Intelligent monitoring method and system for group distribution cabinets based on highly integrated power modules
By building a CFD model of the distribution cabinet and using cGAN, isolated forest algorithm and graph neural network, the prediction of the internal temperature field and component life prediction of the distribution cabinet are solved, and the problem of insufficient monitoring capabilities of group equipment in the existing technology is solved. Improve the intelligence and thermal runaway prevention capabilities of the distribution cabinet.
Patent Information
- Application Number
- CN202411317850.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-20
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-20
AI Technical Summary
The existing intelligent distribution cabinet monitoring system has limited capabilities in monitoring and fault diagnosis of grouped equipment, and cannot effectively predict the health status of equipment and prevent thermal runaway, making it difficult to meet the needs of modern power systems.
By building the CFD model of the distribution cabinet for temperature simulation, using cGAN to build a temperature field prediction model, combining the isolated forest algorithm and graph neural network, the prediction of the internal temperature field of the distribution cabinet, local hot spot identification, thermal runaway evaluation and component life are realized. Forecast, formulate regulation strategies and maintenance warnings.
It improves the intelligence and comprehensiveness of monitoring of group distribution cabinets, can effectively identify the risk of thermal runaway and abnormal operating conditions, predict the remaining life of the components, provide timely maintenance warnings, and improves the health management and safety of the equipment.
Smart Images

Figure CN118842190B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent monitoring of group distribution cabinets, and in particular to an intelligent monitoring method and system for group distribution cabinets based on a highly integrated power module. Background Art
[0002] Due to the rapid development of IT technology and the increasing popularity of IT applications, especially the emergence of cloud computing and virtualization technology, the market has a large demand for highly integrated grouped low-voltage distribution cabinets. Grouped low-voltage distribution cabinets can be installed in various data centers and communication centers, or even any place with high-reliability power supply. With the development of power systems and the advancement of modernization, the intelligent management of distribution equipment has become an important part of power system operation and maintenance.
[0003] However, the current intelligent distribution cabinet monitoring system still has some shortcomings. Traditional monitoring methods mainly focus on the status monitoring of a single device, while the overall status monitoring and fault diagnosis capabilities of the distribution cabinet group equipment are limited and cannot fully meet the needs of modern power systems. Secondly, with the high degree of integration of distribution equipment, the thermal management problem inside the equipment has become more prominent. High-temperature operation may accelerate the aging of the equipment and even cause serious safety accidents. The existing monitoring system lacks effective means for temperature monitoring and early warning, making it difficult to achieve a comprehensive assessment and prediction of the health status of the equipment. Therefore, how to improve the intelligence and comprehensiveness of monitoring of group distribution cabinets is an urgent problem to be solved. Summary of the invention
[0004] The present invention overcomes the defects of the prior art and provides a method and system for intelligent monitoring of group distribution cabinets based on highly integrated power modules, an important purpose of which is to improve the intelligence and comprehensiveness of monitoring of group distribution cabinets.
[0005] To achieve the above-mentioned purpose, the first aspect of the present invention provides an intelligent monitoring method for group distribution cabinets based on a highly integrated power module, comprising:
[0006] Construct a CFD model of a power distribution cabinet, preset several operating conditions of a group of power distribution cabinets, simulate the temperature inside the cabinet and form a simulation data set of the temperature inside the cabinet, construct a temperature field prediction model through cGAN and use the simulation data set of the temperature inside the cabinet for model training;
[0007] Acquire real-time operation monitoring information, acquire operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, use the temperature field prediction model to predict the temperature field inside the power distribution cabinet, identify local hot spots according to the prediction results, and obtain local hot spot identification information;
[0008] Determine the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluate whether the hot spot area will have thermal runaway, and formulate a control strategy to prevent thermal runaway;
[0009] Acquire the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, use the isolation forest algorithm to detect abnormal operation data of the internal components of the distribution cabinet, and analyze the abnormal operation status to obtain abnormal operation status analysis information;
[0010] A component life prediction model is constructed using graph neural networks to obtain the operating environment monitoring information of the target group distribution cabinet. The remaining life of the internal components of the target group distribution cabinet is predicted based on the abnormal operating status analysis information, and maintenance warnings are issued.
