A 10kV distribution network safe operation management system based on intelligent monitoring

Abstract

The present invention relates to the technical field of intelligent management systems for distribution networks, specifically a 10kV distribution network safe operation management system based on intelligent monitoring. Through multimodal data fusion, edge computing, and cloud-based federated decision-making technologies, it achieves real-time status monitoring, risk prediction, and dynamic protection strategy generation for the distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network management systems, and more particularly, to a 10kV distribution network safe operation management system based on intelligent monitoring. Background Art

[0002] With the advancement of smart grid construction, the safe operation and management technology of 10kV distribution networks continues to develop. However, the existing management systems have the following shortcomings: most traditional systems use SCADA systems, which rely solely on electrical quantity monitoring (such as voltage and current) and cannot perceive the coupling risk of equipment physical status (such as temperature and deformation) and environmental factors (such as SF6 concentration). The monitoring dimension is single, and centralized data processing leads to fault prediction delays greater than 200ms, which cannot meet the rapid response requirements of transient faults (such as arc flash). Data in each area is stored independently, the accuracy of global risk assessment is limited, and the misjudgment rate is high. Therefore, a 10kV distribution network safe operation and management system based on intelligent monitoring is provided. Through multimodal data fusion, edge computing and cloud-based federated decision-making technology, real-time status monitoring, risk prediction and dynamic protection strategy generation of the distribution network are achieved. Summary of the Invention

[0003] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a 10kV distribution network safe operation management system based on intelligent monitoring to solve the problems existing in the above-mentioned background technology.

[0004] The above technical objectives of the present invention are achieved through the following technical solutions: A 10kV distribution network safe operation management system based on intelligent monitoring, comprising: several intelligent monitoring sub-terminals arranged on each distribution network node and a cloud decision-making platform arranged on the system terminal; several of the intelligent monitoring sub-terminals are respectively communicated with the cloud decision-making platform; the intelligent monitoring sub-terminals include: a data acquisition module for collecting multimodal data on the distribution network nodes and an edge computing cluster for extracting features from the collected multimodal data and then transmitting them; the cloud decision platform includes: a global risk prediction module for performing risk prediction after receiving the feature vector transmitted by the edge computing cluster and a dynamic protection strategy generation module for generating corresponding protection adjustment plans based on the calculation results of the global risk prediction module.

[0005] Optionally, the multimodal data includes: electrical parameters: three-phase current signals, three-phase voltage signals, fundamental components of three-phase voltage, fundamental components of three-phase current, effective values ​​of harmonics, total harmonic distortion and phase coupling relationship between harmonics of nodes in the distribution network; thermal imaging parameters: visible light image of equipment, infrared image of equipment, maximum surface temperature of equipment, temperature gradient, area ratio of hot spot area and standard deviation of temperature distribution.

[0006] Optionally, the edge computing cluster includes: a harmonic depth analysis unit for extracting harmonic distortion features and time series correlations of the multimodal data and a thermal imaging feature modeling unit for identifying device hotspot areas based on image segmentation technology and quantifying temperature gradient distribution to extract thermal imaging features; the harmonic depth analysis unit includes: an optimized window function processing unit for suppressing spectral leakage using an optimized window function with a fourth-order cosine component and an inter-harmonic correlation analysis unit for performing time series-based inter-harmonic correlation feature extraction.

[0007] Optionally, the feature extraction of the harmonic depth analysis unit includes: the optimized window function processing unit uses the optimized window function to perform a windowed Fourier transform on the voltage signal to suppress spectrum leakage; the inter-harmonic correlation analysis unit calculates the odd harmonic skewness and the Pearson correlation coefficient matrix of adjacent harmonics, extracts the distribution characteristics of harmonic energy in the 30-1500Hz frequency band, and then calculates and obtains the harmonic distortion characteristics.

