate of the i-th related device.

[0052] Based on the above embodiment, the expanded feature vector can be obtained by calculating the change rate of each feature, which can more comprehensively reflect the impact of environmental changes on the power system.

[0053] Optionally, performing feature dimensionality reduction on the first environmental feature to obtain the second environmental feature includes:

[0054] Performing dimensionality reduction on the first environmental feature through a deep autoencoder to determine preliminary low-dimensional features;

[0055] Determining a feature weight according to the preliminary low-latitude feature and the first environmental feature, and determining a weighted low-latitude feature according to the preliminary low-latitude feature and the feature weight;

[0056] Performing nonlinear transformation on the weighted low-dimensional features through principal component analysis to obtain a low-dimensional matrix;

[0057] A second environmental feature corresponding to the first environmental feature is determined according to the weighted low-dimensional feature and the low-dimensional matrix.

[0058] Specifically, the following is an exemplary explanation of performing feature dimensionality reduction on the first environmental feature to obtain the second environmental feature:

[0059] 1. Build a deep autoencoder. The deep encoder can map high-dimensional input to a low-dimensional latent space; the decoder can then reconstruct the low-dimensional latent representation back to the high-dimensional input.

[0060] In an embodiment of the present invention, the deep autoencoder may have the function of feature dimensionality reduction. Optionally, the deep autoencoder may be trained using the reconstruction error as the loss function, and the implementation formula is as follows:

[0061]

[0062] Among them, f decoder represents the encoder function, f encoder represents the decoder function, θ e represents the encoder parameters, θ d represents the decoder parameters, and n represents the number of samples for training the deep autoencoder.

[0063] 2. Adaptive feature weight determination. Based on the trained autoencoder, the influence of the input features on the latent space representation is calculated through sensitivity analysis or gradient backpropagation method. The formula is as follows:

[0064]

[0065] Where h represents the preliminary low-latitude feature; Represents h for input feature x i The partial derivative of x i Indicates the first environmental feature.

[0066] The feature weight is determined based on the calculated influence, and the weighted low-dimensional features corresponding to the preliminary low-dimensional features are determined according to the feature weight, and the realization is publicized as follows:

[0067] h w =w⊙h;

[0068] Among them, h wRepresents weighted low-dimensional features, w represents feature weights, and ⊙ represents element-level multiplication operations.

[0069] 3. Nonlinear principal component analysis (PCA) transformation: The PCA algorithm is used to further reduce the dimensionality of weighted low-dimensional features.

[0070] 1) Construct a weighted covariance matrix and implement the formula as follows:

[0071]

[0072] Among them, μh w represents the mean of weighted low-dimensional features, represents the weighted covariance matrix.

[0073] 2) Eigenvalue decomposition. Perform eigenvalue decomposition on the weighted covariance matrix, select the eigenvectors corresponding to the first d largest eigenvalues in the decomposition result, and determine the low-dimensional matrix based on the selected eigenvectors. The reference relationship is as follows:

[0074]

[0075] Among them, k represents the dimension of the weighted covariance matrix, d represents the dimension of the low-dimensional matrix, and W represents the low-dimensional matrix.

[0076] 3) Low-dimensional space projection. Determine the second environmental feature corresponding to the first environmental feature based on the weighted low-dimensional feature and the low-dimensional matrix, using the following formula:

[0077] Z=h w W;

[0078] Wherein, Z represents the second environmental feature.

[0079] In further detail, the calculation of the low-dimensional matrix W is specifically explained:

[0080] 1. Calculate the weighted eigenvalue and weighted eigenvector of the weighted low-dimensional features. The implementation formula is as follows:

[0081]

[0082] Among them, v represents the weighted eigenvector with a dimension of k×1; λ represents the weighted eigenvalue, represents the weighted covariance matrix.

[0083] The steps used to solve the above formula are as follows:

[0084] The characteristic equation is constructed and the implementation formula is as follows:

[0085]

[0086] Where det represents the determinant operator and I represents the identity matrix with dimension k×k.