[0011] In this solution, the construction of the distribution cabinet CFD model, preset several group distribution cabinet operating conditions, simulate the temperature inside the cabinet and form a cabinet temperature simulation data set, build a temperature field prediction model through cGAN and use the cabinet temperature simulation data set for model training, specifically including:
[0012] Obtain the distribution cabinet structure information of the target group distribution cabinet, use the three-dimensional modeling technology to construct a geometric model of the target group distribution cabinet according to the distribution cabinet structure information, and use data retrieval to obtain and define the material properties of the target group distribution cabinet;
[0013] Based on the structure information of the distribution cabinet, the structural characteristics of the target group distribution cabinet are extracted, including the structural characteristics of the components and the heat dissipation structure characteristics, and the structural characteristic information of the distribution cabinet is obtained. The heating components are analyzed in combination with the equipment specification of the target group distribution cabinet, and the heat source definition and heat source type definition are performed to obtain the heat source analysis information in the cabinet;
[0014] Using CFD software combined with the geometric model of the target group distribution cabinet to build a CFD model of the distribution cabinet, and meshing the CFD model of the distribution cabinet, and defining the heat source area in the CFD model of the distribution cabinet according to the heat source analysis information in the cabinet;
[0015] Based on the structural characteristic information of the power distribution cabinet, the heat dissipation structure characteristics are obtained to set the fluid boundary conditions, including natural convection boundary conditions and forced convection boundary conditions, and the radiation boundary conditions are set according to the heat source analysis information and the heat source area in the cabinet;
[0016] Preset several group distribution cabinet operating conditions, set the global initial temperature, use the k-ε turbulence model and the steady-state solver to simulate the cabinet temperature of the target group distribution cabinet under different operating conditions, and form a cabinet temperature simulation data set;
[0017] A temperature field prediction model is constructed based on cGAN, and a training data set is constructed according to the cabinet temperature simulation data set. The operating condition attributes corresponding to each cabinet temperature simulation data are used as conditional constraints to train the temperature field prediction model and obtain a temperature field prediction model that meets the expectations.
[0018] In this solution, the operation condition characteristics and temperature monitoring characteristics are obtained based on the real-time operation monitoring information, the temperature field prediction model is used to predict the temperature field inside the distribution cabinet, and local hotspots are identified according to the prediction results, specifically including:
[0019] Monitor the target group distribution cabinet to obtain real-time operation monitoring information, perform data preprocessing on the real-time operation monitoring information, and input the preprocessed real-time operation monitoring information into the temperature field prediction model for analysis;
[0020] Extract features from the input real-time operation monitoring data, obtain the operating condition features and temperature monitoring features of the target group distribution cabinets, and form constraint conditions based on the operating condition features;
[0021] The generator generates the initial temperature field prediction result according to the temperature monitoring characteristics, inputs the generated initial temperature field prediction result into the discriminator for judgment, and determines whether to accept the current prediction result in combination with the constraint conditions;
[0022] If the judgment result of the discriminator is acceptance, the corresponding temperature field prediction result is output; if the judgment result of the discriminator is rejection, the generator continues to generate the next prediction result until the constraint condition is met and the final temperature field prediction result is output to obtain temperature field prediction information;
[0023] Generate a temperature distribution gradient map of the target grouped power distribution cabinet according to the temperature field prediction information, divide the internal area of the target grouped power distribution cabinet into a heat source area and a non-heat source area, and set a normal operating temperature range;
[0024] Local hotspot identification is performed through the temperature distribution gradient map according to the normal operating temperature range corresponding to each area to obtain local hotspot identification information.