[0008] Optionally, the feature extraction of the thermal imaging feature modeling module includes: the thermal imaging feature modeling module identifies key areas of the equipment based on an image segmentation algorithm, delineates the temperature analysis hot temperature area ROI, calculates the maximum temperature gradient and radial attenuation coefficient within the ROI, and counts the proportion of pixels exceeding a preset temperature threshold and the spatial distribution discreteness to thereby calculate the thermal imaging feature.

[0009] Optionally, the intelligent monitoring sub-terminal also includes: a gas detection module for detecting changes in SF6 concentration and triggering a gas insulation degradation warning, a temperature and humidity sensing module for detecting temperature and humidity, and an environmental compensation module for dynamically calibrating electrical quantity measurement values ​​based on the detection data of the humidity sensing module.

[0010] Optionally, the global risk prediction module includes: a federal feature fusion engine, used to receive feature vectors transmitted by several of the intelligent monitoring sub-terminals, and use feature fusion to generate a multidimensional joint representation vector and then transmit it; a federated learning aggregator, used to receive several of the transmitted multidimensional joint representation vectors, aggregate several of the multidimensional joint representation vectors at regular time intervals, and output a global risk prediction model; a risk assessment output layer, used to predict and generate a three-dimensional risk situation map through the global risk prediction model, and synchronously output the risk level and confidence interval.

[0011] Optionally, the dynamic protection strategy generation module includes: a digital twin engine, which generates a dynamic adjustment strategy through the three-dimensional risk situation map; a blockchain consensus executor, which receives the generated dynamic adjustment strategy and uses an improved practical Byzantine fault tolerance algorithm to generate dynamic adjustment strategy instructions, and the dynamic adjustment strategy instructions need to be verified by more than 2 / 3 of the nodes before execution; a strategy feedback optimizer, which dynamically optimizes strategy parameters according to the execution effect, and simultaneously generates operation and maintenance recommendation strategies.

[0012] In summary, the present invention has the following beneficial effects:

[0013] 1. The system makes judgments from the perspective of multimodal data and can simultaneously analyze the correlation between electrical quantities, equipment temperature and environmental parameters, which significantly improves the fault recognition rate of the system and reduces the possibility of misjudgment.

[0014] 2. At the same time, the intelligent monitoring sub-terminal set up at each configuration node accelerates feature extraction and optimizes data response speed, avoiding data delays caused by centralized data processing in traditional systems, meeting transient fault response requirements, and further improving system reaction speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the system operation logic of the present invention. DETAILED DESCRIPTION

[0016] To make the objectives, features, and advantages of the present invention more readily apparent, the following detailed description of the present invention is provided with reference to the accompanying drawings. The accompanying drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein.

[0017] In the present invention, unless otherwise expressly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of such features.

[0018] In the present invention, unless otherwise expressly specified and limited, a first feature being "above" or "below" a second feature may include the first and second features being in direct contact, or may include the first and second features being in contact not directly but through another feature between them. Moreover, a first feature being "above," "above," and "above" a second feature includes the first feature being directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below," "below," and "below" a second feature includes the first feature being directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature. The terms "vertical," "horizontal," "left," "right," "above," "below," and similar expressions are for illustrative purposes only and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore should not be understood as limiting the present invention.

[0019] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0020] The present invention provides a 10kV distribution network safe operation management system based on intelligent monitoring, such as Figure 1 As shown, it includes: several intelligent monitoring sub-terminals arranged on each distribution network node and a cloud decision platform arranged on the system terminal; several of the intelligent monitoring sub-terminals are respectively communicated with the cloud decision platform; the intelligent monitoring sub-terminals include: a data acquisition module for collecting multimodal data on the distribution network nodes and an edge computing cluster for extracting features from the collected multimodal data and then transmitting them; the cloud decision platform includes: a global risk prediction module for performing risk prediction after receiving the feature vector transmitted by the edge computing cluster and a dynamic protection strategy generation module for generating a corresponding protection adjustment plan according to the calculation result of the global risk prediction module; thereby realizing the end-edge-cloud fault judgment logic and further increasing the scenario adaptability.

[0021] During the specific implementation process, an intelligent monitoring sub-terminal is installed on each node of the distribution network (which can be a transmission node or a transfer node, etc.), and the multimodal data on the node is obtained through the data acquisition module. After feature extraction is performed through the edge computing cluster of the intelligent monitoring sub-terminal, the feature vector is transmitted to the cloud decision-making platform. The global risk prediction module calculates and predicts the transmitted feature vector, predicts the danger level, and generates a dynamic protection strategy through the dynamic protection strategy generation module, thereby protecting and ensuring the normal operation of the distribution network and avoiding serious property losses.

[0022] Furthermore, the multimodal data includes: electrical parameters: three-phase current signals, three-phase voltage signals, fundamental components of three-phase voltage, fundamental components of three-phase current, effective values ​​of harmonics, total harmonic distortion rate and phase coupling relationship between harmonics of nodes in the distribution network; thermal imaging parameters: visible light image of equipment, infrared image of equipment, maximum surface temperature of equipment, temperature gradient, area ratio of hot spot area and standard deviation of temperature distribution.

[0023] In Example 1, the three-phase voltage / current signals in the frequency band of 0.1Hz-10MHz are collected through the wide-band voltage sensing unit on the data acquisition module, and the fundamental and harmonic components are extracted synchronously to obtain the electrical parameters in the multimodal data; the visible light image and infrared thermal image of the device are captured at a frame rate of 25Hz through the multispectral imaging unit on the data acquisition module to obtain the thermal imaging parameters in the multimodal data.

[0024] Furthermore, the edge computing cluster includes: a harmonic depth analysis unit for extracting the harmonic distortion characteristics and time series correlation of the multimodal data and a thermal imaging feature modeling unit for identifying the hot spots of the equipment based on image segmentation technology and quantifying the temperature gradient distribution to extract thermal imaging features; the harmonic depth analysis unit includes: an optimized window function processing unit for suppressing spectral leakage using an optimized window function with a fourth-order cosine component and an inter-harmonic correlation analysis unit for performing time series-based inter-harmonic correlation feature extraction.

[0025] Furthermore, the feature extraction of the harmonic depth analysis unit includes: the optimized window function processing unit uses the optimized window function to perform a windowed Fourier transform on the voltage signal to suppress spectrum leakage; the inter-harmonic correlation analysis unit calculates the odd harmonic skewness and the Pearson correlation coefficient matrix of adjacent harmonics, extracts the distribution characteristics of harmonic energy in the 30-1500Hz frequency band, and then calculates and obtains the harmonic distortion characteristics.

[0026] Furthermore, the feature extraction of the thermal imaging feature modeling module includes: the thermal imaging feature modeling module identifies the key areas of the equipment based on the image segmentation algorithm, delineates the temperature analysis hot temperature area ROI, calculates the maximum temperature gradient and radial attenuation coefficient in the ROI, counts the proportion of pixels exceeding the preset temperature threshold and the spatial distribution discreteness, and then calculates the thermal imaging feature.

[0027] In the specific implementation process, the maximum temperature gradient and radial attenuation coefficient within the ROI are calculated, and the proportion of pixels exceeding the preset temperature threshold and the spatial distribution dispersion are counted to calculate the thermal imaging characteristics, where the infrared thermal image temperature modeling formula is expressed as: , Expressed as the actual temperature of the device surface coordinate (x, y), It is represented by the radiation energy received by the infrared sensor. Expressed as the device surface emissivity, is expressed as the Stefan-Boltzmann constant, Expressed as ambient temperature; the temperature gradient within the ROI is calculated based on the modeling results and expressed as ,in It is expressed as the maximum temperature in the ROI area. It is expressed as the average temperature in the ROI area. It is expressed as the radial distance from the center to the edge of the hotspot, and then the thermal image characteristics are obtained.

[0028] Furthermore, the intelligent monitoring sub-terminal also includes: a gas detection module for detecting changes in SF6 concentration and triggering a gas insulation degradation warning, a temperature and humidity sensing module for detecting temperature and humidity, and an environmental compensation module for dynamically calibrating electrical quantity measurement values ​​based on the detection data of the humidity sensing module.

[0029] In the second embodiment, in order to deal with the measurement deviation caused by environmental factors at the distribution network node, the data environment compensation formula is adopted, which is expressed as ,in Expressed as the effective value of the voltage after compensation, Expressed as the original voltage value, Expressed as the temperature compensation coefficient, Expressed as ambient temperature, Expressed as the temperature compensation coefficient, Expressed as ambient relative humidity.

[0030] Furthermore, the global risk prediction module includes: a federal feature fusion engine, used to receive feature vectors transmitted by several of the intelligent monitoring sub-terminals, and use feature fusion to generate a multidimensional joint representation vector and then transmit it; a federated learning aggregator, used to receive several of the transmitted multidimensional joint representation vectors, aggregate several of the multidimensional joint representation vectors at regular time intervals, and output a global risk prediction model; a risk assessment output layer, used to predict and generate a three-dimensional risk situation map through the global risk prediction model, and synchronously output the risk level and confidence interval.

[0031] In Example 3, the federated feature fusion engine consists of a cross-modal attention layer and a spatiotemporal convolutional network. The cross-modal attention layer adopts a multi-head attention mechanism, and the weight distribution formula is: ,in is represented as a trainable parameter matrix, q is represented as a query vector, Represented as a key vector, the spatiotemporal convolutional network consists of three layers of dilated convolutions with dilation factors d = 1, 2, and 4 to capture harmonic timing characteristics. It receives 18-dimensional feature vectors (harmonic distortion rate, temperature gradient) from each edge computing cluster and generates a 128-dimensional joint representation vector through feature fusion.

[0032] The federated learning aggregator consists of a dynamic weighting unit and a differential injection unit. The dynamic weighting unit calculates the aggregation weight according to the node data quality (data loss rate) and model accuracy, which is expressed as ,in Expressed as the aggregate weight of the kth edge node, Expressed as the prediction accuracy of the k-th local node model, The data missing rate of the kth node, K is represented by the total number of edge nodes participating in federated learning; the differential injection unit is used to add Laplace noise, which is expressed as ,in Expressed as a Laplace distribution; aggregate edge node model parameters periodically at regular intervals to output a high-precision global risk prediction model;

[0033] The risk assessment output layer consists of a Softmax classifier and an interpretable analysis unit. The Softmax classifier is used to compare the prediction results of the global risk prediction model with the preset warning threshold (warning ≥ 0.7, failure ≥ 0.9) and output three-level probabilities of normal / warning / failure. The interpretable analysis unit quantifies the contribution of each feature based on the SHAP value; it generates a three-dimensional risk situation map (spatial resolution 0.5m, time granularity 50ms) and simultaneously outputs the risk level and confidence interval. For example, the failure probability of cable connector T23 is 92% ± 3%.

[0034] Furthermore, the dynamic protection strategy generation module includes: a digital twin engine, which generates a dynamic adjustment strategy through the three-dimensional risk situation map; a blockchain consensus executor, which receives the generated dynamic adjustment strategy and uses an improved practical Byzantine fault tolerance algorithm to generate dynamic adjustment strategy instructions. The dynamic adjustment strategy instructions must be verified by more than 2 / 3 of the nodes before execution; a strategy feedback optimizer, which dynamically optimizes strategy parameters according to the execution effect and simultaneously generates operation and maintenance recommendation strategies.

[0035] In the fourth embodiment, the digital twin engine is composed of a real-time state mapping unit, wherein the real-time state mapping unit generates a dynamic adjustment strategy through a three-dimensional risk situation map generated in real time;

[0036] The blockchain consensus executor consists of a PBFT consensus network and a smart contract library. The PBFT consensus network contains several verification nodes, with the number of verification nodes N ≥ 4 and the fault tolerance f = 1 (satisfying N ≥ 3f + 1). The smart contract library contains 12 preset policy instructions, such as overcurrent protection and arc isolation, to extract the preset policy instructions corresponding to the dynamic adjustment policy, thereby generating dynamic adjustment policy instructions. During the implementation of dynamic policy instructions, verification nodes of the PBFT consensus network are required to verify the legitimacy of the protection instructions (requires confirmation by 2 / 3 of the nodes). After verification, the operation hash value is recorded for subsequent traceability operations.

[0037] The policy feedback optimizer consists of a reinforcement learning agent processor and a historical case library. The reinforcement learning adopts the DQN algorithm, and the reward function is designed as The historical case library is used to store historical fault handling records; it dynamically optimizes strategy parameters based on execution results and generates operation and maintenance recommendations (such as "Line L12 requires priority inspection").

[0038] In other embodiments, to ensure effective response to transient faults, the digital twin engine can also receive real-time parameters to make corresponding protection setting adjustment coefficient strategies, which are expressed as , where k represents the protection setting adjustment coefficient, Indicates the maximum operating temperature allowed for the device. Indicates real-time monitoring of temperature. It is expressed as the rate of change of total harmonic distortion over time. The protection setting adjustment coefficient strategy is obtained through rapid response, thereby coping with transient faults.

[0039] In one specific embodiment, if a cable joint in any distribution network overheats, during the data collection process, the broadband sensor detects a sudden increase in the current harmonic distortion rate, and the infrared thermal image shows that the joint temperature rises from 45°C to 82°C, with a temperature gradient of 15°C / cm; edge computing: the inter-harmonic correlation analysis unit identifies the abnormal amplitude of the third harmonic, and the thermal image feature modeling unit calculates the hotspot area ratio of up to 12%; on the cloud decision platform, the federated feature fusion engine integrates data from 10 edge nodes, and the federated learning aggregator aggregates the data and outputs it to the risk assessment output layer, outputting a risk value of 0.93. The digital twin engine generates a dynamic adjustment strategy: dynamically adjusts the overcurrent protection threshold of the line from 120% of the rated value to 95%, and the blockchain consensus executor triggers the circuit breaker to trip to prevent the cable insulation layer from melting. The entire event process (including multimodal data) is stored on the chain to generate a unique hash identifier.

[0040] In one embodiment, if a lightning strike occurs at any distribution network node, during the data collection process, the wide-band sensor captures 10MHz high-frequency transient current, the multispectral imaging unit records the arc light trajectory, the inter-harmonic correlation analysis unit uses the Nuttall window function to process the signal and extract the wave head time ≤1μs, the thermal image feature modeling unit locates the arc spot coordinates, the federal feature fusion engine aggregates the meteorological radar data, and corrects the lightning strike point to tower T23 (error ±0.3m). The digital twin engine generates an inspection work order and pushes it to the operation and maintenance terminal.

[0041] The present invention provides a 10kV distribution network safe operation management system based on intelligent monitoring. The system makes judgments from the perspective of multimodal data and can simultaneously analyze the correlation between electrical quantities, equipment temperature and environmental parameters, so that the fault recognition rate determined by the system is significantly improved and the possibility of misjudgment is reduced. At the same time, the intelligent monitoring sub-terminal set at each configuration node speeds up feature extraction, optimizes data response speed, avoids data delays caused by centralized data processing in traditional systems, meets transient fault response requirements, and further improves system reaction speed.

[0042] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A 10kV distribution network safe operation management system based on intelligent monitoring, characterized in that: include: Several intelligent monitoring sub-terminals installed at each distribution network node and a cloud-based decision-making platform installed at the system terminal; The plurality of intelligent monitoring sub-terminals are respectively connected to the cloud decision-making platform for communication; The intelligent monitoring sub-terminal includes: a data acquisition module for collecting multimodal data on the distribution network nodes and an edge computing cluster for extracting features from the collected multimodal data and then transmitting it; The cloud-based decision-making platform includes: a global risk prediction module for performing risk prediction after receiving the feature vector transmitted by the edge computing cluster, and a dynamic protection strategy generation module for generating a corresponding protection adjustment plan based on the calculation result of the global risk prediction module; The global risk prediction module includes: a federated feature fusion engine configured to receive feature vectors transmitted by a plurality of the intelligent monitoring sub-terminals, generate a multi-dimensional joint representation vector by feature fusion, and then transmit the generated multi-dimensional joint representation vector; a federated learning aggregator, configured to receive the transmitted multidimensional joint representation vectors, aggregate the multidimensional joint representation vectors at regular intervals, and output a global risk prediction model; The risk assessment output layer is used to generate a three-dimensional risk situation map through the global risk prediction model and simultaneously output the risk level and confidence interval; The dynamic protection strategy generation module includes: A digital twin engine generates dynamic adjustment strategies based on the three-dimensional risk situation map; The blockchain consensus executor receives the generated dynamic adjustment strategy and uses an improved practical Byzantine fault-tolerant algorithm to generate dynamic adjustment strategy instructions. The dynamic adjustment strategy instructions must be verified by more than 2 / 3 of the nodes before they are executed. The strategy feedback optimizer dynamically optimizes strategy parameters based on execution results and simultaneously generates operation and maintenance recommendation strategies.

2. A 10kV distribution network safe operation management system based on intelligent monitoring according to claim 1, characterized in that: The multimodal data includes: Electrical parameters: three-phase current signal, three-phase voltage signal, fundamental component of three-phase voltage, fundamental component of three-phase current, effective value of harmonics, total harmonic distortion rate and phase coupling relationship between harmonics of the nodes of the distribution network; Thermal imaging parameters: visible light image of the device, infrared image of the device, maximum surface temperature of the device, temperature gradient, area ratio of hotspot area, and standard deviation of temperature distribution.

3. A 10kV distribution network safe operation management system based on intelligent monitoring according to claim 2, characterized in that: The edge computing cluster includes: a harmonic depth analysis unit for extracting harmonic distortion features and time series correlation of the multimodal data; and a thermal imaging feature modeling unit for identifying device hot spots based on image segmentation technology and quantifying temperature gradient distribution to extract thermal imaging features; The harmonic depth analysis unit includes: an optimized window function processing unit for suppressing spectrum leakage by using an optimized window function having a fourth-order cosine component; and an inter-harmonic correlation analysis unit for performing time series-based inter-harmonic correlation feature extraction.

4. A 10kV distribution network safe operation management system based on intelligent monitoring according to claim 3, characterized in that: The feature extraction of the harmonic depth analysis unit includes: The optimized window function processing unit uses the optimized window function to perform a windowed Fourier transform on the voltage signal to suppress spectrum leakage; The inter-harmonic correlation analysis unit calculates the odd harmonic skewness and the Pearson correlation coefficient matrix of adjacent harmonics, extracts the distribution characteristics of harmonic energy in the 30-1500 Hz frequency band, and then calculates and obtains the harmonic distortion characteristics.

5. The 10kV distribution network safe operation management system based on intelligent monitoring according to claim 3 is characterized in that: The feature extraction of the thermal imaging feature modeling unit includes: The thermal imaging feature modeling unit identifies key areas of the device based on an image segmentation algorithm, delineates a temperature analysis hot temperature area ROI, calculates the maximum temperature gradient and radial attenuation coefficient within the ROI, and counts the proportion of pixels exceeding a preset temperature threshold and the spatial distribution discreteness to thereby calculate the thermal imaging feature.

6. A 10kV distribution network safe operation management system based on intelligent monitoring according to claim 1, characterized in that: The intelligent monitoring sub-terminal also includes: a gas detection module for detecting changes in SF6 concentration and triggering a gas insulation degradation warning, a temperature and humidity sensing module for detecting temperature and humidity, and an environmental compensation module for dynamically calibrating electrical quantity measurements based on the detection data of the humidity sensing module.

Images