[0087] By solving the characteristic equation, k values of the eigenvalue λ are obtained;

[0088] Weighted eigenvector calculation. For each weighted eigenvalue λ, the corresponding eigenvector v is calculated through a set of linear equations. The implementation formula is as follows:

[0089]

[0090] in, represents the weighted covariance matrix, I represents the identity matrix, λ represents the weighted eigenvalue, and v represents the weighted eigenvector.

[0091] 2. From the calculated eigenvalues and eigenvectors, select the first d largest eigenvalues λ1,λ2,…,λ d and its corresponding eigenvectors v1,v2,…,v d ; Combine the selected d eigenvectors into a matrix W = [v1, v2, ..., v d ], and get the low-dimensional matrix W.

[0092] Based on the above embodiment scheme, the eigenvalues and eigenvectors of the weighted covariance matrix can be effectively calculated through eigenvalue decomposition. Selecting the eigenvectors corresponding to the first d largest eigenvalues to determine the low-dimensional matrix W can effectively reduce the feature dimension and improve the efficiency of model training and prediction. In an embodiment of the present invention, the processing method for training samples can be the same as the processing method for device environment data, based on this, it can help improve the generalization ability and robustness of the model.

[0093] In an embodiment of the present invention, a deep autoencoder can be used to map high-dimensional input to a low-dimensional latent space representation, including multiple fully connected layers, each of which is connected to a ReLU activation function.

[0094] For the decoder corresponding to the deep autoencoder, it can be used to reconstruct the low-dimensional potential representation back to the high-dimensional input. Its structure is opposite to that of the encoder. The number of fully connected layers used is the same as the number of encoder layers, and each layer is followed by a ReLU activation function.

[0095] Based on the above-described embodiment scheme, by constructing a deep autoencoder to reduce the dimensionality of environmental features, feature redundancy is reduced and the efficiency of model prediction is improved. Using the reconstruction error as the loss function to train the autoencoder can effectively optimize encoder performance. By calculating the influence degree through sensitivity analysis or gradient backpropagation methods, features that have a significant impact on network performance can be more accurately identified. Nonlinear PCA transformation further reduces the dimensionality and improves the model's prediction and generalization capabilities.

[0096] S130: Perform network prediction on the input second environmental feature through a network prediction model to obtain a network prediction result.

[0097] The network prediction model can be understood as a machine model with data prediction capabilities. In an embodiment of the present invention, for the network prediction model, the input data can be a second environmental feature related to the device environment data. The output data can be the network prediction result. The network prediction result may include predicted network load and / or network stability parameters. The predicted network load can be a network load condition. For example, the predicted network load can be a predicted load value. The network stability parameter can characterize the degree of network stability. For example, the network stability parameter can be a predicted stability value.

[0098] In an embodiment of the present invention, the data prediction model is obtained by training a Long Short Term Memory (LSTM) network based on a training sample set.

[0099] The following is a detailed description of the LSTM model structure in the embodiment of the present invention:

[0100] 1. Input layer, which can be used to input the second environment features.

[0101] The first LSTM layer includes 128 LSTM units, the activation function is tanh, the dropout rate is 0.2 to reduce overfitting, and the output of each time step is 128 dimensions.

[0102] The second LSTM layer receives the output of the first LSTM layer, consists of 64 LSTM units, uses the tanh activation function, and has a dropout rate of 0.2 to reduce overfitting. The output of each time step is 64 dimensions.

[0103] The third LSTM layer receives the output of the second LSTM layer, consists of 32 LSTM units, uses the tanh activation function, and has a dropout rate of 0.2 to reduce overfitting. The output of each time step is 32 dimensions.

[0104] The first fully connected layer receives the output of the third LSTM layer, consists of 64 units, uses ReLU as the activation function, and has a dropout rate of 0.5; each time step outputs a 64-dimensional image.

[0105] The second fully connected layer receives the output of the first fully connected layer, consists of 32 units, uses ReLU as the activation function, and has a dropout rate of 0.5; each time step outputs 32 dimensions.

[0106] 2. Network load prediction branch. Receives the output of the second fully connected layer, which uses one unit and a linear activation function, and outputs the network load prediction value for each network path.

[0107] 3. Network path stability prediction branch: This branch receives the output of the second fully connected layer, which is selected based on the number of paths and uses a Sigmoid activation function. It outputs a stability prediction value for each network path, ranging from 0 to 1.

[0108] Based on the above implementation scheme, the LSTM model has powerful time series prediction capabilities. Through multiple LSTM layers and fully connected layers, it can effectively capture the impact of environmental changes on network load and network path stability. The application of the ReLU activation function and the Dropout layer can enhance the model's nonlinear expression capabilities and reduce overfitting. The design of the network load prediction branch and the path network stability prediction branch allows for branching network load and path network stability prediction, effectively improving the model's prediction accuracy and practicality.

[0109] S140: Adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure.

[0110] In this embodiment of the present invention, the preliminary network structure may include a preliminary network path and / or a preliminary network bandwidth. Correspondingly, the target network structure may include a target network path and / or a target network bandwidth. The preliminary network path and the target network path may be different or the same. The preliminary network bandwidth and the target network bandwidth may be different or the same.

[0111] The technical solution of the embodiment of the present invention is to collect device environment data corresponding to node devices in the target power system under the preliminary network structure in real time, wherein the device environment data includes environment-related data and / or device-related data; perform feature extraction on the device environment data to obtain a first environment feature, perform feature dimensionality reduction on the first environment feature to obtain a second environment feature; perform network prediction on the input second environment feature through a network prediction model to obtain a network prediction result, wherein the network prediction result includes predicted network load and / or network stability parameters, and the data prediction model is obtained by training a long short-term memory network based on a training sample set; and adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth. The power system network structure can be adjusted in real time and dynamically based on the device environment data, avoiding the limitations of traditional static network structure adjustment methods. By adaptively adjusting the network structure in real time according to the real-time collected device environment data, the high performance and high stability of the system network can be continuously maintained, ensuring the continuous and efficient operation of the power system.

[0112] Example 2

[0113] Figure 2 This is a flowchart of a network structure adjustment method based on environmental data provided by the second embodiment of the present invention. This embodiment is directed to the above embodiment, which adjusts the preliminary network structure in real time according to the network prediction results and determines the target network structure for refinement. Figure 2 As shown, the method includes:

[0114] S210 , collecting device environment data corresponding to node devices in the target power system under the preliminary network structure in real time.

[0115] S220 , performing feature extraction on the device environment data to obtain a first environment feature, and performing feature dimensionality reduction on the first environment feature to obtain a second environment feature.

[0116] S230: Perform network prediction on the input second environmental feature through a network prediction model to obtain a network prediction result.

[0117] S240: Determine multiple candidate network paths in the target power system, and determine a broadband demand coefficient and a path broadband limit corresponding to the target power system.

[0118] The candidate network path may be understood as a candidate network transmission path.

[0119] The broadband demand coefficient may be understood as a broadband calculation coefficient related to the target power system.

[0120] The path bandwidth limit may be understood as a bandwidth limit associated with a network path in the target power system. Optionally, the path bandwidth limit includes a maximum bandwidth limit and / or a minimum bandwidth limit.

[0121] In the embodiment of the present invention, the bandwidth requirement coefficient and the path bandwidth limit may be preset according to scenario requirements, and are not specifically limited herein.

[0122] S250. For each candidate network path, determine the target network path in the candidate network path and the target network bandwidth corresponding to the target network path according to the bandwidth demand coefficient, the path bandwidth limit, and the network prediction result.

[0123] Optionally, the network prediction result includes a sub-prediction result corresponding to each candidate network path, and the sub-prediction result includes a sub-prediction load related to the predicted network load and a sub-prediction stability parameter related to the network stability parameter;

[0124] The determining, according to the broadband demand coefficient, the path bandwidth limit, and the network prediction result, the target network path in the candidate network paths and the target network bandwidth corresponding to the target network path includes:

[0125] determining the target network path among the candidate network paths according to the sub-prediction stability parameter;

[0126] Determine a required broadband parameter according to the sub-predicted load corresponding to the target network path and the broadband demand coefficient, and determine a broadband stability parameter according to the required broadband parameter and the sub-predicted stability parameter corresponding to the target network path;

[0127] A target broadband parameter corresponding to the target network path is determined according to the path broadband limit and the broadband stability parameter, and the target network bandwidth is determined according to the target broadband parameter.

[0128] The sub-prediction stability parameter may represent the stability of each candidate network path.

[0129] Optionally, the target network path may be a network path with the highest stability among the candidate network paths.

[0130] The required bandwidth parameter may represent the bandwidth required by the target network path under the sub-predicted load.

[0131] The broadband stability parameter may represent the bandwidth required by the network path under the sub-predicted load while ensuring the stability of the network path.

[0132] The target bandwidth parameter may include the path bandwidth limit or the bandwidth stability parameter.

[0133] Specifically, the following is an exemplary explanation of determining the stable broadband parameter corresponding to each candidate network path according to the broadband demand coefficient, the path broadband limit, and the network prediction result:

[0134] 1. Determine the candidate network path with the highest network stability represented by the sub-prediction stability parameter as the target network path.

[0135] 2. Calculate the bandwidth requirements of the network path. The implementation formula is as follows:

[0136] B requiredpath,i =L path,i ×B factor ;

[0137] Among them, B requiredpath,i represents the bandwidth requirement parameter of the i-th path, L path,i represents the sub-predicted load of the i-th path, B factor represents the broadband demand coefficient.

[0138] The i-th path may be a target network path.

[0139] 3. Based on the predicted network stability parameters, adjust the bandwidth requirements. The implementation formula is as follows:

[0140] B adjustedpath,i =B requiredpath,i ×S path,i ;

[0141] Among them, B adjustedpath,i represents the bandwidth requirement of the ith path after considering stability (i.e., the broadband stability parameter); S path,i represents the sub-prediction stability parameter of the i-th path.

[0142] 4. Limit the adjusted bandwidth demand according to the bandwidth limit. The implementation formula is as follows:

[0143] B finalpath,i =max(B min ,min(B max ,B adjustedpath,i ));

[0144] Among them, B finalpath,i represents the target bandwidth of the i-th path (i.e., the target bandwidth parameter), B min Indicates the minimum bandwidth limit of the path; B max Indicates the maximum bandwidth limit of the path.

[0145] Furthermore, the target broadband parameter is used as the target network bandwidth.

[0146] S260: Adjust the preliminary network structure in real time according to the target network path and the target network bandwidth to determine a target network structure.

[0147] Based on the above embodiment scheme, it is realized that according to the sub-prediction stability parameters related to the network stability parameters, the most stable network path is selected as the target network path for data transmission in the power system, which can effectively reduce the probability of data loss and delay due to network fluctuations, ensure the efficient operation of the power system, and improve the overall reliability of the network structure; and the bandwidth demand under the stability of the network path is calculated in combination with the predicted sub-prediction stability parameters, ensuring the reasonable allocation of bandwidth of the target network path, improving the utilization rate of bandwidth, avoiding the waste of bandwidth resources, and improving the overall performance of the power system.

[0148] The technical solution of the embodiment of the present invention determines the broadband demand coefficient and path broadband limit corresponding to the target power system by determining multiple candidate network paths in the target power system; for each candidate network path, the target network path in the candidate network path and the target network bandwidth corresponding to the target network path are determined based on the broadband demand coefficient, the path broadband limit, and the network prediction result; the preliminary network structure is adjusted in real time based on the target network path and the target network bandwidth to determine the target network structure. The determination and adjustment of the target network structure based on multi-dimensional data (network stability-related prediction value, load prediction value, broadband limit, and broadband demand-related coefficient) is realized.

[0149] Example 3

[0150] Figure 3 This is a flowchart of a network structure adjustment method based on environmental data provided in the third embodiment of the present invention. This embodiment is an addition to the above embodiment for determining the target network structure. Figure 3 As shown, the method includes:

[0151] S310 , collecting device environment data corresponding to node devices in the target power system under the preliminary network structure in real time.

[0152] S320: Perform feature extraction on the device environment data to obtain a first environment feature, and perform feature dimensionality reduction on the first environment feature to obtain a second environment feature.

[0153] S330: Perform network prediction on the input second environmental feature through a network prediction model to obtain a network prediction result.

[0154] S340: Adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure.

[0155] S350, determining target performance parameters of the node devices in the target power system under the target network structure; and determining preliminary performance parameters of the node devices in the target power system under the preliminary network structure.

[0156] The target performance parameter may represent the performance of the node device in the target power system under the target network structure. Correspondingly, the target performance parameter may represent the performance of the node device in the target power system under the preliminary network structure.

[0157] Optionally, determining target performance parameters of the node devices in the target power system under the target network structure includes:

[0158] Determining performance-related data corresponding to the node device in the target power system under the target network structure, wherein the performance-related data includes at least one of transmission delay, packet loss rate, throughput, and bandwidth utilization;

[0159] The target performance parameter is determined based on the performance-related data.

[0160] The performance-related data may be understood as performance parameters related to the node device. Optionally, the performance-related data may include device performance parameters and / or network performance parameters.

[0161] Exemplarily, the performance-related data includes at least one of transmission delay, packet loss rate, throughput, and bandwidth utilization.

[0162] S360: Determine a performance improvement parameter based on the preliminary performance parameter and the target performance parameter, and if the performance improvement parameter does not meet a preset performance condition, determine to update the prediction model.

[0163] The performance improvement parameter may represent the degree of performance improvement of the target power system after the network structure is adjusted.

[0164] The updated prediction model is obtained by training the long short-term memory network based on the updated sample set. In an embodiment of the present invention, the updated sample set and the training sample set may be different.

[0165] Specifically, the following is an exemplary description of the specific steps for determining the performance improvement parameters:

[0166] 1. Determine the target performance parameters based on performance-related data. The implementation formula is as follows:

[0167]

[0168] Among them, α, β, γ, and δ represent multiple weight coefficients respectively, Th represents throughput; PLR represents packet loss rate; Ly represents latency; BU represents bandwidth utilization; and S represents the target performance parameter.

[0169] 2. Determine the channel improvement percentage for each channel related to the network structure. The implementation formula is as follows:

[0170]

[0171] Among them, IP KPI Indicates the channel improvement percentage of performance-related data KPI, S old represents the initial performance parameters before network structure adjustment, S new Represents the target performance parameter after network structure adjustment.

[0172] 3. Based on IP KPI Calculate the performance improvement parameter to measure the degree of network performance improvement after network structure adjustment. The implementation formula is as follows:

[0173]

[0174] Among them, w i represents the weight of the KPI of the i-th channel, and OS represents the performance improvement parameter.

[0175] 4. Determine whether the performance meets the preset performance conditions based on the performance improvement parameters. If not, collect and update the sample set again, retrain the LSTM model with the updated sample set, adjust the model parameters, obtain an updated prediction model, and adjust the network structure with the updated prediction model.

[0176] Based on the above embodiment scheme, by setting key performance indicators such as transmission delay, packet loss rate, throughput and bandwidth utilization, it is possible to comprehensively evaluate the performance improvement of the network, and to update the network structure adjustment based on the evaluation results, effectively ensuring the data transmission quality of the power system and ensuring the timeliness and accuracy of data transmission.

[0177] S370: Perform network prediction on the second environmental feature using the updated prediction model to obtain an updated prediction result, and determine an updated network structure according to the updated prediction result.

[0178] The technical solution of the embodiment of the present invention is to determine the target performance parameters of the node devices in the target power system under the target network structure; and to determine the preliminary performance parameters of the node devices in the target power system under the preliminary network structure; determine the performance improvement parameters based on the preliminary performance parameters and the target performance parameters, and determine an updated prediction model when the performance improvement parameters do not meet the preset performance conditions, wherein the updated prediction model is obtained by training the long short-term memory network based on the updated sample set; perform network prediction on the second environmental feature through the updated prediction model to obtain an updated prediction result, and determine an updated network structure based on the updated prediction result. The present invention can verify the degree of network performance improvement of the network structure adjustment, and when the degree of network performance improvement is low, perform network structure adjustment based on the backup prediction model to ensure a high degree of network performance improvement under real-time network structure adjustment.

[0179] Example 4

[0180] Figure 4 This is a schematic diagram of a network structure adjustment device based on environmental data provided by the fourth embodiment of the present invention. Figure 4 As shown, the device includes: a data collection module 410, a feature processing module 420, a network prediction module 430 and a network adjustment module 440.

[0181] Among them, the data acquisition module 410 is used to collect in real time the equipment environment data corresponding to the node equipment in the target power system under the preliminary network structure, wherein the equipment environment data includes environment-related data and / or equipment-related data; the feature processing module 420 is used to extract features from the equipment environment data to obtain a first environment feature, and perform feature dimensionality reduction on the first environment feature to obtain a second environment feature; the network prediction module 430 is used to perform network prediction on the input second environment feature through a network prediction model to obtain a network prediction result, wherein the network prediction result includes predicted network load and / or network stability parameters, and the data prediction model is obtained by training the long short-term memory network based on the training sample set; the network adjustment module 440 is used to adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth.

[0182] The technical solution of the embodiment of the present invention is to collect device environment data corresponding to node devices in the target power system under the preliminary network structure in real time, wherein the device environment data includes environment-related data and / or device-related data; perform feature extraction on the device environment data to obtain a first environment feature, perform feature dimensionality reduction on the first environment feature to obtain a second environment feature; perform network prediction on the input second environment feature through a network prediction model to obtain a network prediction result, wherein the network prediction result includes predicted network load and / or network stability parameters, and the data prediction model is obtained by training a long short-term memory network based on a training sample set; and adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth. The power system network structure can be adjusted in real time and dynamically based on the device environment data, avoiding the limitations of traditional static network structure adjustment methods. By adaptively adjusting the network structure in real time according to the real-time collected device environment data, the high performance and high stability of the system network can be continuously maintained, ensuring the continuous and efficient operation of the power system.

[0183] Optionally, the feature processing module 420 is specifically configured to:

[0184] Performing feature extraction on the device environment data to determine a plurality of device environment features, and determining, for each of the device environment features, a feature change rate corresponding to the device environment feature;

[0185] A target feature vector is constructed according to the device environment feature and the feature change rate, and the target feature vector is used as the first environment feature of the device environment data.

[0186] Optionally, the feature processing module 420 is specifically configured to:

[0187] Performing dimensionality reduction on the first environmental feature through a deep autoencoder to determine preliminary low-dimensional features;

[0188] Determining a feature weight according to the preliminary low-latitude feature and the first environmental feature, and determining a weighted low-latitude feature according to the preliminary low-latitude feature and the feature weight;

[0189] Performing nonlinear transformation on the weighted low-dimensional features through principal component analysis to obtain a low-dimensional matrix;

[0190] A second environmental feature corresponding to the first environmental feature is determined according to the weighted low-dimensional feature and the low-dimensional matrix.

[0191] Optionally, the network adjustment module 440 includes: a coefficient limit determination unit, a network structure determination unit, and a network structure adjustment unit;

[0192] The coefficient limit determination unit is configured to determine a plurality of candidate network paths in the target power system, and determine a broadband demand coefficient and a path broadband limit corresponding to the target power system;

[0193] The network structure determining unit is configured to determine, for each candidate network path, the target network path among the candidate network paths and the target network bandwidth corresponding to the target network path according to the bandwidth demand coefficient, the path bandwidth limit, and the network prediction result;

[0194] The network structure adjustment unit is used to adjust the preliminary network structure in real time according to the target network path and the target network bandwidth to determine the target network structure.

[0195] Optionally, the network prediction result includes a sub-prediction result corresponding to each candidate network path, and the sub-prediction result includes a sub-prediction load related to the predicted network load and a sub-prediction stability parameter related to the network stability parameter;

[0196] The network structure determining unit is specifically configured to:

[0197] determining the target network path among the candidate network paths according to the sub-prediction stability parameter;

[0198] Determine a required broadband parameter according to the sub-predicted load corresponding to the target network path and the broadband demand coefficient, and determine a broadband stability parameter according to the required broadband parameter and the sub-predicted stability parameter corresponding to the target network path;

[0199] A target broadband parameter corresponding to the target network path is determined according to the path broadband limit and the broadband stability parameter, and the target network bandwidth is determined according to the target broadband parameter.

[0200] Optionally, the network structure adjustment device based on environmental data further includes: a performance parameter determination module, a prediction model update module and a network structure update module;

[0201] The performance parameter determination module is configured to, after determining the target network structure, determine the target performance parameters of the node devices in the target power system under the target network structure; and determine the preliminary performance parameters of the node devices in the target power system under the preliminary network structure;

[0202] The prediction model updating module is configured to determine a performance improvement parameter based on the preliminary performance parameter and the target performance parameter, and to determine an updated prediction model if the performance improvement parameter does not meet a preset performance condition, wherein the updated prediction model is obtained by training the long short-term memory network based on an updated sample set;

[0203] The network structure updating module is used to perform network prediction on the second environmental feature through the updated prediction model to obtain an updated prediction result, and determine an updated network structure according to the updated prediction result.

[0204] Optionally, the performance parameter determination module is specifically configured to:

[0205] Determining performance-related data corresponding to the node device in the target power system under the target network structure, wherein the performance-related data includes at least one of transmission delay, packet loss rate, throughput, and bandwidth utilization;

[0206] The target performance parameter is determined based on the performance-related data.

[0207] The network structure adjustment device based on environmental data provided in an embodiment of the present invention can execute the network structure adjustment method based on environmental data provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0208] Example 5

[0209] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0210] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0211] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0212] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors that run machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the network structure adjustment method based on environmental data.

[0213] In some embodiments, the network structure adjustment method based on environmental data can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the network structure adjustment method based on environmental data described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the network structure adjustment method based on environmental data in any other appropriate manner (for example, by means of firmware).

[0214] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0215] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0216] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0217] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0218] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0219] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0220] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0221] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A network structure adjustment method based on environmental data, characterized in that: include: Real-time collection of device environment data corresponding to node devices in the target power system under the preliminary network structure, wherein the device environment data includes environment-related data and / or device-related data; Performing feature extraction on the device environment data to obtain a first environment feature, and performing feature dimensionality reduction on the first environment feature to obtain a second environment feature; Performing network prediction on the input second environmental feature using a network prediction model to obtain a network prediction result, wherein the network prediction result includes a predicted network load and / or a network stability parameter, the data prediction model being obtained by training a long short-term memory network based on a training sample set; The preliminary network structure is adjusted in real time according to the network prediction result to determine a target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth.

2. The method according to claim 1, characterized in that The extracting features of the device environment data to obtain a first environment feature includes: Performing feature extraction on the device environment data to determine a plurality of device environment features, and determining, for each of the device environment features, a feature change rate corresponding to the device environment feature; A target feature vector is constructed according to the device environment feature and the feature change rate, and the target feature vector is used as the first environment feature of the device environment data.

3. The method according to claim 1, characterized in that The performing feature dimensionality reduction on the first environmental feature to obtain the second environmental feature includes: Performing dimensionality reduction on the first environmental feature through a deep autoencoder to determine preliminary low-dimensional features; Determining a feature weight according to the preliminary low-latitude feature and the first environmental feature, and determining a weighted low-latitude feature according to the preliminary low-latitude feature and the feature weight; Performing nonlinear transformation on the weighted low-dimensional features through principal component analysis to obtain a low-dimensional matrix; A second environmental feature corresponding to the first environmental feature is determined according to the weighted low-dimensional feature and the low-dimensional matrix.

4. The method according to claim 1, wherein The step of adjusting the preliminary network structure in real time according to the network prediction result to determine the target network structure includes: Determining a plurality of candidate network paths in the target power system, and determining a broadband demand coefficient and a path broadband limit corresponding to the target power system; For each candidate network path, determining the target network path in the candidate network path and the target network bandwidth corresponding to the target network path according to the bandwidth demand coefficient, the path bandwidth limit, and the network prediction result; The preliminary network structure is adjusted in real time according to the target network path and the target network bandwidth to determine the target network structure.

5. The method according to claim 4, characterized in that The network prediction result includes a sub-prediction result corresponding to each candidate network path, wherein the sub-prediction result includes a sub-prediction load related to the predicted network load and a sub-prediction stability parameter related to the network stability parameter; The determining, according to the broadband demand coefficient, the path bandwidth limit, and the network prediction result, the target network path in the candidate network paths and the target network bandwidth corresponding to the target network path includes: determining the target network path among the candidate network paths according to the sub-prediction stability parameter; Determine a required broadband parameter according to the sub-predicted load corresponding to the target network path and the broadband demand coefficient, and determine a broadband stability parameter according to the required broadband parameter and the sub-predicted stability parameter corresponding to the target network path; A target broadband parameter corresponding to the target network path is determined according to the path broadband limit and the broadband stability parameter, and the target network bandwidth is determined according to the target broadband parameter.

6. The method according to claim 1, characterized in that After determining the target network structure, the method further includes: Determining target performance parameters of the node devices in the target power system under the target network structure; and determining preliminary performance parameters of the node devices in the target power system under the preliminary network structure; determining a performance improvement parameter based on the preliminary performance parameter and the target performance parameter, and determining an updated prediction model if the performance improvement parameter does not meet a preset performance condition, wherein the updated prediction model is obtained by training the long short-term memory network based on an updated sample set; The updated prediction model is used to perform network prediction on the second environmental feature to obtain an updated prediction result, and an updated network structure is determined according to the updated prediction result.

7. The method according to claim 6, characterized in that The determining of target performance parameters of the node devices in the target power system under the target network structure includes: Determining performance-related data corresponding to the node device in the target power system under the target network structure, wherein the performance-related data includes at least one of transmission delay, packet loss rate, throughput, and bandwidth utilization; The target performance parameter is determined based on the performance-related data.

8. A network structure adjustment device based on environmental data, characterized in that: include: A data acquisition module, configured to acquire in real time device environment data corresponding to node devices in the target power system under the preliminary network structure, wherein the device environment data includes environment-related data and / or device-related data; a feature processing module, configured to extract features from the device environment data to obtain a first environment feature, and perform feature dimensionality reduction on the first environment feature to obtain a second environment feature; a network prediction module, configured to perform network prediction on the input second environmental feature using a network prediction model to obtain a network prediction result, wherein the network prediction result includes a predicted network load and / or a network stability parameter, wherein the data prediction model is obtained by training a long short-term memory network based on a training sample set; The network adjustment module is used to adjust the preliminary network structure in real time according to the network prediction result to determine the target network structure, wherein the target network structure includes a target network path and / or a target network bandwidth.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the network structure adjustment method based on environmental data according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the network structure adjustment method based on environmental data according to any one of claims 1 to 7 when executed.

Images