[0025] In this solution, the hot spot area inside the target power distribution cabinet is determined according to the local hot spot identification information, whether the hot spot area will have thermal runaway is evaluated, and a control strategy is formulated to prevent thermal runaway, which specifically includes:
[0026] Acquire local hotspot identification information, extract hotspot area position characteristics and hotspot area temperature characteristics according to the local hotspot identification information, and obtain hotspot area characteristic information;
[0027] Acquire the temperature tolerance of the components in the hot spot area according to the characteristic information of the hot spot area, make a judgment based on the temperature characteristics corresponding to the hot spot area, and analyze whether the temperature of the target hot spot area exceeds the temperature tolerance;
[0028] If the temperature tolerance of the regional components is within the limit, the corresponding hot spot area is a temperature controllable area; if the temperature tolerance of the regional components is exceeded, thermal runaway analysis is performed on the target hot spot area;
[0029] Extracting temperature characteristics of the target hot spot area according to the characteristic information of the hot spot area, calculating the temperature change rate of the target hot spot area through the extracted temperature characteristics, analyzing the temperature change of the target hot spot area in the future time period based on the temperature change rate of the target area, and generating a temperature change trend graph;
[0030] Constructing a thermal runaway assessment model, extracting the temperature prediction values corresponding to the target hot spot area at each future time node according to the temperature change area map, inputting them into the thermal runaway assessment model for analysis, and obtaining thermal runaway assessment information;
[0031] The time node of the occurrence of thermal runaway is extracted according to the thermal runaway assessment information, the time deviation from the actual time is calculated, and a judgment is made with a preset threshold, and a corresponding control strategy is obtained according to the judgment result to prevent thermal runaway.
[0032] In this solution, the operation characteristics of the internal components of the target group distribution cabinet are obtained according to the real-time operation monitoring information, the abnormal operation data of the internal components of the distribution cabinet are detected by using the isolation forest algorithm, and the abnormal operation status is analyzed to obtain the abnormal operation status analysis information, which specifically includes:
[0033] Acquire real-time operation monitoring information of the target group distribution cabinet, extract internal component operation characteristics of the target group distribution cabinet during daily operation according to the real-time operation monitoring information, and obtain internal component operation characteristic information;
[0034] An anomaly detection model is constructed based on the isolation forest algorithm, the operating characteristic information of the internal components is processed in a time series manner to form a time series sequence, and the time series sequence is input into the anomaly detection model for abnormal operation detection;
[0035] The RANSAC algorithm is introduced to randomly select a number of sample data according to the constructed time series, perform linear fitting through the selected sample data, and calculate the difference between the actual observed value and the fitted predicted value according to the linear fitting to generate residual features;
[0036] Extract features from the constructed time series, build a new feature matrix based on the residual features, determine the total number of trees in the isolated forest and build a feature space based on the new feature matrix;
[0037] Construct a random binary search tree to detect low-density and isolated data points in the feature space, which are defined as isolated points. Calculate the path length of each isolated point in the corresponding tree as the abnormal score, and compare it with the preset threshold to obtain abnormal operation detection information;
[0038] Performing feature extraction on the abnormal operation detection information, extracting abnormal operation features of internal components of the target grouped power distribution cabinet, and obtaining a normal operation parameter benchmark based on the extracted abnormal operation features;
[0039] Calculate the deviation between the abnormal operation characteristics and the normal operation parameter benchmark, perform abnormal operation status analysis and analyze the periodicity of abnormal operation phenomena to obtain abnormal operation status analysis information.
[0040] In this solution, the component life prediction model is constructed using a graph neural network to obtain the operating environment monitoring information of the target group distribution cabinet, and the remaining life of the internal components of the target group distribution cabinet is predicted in combination with the abnormal operating status analysis information, and maintenance warning is performed, specifically including:
[0041] Based on big data retrieval, historical evaluation instances of distribution cabinet components with different life spans are obtained, life evaluation features are extracted for each historical evaluation instance, operating status features and operating environment features of the distribution cabinet components corresponding to each historical evaluation instance are obtained, and life evaluation feature information is obtained;
[0042] Based on the component type, life assessment characteristics and life degree, entity triples are formed, a life assessment knowledge graph is formed according to the entity triples, the life assessment knowledge graph is represented by a graph, the graph representation of the life assessment knowledge graph is learned based on a graph neural network, and a component life prediction model is constructed;
[0043] Acquire the operating environment monitoring information and abnormal operating condition analysis information of the target group distribution cabinet, and input them into the component life prediction model to perform life assessment on the internal components of the target group distribution cabinet;
[0044] Generate a target node and a node feature matrix of the target node according to the input data, perform first-order sampling on the target node to obtain neighbor nodes and form an adjacency matrix, obtain neighbors of the neighbor nodes through the adjacency matrix, use the attention mechanism to obtain attention coefficients of neighbors of the neighbor nodes, aggregate the adjacency matrix, and update the feature vectors of the neighbor nodes;
[0045] The updated adjacency matrix is obtained based on the feature vectors of the updated neighbor nodes, the attention coefficients of the neighbor nodes are obtained through the shared attention parameters, the node feature matrix of the target node is updated, and the final node feature matrix of the target node is obtained by stacking multiple layers of graph convolutional layers;
[0046] The life prediction results of the internal components of the target grouped distribution cabinet are generated according to the final node feature matrix of the target node to obtain the life prediction result information; the life prediction result information is judged with the preset threshold to analyze whether the internal components of the target grouped distribution cabinet need to be repaired, and a maintenance warning prompt is issued.
[0047] A second aspect of the present invention provides an intelligent monitoring system for a group distribution cabinet based on a highly integrated power module, the system comprising: a memory, a processor, the memory containing an intelligent monitoring method program for a group distribution cabinet based on a highly integrated power module, the intelligent monitoring method program for a group distribution cabinet based on a highly integrated power module being executed by the processor to implement the following steps:
[0048] Construct a CFD model of a power distribution cabinet, preset several operating conditions of a group of power distribution cabinets, simulate the temperature inside the cabinet and form a simulation data set of the temperature inside the cabinet, construct a temperature field prediction model through cGAN and use the simulation data set of the temperature inside the cabinet for model training;
[0049] Acquire real-time operation monitoring information, acquire operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, use the temperature field prediction model to predict the temperature field inside the power distribution cabinet, identify local hot spots according to the prediction results, and obtain local hot spot identification information;
[0050] Determine the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluate whether the hot spot area will have thermal runaway, and formulate a control strategy to prevent thermal runaway;
[0051] Acquire the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, use the isolation forest algorithm to detect abnormal operation data of the internal components of the distribution cabinet, and analyze the abnormal operation status to obtain abnormal operation status analysis information;
[0052] A component life prediction model is constructed using graph neural networks to obtain the operating environment monitoring information of the target group distribution cabinet. The remaining life of the internal components of the target group distribution cabinet is predicted based on the abnormal operating status analysis information, and maintenance warnings are issued.
[0053] The present invention discloses an intelligent monitoring method and system for grouped distribution cabinets based on highly integrated power modules, including: constructing a CFD simulation model of a target grouped distribution cabinet to simulate and analyze the temperature inside the cabinet under different working conditions; constructing a temperature field prediction model through cGAN and using the simulation data of the temperature inside the cabinet to train the model; based on the operating condition characteristics and temperature monitoring characteristics of the target grouped distribution cabinet, using the temperature field prediction model to predict the temperature inside the cabinet and identify the existing local hot spots; determining the hot spot area inside the target distribution cabinet according to the local hot spot information, evaluating whether the hot spot area will have a thermal runaway condition, and formulating a control strategy to prevent thermal runaway. Obtaining the operating data of the components inside the distribution cabinet, detecting and analyzing the abnormal operating conditions of the components inside the distribution cabinet; predicting the remaining life of the components in combination with the abnormal operating conditions and the operating environment, and performing maintenance warnings to improve the intelligence and comprehensiveness of monitoring the grouped distribution cabinet. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.
[0055] Figure 1 A flow chart of an intelligent monitoring method for group distribution cabinets based on highly integrated power modules provided by one embodiment of the present invention;
[0056] Figure 2 A flow chart of intelligent monitoring of target group distribution cabinets provided by an embodiment of the present invention;
[0057] Figure 3 A block diagram of an intelligent monitoring system for group distribution cabinets based on highly integrated power modules provided by one embodiment of the present invention;
[0058] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0059] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.
[0061] Figure 1 A flow chart of an intelligent monitoring method for group distribution cabinets based on highly integrated power modules provided by one embodiment of the present invention;
[0062] like Figure 1 As shown, the present invention provides a flow chart of an intelligent monitoring method for group distribution cabinets based on a highly integrated power module, including:
[0063] S102, constructing a CFD model of a power distribution cabinet, presetting a number of operating conditions of a group of power distribution cabinets, simulating the temperature inside the cabinet and forming a simulation data set of the temperature inside the cabinet, constructing a temperature field prediction model through cGAN and using the simulation data set of the temperature inside the cabinet for model training;
[0064] S104, acquiring real-time operation monitoring information, acquiring operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, predicting the temperature field inside the power distribution cabinet using the temperature field prediction model, identifying local hot spots according to the prediction results, and obtaining local hot spot identification information;
[0065] S106, determining the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluating whether the hot spot area will have a thermal runaway condition, and formulating a control strategy to prevent thermal runaway;
[0066] S108, obtaining the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, detecting abnormal operation data of the internal components of the distribution cabinet using an isolation forest algorithm, and analyzing the abnormal operation status to obtain abnormal operation status analysis information;
[0067] S110, using graph neural network to build a component life prediction model, obtain the operating environment monitoring information of the target group distribution cabinet, combine the abnormal operating status analysis information to predict the remaining life of the internal components of the target group distribution cabinet, and issue a maintenance warning.
[0068] It should be noted that the present invention provides an intelligent monitoring method and system for grouped distribution cabinets based on highly integrated power modules. Based on the distribution cabinet structure information of the target grouped distribution cabinet, a CFD simulation model of the target grouped distribution cabinet is constructed to simulate and analyze the temperature inside the cabinet under different working conditions, and form a simulation data set of the temperature inside the cabinet; a temperature field prediction model is constructed by cGAN and the cabinet temperature simulation data set is used for model training; based on the operating condition characteristics and temperature monitoring characteristics of the target grouped distribution cabinet, the temperature inside the cabinet is predicted using the temperature field prediction model, and the existing local hot spots are identified; the hot spot area inside the target distribution cabinet is determined according to the local hot spot information, and whether the hot spot area will have a thermal runaway condition is evaluated, and a control strategy is formulated to prevent thermal runaway. According to the real-time operation monitoring information, the operating characteristics of the components inside the target grouped distribution cabinet are obtained, and the abnormal operation data of the components inside the distribution cabinet are detected using the isolation forest algorithm, and the abnormal operation status is analyzed; the component life prediction model is constructed using a graph neural network, and the remaining life of the component is predicted in combination with the abnormal operation status and the operation environment, and maintenance warnings are performed to improve the intelligence and comprehensiveness of monitoring the grouped distribution cabinet.
[0069] Furthermore, in a preferred embodiment of the present invention, the construction of the power distribution cabinet CFD model, presetting a number of grouped power distribution cabinet operating conditions, performing cabinet temperature simulation and forming a cabinet temperature simulation data set, constructing a temperature field prediction model through cGAN and using the cabinet temperature simulation data set for model training, specifically includes:
[0070] Obtain the distribution cabinet structure information of the target group distribution cabinet, use the three-dimensional modeling technology to construct a geometric model of the target group distribution cabinet according to the distribution cabinet structure information, and use data retrieval to obtain and define the material properties of the target group distribution cabinet;
[0071] Based on the structure information of the distribution cabinet, the structural characteristics of the target group distribution cabinet are extracted, including the structural characteristics of the components and the heat dissipation structure characteristics, and the structural characteristic information of the distribution cabinet is obtained. The heating components are analyzed in combination with the equipment specification of the target group distribution cabinet, and the heat source definition and heat source type definition are performed to obtain the heat source analysis information in the cabinet;
[0072] Using CFD software combined with the geometric model of the target group distribution cabinet to build a CFD model of the distribution cabinet, and meshing the CFD model of the distribution cabinet, and defining the heat source area in the CFD model of the distribution cabinet according to the heat source analysis information in the cabinet;
[0073] Based on the structural characteristic information of the power distribution cabinet, the heat dissipation structure characteristics are obtained to set the fluid boundary conditions, including natural convection boundary conditions and forced convection boundary conditions, and the radiation boundary conditions are set according to the heat source analysis information and the heat source area in the cabinet;
[0074] Preset several group distribution cabinet operating conditions, set the global initial temperature, use the k-ε turbulence model and the steady-state solver to simulate the cabinet temperature of the target group distribution cabinet under different operating conditions, and form a cabinet temperature simulation data set;
[0075] A temperature field prediction model is constructed based on cGAN, and a training data set is constructed according to the cabinet temperature simulation data set. The operating condition attributes corresponding to each cabinet temperature simulation data are used as conditional constraints to train the temperature field prediction model and obtain a temperature field prediction model that meets the expectations.
[0076] It should be noted that by obtaining the structural information of the target group distribution cabinet, the corresponding geometric model is constructed using the three-dimensional modeling technology. The material properties of the distribution cabinet are obtained and defined based on the data retrieval technology to ensure that the geometric model is consistent with the physical characteristics of the actual equipment. Next, the key heat sources in the distribution cabinet are identified by the structural characteristics of the distribution cabinet, and the heat sources are defined and classified in combination with the equipment specification, such as body heat source, point heat source, etc. Further understand the internal thermal environment of the distribution cabinet and provide a basis for subsequent heat source modeling. Subsequently, the CFD model of the distribution cabinet is created based on the constructed three-dimensional geometric model using CFD software. The model is meshed to improve the accuracy and efficiency of the simulation. At the same time, the corresponding heat source area is defined in the CFD model. Subsequently, based on the structural feature information of the distribution cabinet, the fluid boundary conditions are set, including natural convection boundary conditions and forced convection boundary conditions, and the radiation boundary conditions are defined in combination with the heat source analysis information in the cabinet. In order to more comprehensively simulate the heat transfer process inside the distribution cabinet. Then, the simulation is performed based on the working conditions, which are the distribution cabinet environment and the distribution cabinet operation mode and status. The k-ε turbulence model and steady-state solver are used to simulate the temperature inside the cabinet under each working condition to form a cabinet temperature simulation data set. Finally, in order to achieve rapid prediction of the temperature field of the distribution cabinet, a temperature field prediction model is constructed based on the conditional generative adversarial network (cGAN). The previously constructed cabinet temperature simulation data set is used for training, where the operating conditions corresponding to each simulation data are used as conditional constraints. The Adam optimizer is used for gradient optimization, and the network weights are initialized to 0. The learning rate decay strategy is used to make the learning rate value linearly decrease as the training deepens, so as to obtain a model that meets the expectations.
[0077] Further, in a preferred embodiment of the present invention, the operation condition characteristics and temperature monitoring characteristics are obtained based on the real-time operation monitoring information, the temperature field prediction model is used to predict the temperature field inside the distribution cabinet, and local hotspots are identified according to the prediction results, specifically including:
[0078] Monitor the target group distribution cabinet to obtain real-time operation monitoring information, perform data preprocessing on the real-time operation monitoring information, and input the preprocessed real-time operation monitoring information into the temperature field prediction model for analysis;
[0079] Extract features from the input real-time operation monitoring data, obtain the operating condition features and temperature monitoring features of the target group distribution cabinets, and form constraint conditions based on the operating condition features;
[0080] The generator generates the initial temperature field prediction result according to the temperature monitoring characteristics, inputs the generated initial temperature field prediction result into the discriminator for judgment, and determines whether to accept the current prediction result in combination with the constraint conditions;
[0081] If the judgment result of the discriminator is acceptance, the corresponding temperature field prediction result is output; if the judgment result of the discriminator is rejection, the generator continues to generate the next prediction result until the constraint condition is met and the final temperature field prediction result is output to obtain temperature field prediction information;
[0082] Generate a temperature distribution gradient map of the target grouped power distribution cabinet according to the temperature field prediction information, divide the internal area of the target grouped power distribution cabinet into a heat source area and a non-heat source area, and set a normal operating temperature range;
[0083] