Line loss analysis method and system based on edge computing

Abstract

The application relates to an edge-computing-based line loss analysis method, which comprises the following steps: step S1: arranging sensors in a power distribution network to obtain power distribution network data; step S2: obtaining the power distribution network data based on the sensors arranged in the power distribution network and transmitting the data to an edge computing node; step S3: the edge computing node pre-processes the obtained power distribution network data, adopts an SVM to preliminarily analyze line loss, judges whether there is an abnormality, if there is no abnormality, stores the data into a local database, and if there is an abnormality, uploads the abnormal data to a cloud server; step S4: the cloud server further analyzes the abnormal data uploaded by the edge computing node, classifies line loss abnormalities based on a multilayer perceptron, and obtains a corresponding processing scheme; and step S5: pushing the obtained processing scheme to a mobile terminal of a corresponding personnel. The application can not only improve the safety of line loss data transmission, but also effectively improve line loss analysis quality and efficiency.

Description

Technical Field

[0001] This invention relates to the field of power grid detection, and in particular to a line loss analysis method and system based on edge computing. Background Technology

[0002] Building a resource-saving and environmentally friendly society has increasingly become a focus of public attention. This places new and higher demands on line loss management in power systems. However, due to line loss management issues, significant power losses still exist in power supply management systems. Therefore, how to take proactive and effective measures to improve line loss management and reduce power consumption losses is a crucial issue that relevant power management departments urgently need to address.

[0003] Existing line loss analysis methods often involve collecting distribution data uniformly to a cloud or terminal server, and then performing data analysis through the data analysis model built into the cloud or terminal service. However, due to the large number of users and the large amount of data collected through smart meters, various reasons can easily lead to data loss or corruption during transmission, resulting in inaccurate data and seriously affecting the analysis results. Moreover, the data transmission process is insecure and vulnerable to unauthorized access and attacks, which can lead to the leakage and misuse of sensitive information.

[0004] Moreover, since power grids are mostly located in mountainous areas, the data transmission distance is long and the signal is poor, resulting in low real-time performance and the inability to detect abnormalities in a timely manner, which affects the normal operation quality of the power grid. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a line loss analysis method and system based on edge computing, which can not only improve the security of line loss data transmission, but also effectively improve the quality and efficiency of line loss analysis.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A line loss analysis method based on edge computing includes the following steps: Step S1: Determine the number and location of sensor nodes to be deployed based on the scale and complexity of the power distribution network; determine the number and location of edge computing nodes to be deployed based on the number of sensor nodes and the sampling frequency; and encrypt the sensor data between the sensor nodes and the edge computing nodes using an encryption algorithm. Step S2: Acquire power distribution network data, including current, voltage, and power, based on sensors deployed in the power distribution network, and transmit it to the edge computing node; Step S3: The edge computing node preprocesses the acquired power distribution network data and performs preliminary line loss analysis using SVM to determine if there are any anomalies. If there are no anomalies, the data is stored in the local database; if there are anomalies, the abnormal data is uploaded to the cloud server. Step S4: The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies the line loss anomalies based on the multilayer sensor, and obtains the corresponding processing solutions. Step S5: Push the obtained processing solution to the mobile terminal of the relevant personnel.

[0007] Furthermore, the encryption algorithm is as follows: Generate key pairs on edge computing nodes, including a public key and a private key. The private key is used to decrypt data, and the public key is used to encrypt data. At the sensor nodes, the collected line loss data is encrypted using the public key of the edge computing nodes. The encrypted data can only be decrypted using the private key of the edge computing node, ensuring the confidentiality of the data during transmission. The encrypted data is transmitted to the edge computing node, which uses its private key to decrypt the encrypted data and obtain the original line loss data.

[0008] Furthermore, the generation of the key pair is as follows: (1) Generate a series of round keys based on the preset initial key, which are used for the round functions in the encryption and decryption process; (2) In the initial round, the input data is XORed with the key from the first round; (3) Repeat the function N times, using a different round key in each round; The round function includes four basic operations: byte substitution, row shifting, column obfuscation, and round key addition; Byte substitution: Replace each byte of the input with the corresponding byte in a predefined S-box; Row shift: Perform a circular left shift operation on each line of the input; Column obfuscation: Performs a series of mathematical operations on each column of the input, including multiplication, addition, and XOR operations; Round key addition: Performs an XOR operation between the key of the current round and the input data; (4) The last round of the round function does not include column obfuscation operation, and the output is the encrypted data.

[0009] Furthermore, an access control mechanism is implemented on the edge computing node to ensure that only authorized users can access and use the data. Specifically: Users send their username and password to the edge computing node for authentication, and the edge computing node verifies whether the username and password provided by Alice are correct; If the verification is successful, the edge computing node confirms the user's identity and allows her to continue accessing data and services; Edge computing nodes determine the data and services accessed based on the user's identity and role; Edge computing nodes use role-based access control mechanisms to manage user permissions.

[0010] Furthermore, step S3 is specifically as follows: (1) Perform data cleaning on the acquired distribution network data. Data cleaning includes removing outliers, filling in missing values, and converting data formats. (2) Calculate line loss data based on the preprocessed distribution network data:

[0011] in, Here, Q represents the line loss power, and R and X represent the resistance and reactance in the line impedance Z; U1 and U2 represent the line voltages at both ends of the line; and G1 and G2 represent the line conductance to ground. a is Correlation coefficient; b is the maximum transmission power of the line; c is the line loss at maximum power; e and For sensor measurement error; (3) Construct an SVM-based line loss anomaly detection model, input the calculated line loss power and line loss rate into the line loss anomaly detection model, obtain an anomaly score, which is used to represent the degree of deviation of the sample from the normal state, and determine whether it is an abnormal situation according to the set threshold.

[0012] Furthermore, the line loss anomaly detection model is as follows: Based on historical line loss data, an SVM model is trained using a linear kernel function, and its decision function is: f(x) = w^T x + b Where w is the normal vector of the hyperplane, b is the intercept of the hyperplane, and x is the feature vector of the sample; For a new line loss data x, project it onto the hyperplane to obtain its distance d from the hyperplane: d = |w^T x + b| / ||w|| Where ||w|| is the norm of the hyperplane; The anomaly score is defined as the ratio of the distance from a sample to the hyperplane to the mean distance from all samples in the training dataset to the hyperplane. score(x) = d / mean(d) Where mean(d) is the mean distance of all samples in the training dataset to the hyperplane.

[0013] Furthermore, the perceptron consists of an input layer, hidden layers, and an output layer. Each neuron has a weight and a bias, specifically: The input layer receives normalized line loss data; The hidden layer undergoes a non-linear transformation:

[0014] in, The output of the hidden layer neurons, The weights between the input layer and the hidden layer. Input data; Bias for hidden layer neurons; For activation functions; Classification results of output layer output line loss anomalies:

[0015] in, The output of the output layer neurons. The weights between the hidden layer and the output layer. Bias for output layer neurons; This is the activation function.

[0016] A line loss analysis system based on edge computing includes a cloud server, edge computing nodes, sensors installed in the power distribution network, and mobile terminals. The sensors in the power distribution network acquire power distribution network data and transmit it to the edge computing nodes. The edge computing nodes preprocess the acquired power distribution network data and perform preliminary line loss analysis using SVM to determine if any anomalies exist. If no anomalies are found, the data is stored in a local database; if anomalies are found, the abnormal data is uploaded to the cloud server. The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies line loss anomalies based on a multilayer sensor, obtains corresponding processing solutions, and pushes the obtained processing solutions to the mobile terminals of relevant personnel.

[0017] Furthermore, the sensors include current sensors, voltage sensors, power factor sensors, frequency sensors, and temperature sensors; the sensors are installed in substations, distribution boxes, and cables in the power distribution network to monitor current, voltage, power factor, frequency, and temperature, and transmit the data to edge computing nodes via wireless communication.

[0018] The present invention has the following beneficial effects: This invention combines edge computing and cloud computing, setting up preliminary line loss anomaly detection at edge computing nodes. This allows for more accurate and localized data processing, enabling simple anomaly detection, improving data processing efficiency, and transmitting only abnormal data to the cloud for further analysis. This significantly reduces data transmission volume and saves bandwidth, effectively improving the efficiency and quality of line loss data analysis while ensuring data security. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: refer to Figure 1 In this embodiment, a line loss analysis method based on edge computing includes the following steps: Step S1: Determine the number and location of sensor nodes to be deployed based on the scale and complexity of the power distribution network; determine the number and location of edge computing nodes to be deployed based on the number of sensor nodes and the sampling frequency; and encrypt the sensor data between the sensor nodes and the edge computing nodes using an encryption algorithm. Step S2: Acquire power distribution network data, including current, voltage, and power, based on sensors deployed in the power distribution network, and transmit it to the edge computing node; Step S3: The edge computing node preprocesses the acquired power distribution network data and performs preliminary line loss analysis using SVM to determine if there are any anomalies. If there are no anomalies, the data is stored in the local database; if there are anomalies, the abnormal data is uploaded to the cloud server. Step S4: The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies the line loss anomalies based on the multilayer sensor, and obtains the corresponding processing solutions. Step S5: Push the obtained processing solution to the mobile terminal of the relevant personnel.

[0021] In this embodiment, the encryption algorithm is as follows: Generate key pairs on edge computing nodes, including a public key and a private key. The private key is used to decrypt data, and the public key is used to encrypt data. At the sensor nodes, the collected line loss data is encrypted using the public key of the edge computing nodes. The encrypted data can only be decrypted using the private key of the edge computing node, ensuring the confidentiality of the data during transmission. The encrypted data is transmitted to the edge computing node, which uses its private key to decrypt the encrypted data and obtain the original line loss data.

[0022] In this embodiment, the key pair is generated as follows: (1) Generate a series of round keys based on the preset initial key, which are used for the round functions in the encryption and decryption process; (2) In the initial round, the input data is XORed with the key from the first round; (3) Repeat the function N times, using a different round key in each round; The round function includes four basic operations: byte substitution, row shifting, column obfuscation, and round key addition; Byte substitution: Replace each byte of the input with the corresponding byte in a predefined S-box; Row shift: Perform a circular left shift operation on each line of the input; Column obfuscation: Performs a series of mathematical operations on each column of the input, including multiplication, addition, and XOR operations; Round key addition: Performs an XOR operation between the key of the current round and the input data; (4) The last round of the round function does not include column obfuscation operation, and the output is the encrypted data.

[0023] In this embodiment, an access control mechanism is implemented on the edge computing node to ensure that only authorized users can access and use the data. Specifically: Users send their username and password to the edge computing node for authentication, and the edge computing node verifies whether the username and password provided by Alice are correct; If the verification is successful, the edge computing node confirms the user's identity and allows her to continue accessing data and services; Edge computing nodes determine the data and services accessed based on the user's identity and role; Edge computing nodes use role-based access control mechanisms to manage user permissions.

[0024] In this embodiment, step S3 is as follows: (1) Perform data cleaning on the acquired distribution network data. Data cleaning includes removing outliers, filling in missing values, and converting data formats. (2) Calculate line loss data based on the preprocessed distribution network data:

[0025] in, Here, Q represents the line loss power, and R and X represent the resistance and reactance in the line impedance Z; U1 and U2 represent the line voltages at both ends of the line; and G1 and G2 represent the line conductance to ground. a is Correlation coefficient; b is the maximum transmission power of the line; c is the line loss at maximum power; e and For sensor measurement error; (3) Construct an SVM-based line loss anomaly detection model, input the calculated line loss power and line loss rate into the line loss anomaly detection model, obtain an anomaly score, which is used to represent the degree of deviation of the sample from the normal state, and determine whether it is an abnormal situation according to the set threshold.

[0026] In this embodiment, the line loss anomaly detection model is as follows: Based on historical line loss data, an SVM model is trained using a linear kernel function, and its decision function is: f(x) = w^T x + b Where w is the normal vector of the hyperplane, b is the intercept of the hyperplane, and x is the feature vector of the sample; For a new line loss data x, project it onto the hyperplane to obtain its distance d from the hyperplane: d = |w^T x + b| / ||w|| Where ||w|| is the norm of the hyperplane; The anomaly score is defined as the ratio of the distance from a sample to the hyperplane to the mean distance from all samples in the training dataset to the hyperplane. score(x) = d / mean(d) Where mean(d) is the mean distance of all samples in the training dataset to the hyperplane.

[0027] In this embodiment, the training dataset is set as follows:

[0028] Use this data to train an SVM model, and use a linear kernel function to obtain an SVM-based line loss anomaly detection model; Consider a new sample with a line loss rate of 0.05 and a line loss power of 180. We project it onto a hyperplane and obtain its distance d from the hyperplane as: d = |w^T x + b| / ||w|| = (0.05 0.1 + 180 0.2 - 1) / sqrt(0.1^2 + 0.2^2) = 33.97 The mean (d) of the distances from all samples in the training dataset to the hyperplane is calculated as follows: mean(d) = (0.02 0.1 + 100 0.2 - 1 + 0.03 0.1 + 120 0.2 - 1 + 0.01 0.1+ 80 0.2 - 1 + 0.04 0.1 + 150 0.2 - 1 + 0.02 0.1 + 110 0.2 - 1 + 0.01 0.1 +90 0.2 - 1 + 0.03 0.1 + 130 0.2 - 1 + 0.02 0.1 + 100 0.2 - 1 + 0.01 0.1 + 85 0.2 - 1 + 0.02 0.1 + 105 0.2 - 1) / sqrt(0.1^2 + 0.2^2) = 32.77 Therefore, we can calculate the anomaly score for this sample as follows: score(x) = d / mean(d) = 33.97 / 32.77 = 1.04 If the threshold is set to 1, the sample will be considered an anomaly.

[0029] Furthermore, the perceptron consists of an input layer, hidden layers, and an output layer. Each neuron has a weight and a bias, specifically: The input layer receives normalized line loss data; The hidden layer undergoes a non-linear transformation:

[0030] in, The output of the hidden layer neurons, The weights between the input layer and the hidden layer. Input data; Bias for hidden layer neurons; For activation functions; Classification results of output layer output line loss anomalies:

[0031] in, The output of the output layer neurons. The weights between the hidden layer and the output layer. Bias for output layer neurons; This is the activation function.

[0032] In this embodiment, a line loss analysis system based on edge computing is also provided, including a cloud server, edge computing nodes, sensors installed in the power distribution network, and mobile terminals. The sensors installed in the power distribution network acquire power distribution network data and transmit it to the edge computing nodes. The edge computing nodes preprocess the acquired power distribution network data and perform preliminary line loss analysis using SVM to determine if any anomalies exist. If no anomalies are found, the data is stored in a local database; if anomalies are found, the abnormal data is uploaded to the cloud server. The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies the line loss anomalies based on a multilayer sensor, obtains corresponding processing solutions, and pushes the obtained processing solutions to the mobile terminals of the relevant personnel.

[0033] Preferably, the sensors include a current sensor, a voltage sensor, a power factor sensor, a frequency sensor, and a temperature sensor; the sensors are installed in substations, distribution boxes, cables, etc. in the power distribution network to monitor current, voltage, power factor, frequency, and temperature, and transmit the data to the edge computing node via wireless communication.

[0034] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0035] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0036] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0037] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0038] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A line loss analysis method based on edge computing, characterized in that, Includes the following steps: Step S1: Determine the number and location of sensor nodes to be deployed based on the scale and complexity of the power distribution network; determine the number and location of edge computing nodes to be deployed based on the number of sensor nodes and the sampling frequency; and encrypt the sensor data between the sensor nodes and the edge computing nodes using an encryption algorithm. Step S2: Acquire power distribution network data based on sensors deployed in the power distribution network and transmit it to the edge computing node; Step S3: The edge computing node preprocesses the acquired power distribution network data and performs preliminary line loss analysis using SVM to determine if there are any anomalies. If there are no anomalies, the data is stored in the local database; if there are anomalies, the abnormal data is uploaded to the cloud server. Step S4: The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies the line loss anomalies based on the multilayer sensor, and obtains the corresponding processing solutions. Step S5: Push the obtained processing solution to the mobile terminal of the relevant personnel; The encryption algorithm is as follows: Generate key pairs on edge computing nodes, including a public key and a private key. The private key is used to decrypt data, and the public key is used to encrypt data. At the sensor nodes, the collected line loss data is encrypted using the public key of the edge computing nodes. The encrypted data can only be decrypted using the private key of the edge computing node, ensuring the confidentiality of the data during transmission. The encrypted data is transmitted to the edge computing node, which uses its private key to decrypt the encrypted data and obtain the original line loss data. The key pair is generated as follows: (1) Generate a series of round keys based on the preset initial key, which are used for the round functions in the encryption and decryption process; (2) In the initial round, the input data is XORed with the key from the first round; (3) Repeat the function N times, using a different round key in each round; The round function includes four basic operations: byte substitution, row shifting, column obfuscation, and round key addition; Byte substitution: Replace each byte of the input with the corresponding byte in a predefined S-box; Row shift: Perform a circular left shift operation on each line of the input; Column obfuscation: Performs a series of mathematical operations on each column of the input, including multiplication, addition, and XOR operations; Round key addition: Performs an XOR operation between the key of the current round and the input data; (4) The last round function does not include column obfuscation operations, and the output is the encrypted data; Implement access control mechanisms on edge computing nodes to ensure that only authorized users can access and use data. Specifically: Users send their username and password to the edge computing node for authentication, and the edge computing node verifies whether the username and password provided by Alice are correct; If the verification is successful, the edge computing node confirms the user's identity and allows her to continue accessing data and services; Edge computing nodes determine the data and services accessed based on the user's identity and role; Edge computing nodes use role-based access control mechanisms to manage user permissions; Step S3 is as follows: (1) Perform data cleaning on the acquired distribution network data. Data cleaning includes removing outliers, filling in missing values, and converting data formats. (2) Calculate line loss data based on the preprocessed distribution network data: ； in, Here, Q represents the line loss power, and R and X represent the resistance and reactance in the line impedance Z; U1 and U2 represent the line voltages at both ends of the line; and G1 and G2 represent the line conductance to ground. a is Correlation coefficient; b is the maximum transmission power of the line; c is the line loss at maximum power; e and For sensor measurement error; (3) Construct an SVM-based line loss anomaly detection model, input the calculated line loss power and line loss rate into the line loss anomaly detection model, obtain an anomaly score, which is used to represent the degree of deviation of the sample from the normal state, and determine whether it is an abnormal situation according to the set threshold.

2. The line loss analysis method based on edge computing according to claim 1, characterized in that, The line loss anomaly detection model is as follows: Based on historical line loss data, an SVM model is trained using a linear kernel function, and its decision function is: f(x) = w^T x + b Where w is the normal vector of the hyperplane, b is the intercept of the hyperplane, and x is the feature vector of the sample; For a new line loss data x, project it onto the hyperplane to obtain its distance d from the hyperplane: d = |w^T x + b| / ||w|| Where ||w|| is the norm of the hyperplane; The anomaly score is defined as the ratio of the distance from a sample to the hyperplane to the mean distance from all samples in the training dataset to the hyperplane. score(x) = d / mean(d) Where mean(d) is the mean distance from all samples in the training dataset to the hyperplane.

3. The line loss analysis method based on edge computing according to claim 1, characterized in that, The multilayer perceptron consists of an input layer, a hidden layer, and an output layer. Each neuron has a weight and a bias, specifically: The input layer receives normalized line loss data; The hidden layer undergoes a non-linear transformation: ； in, The output of the hidden layer neurons, The weights between the input layer and the hidden layer. Input data; Bias for hidden layer neurons; For activation functions; Classification results of output layer output line loss anomalies: ； in, The output of the output layer neurons. The weights between the hidden layer and the output layer. Bias for output layer neurons; This is the activation function.

4. The line loss analysis method based on edge computing according to claim 1, characterized in that, The power distribution network data includes current, voltage, and power.

5. A system based on the edge computing-based line loss analysis method according to any one of claims 1-4, characterized in that, The system includes a cloud server, edge computing nodes, sensors installed in the power distribution network, and mobile terminals. The sensors in the power distribution network acquire power distribution network data and transmit it to the edge computing nodes. The edge computing nodes preprocess the acquired power distribution network data and perform preliminary line loss analysis using SVM to determine if any anomalies exist. If no anomalies are found, the data is stored in a local database; if anomalies are found, the abnormal data is uploaded to the cloud server. The cloud server further analyzes the abnormal data uploaded by the edge computing nodes, classifies line loss anomalies based on a multilayer sensor, obtains corresponding processing solutions, and pushes the obtained processing solutions to the mobile terminals of the relevant personnel.

6. The system according to claim 5, characterized in that, The sensors include current sensors, voltage sensors, power factor sensors, frequency sensors, and temperature sensors. The sensors are installed in substations, distribution boxes, and cables in the power distribution network to monitor current, voltage, power factor, frequency, and temperature, and transmit the data to edge computing nodes via wireless communication.

Images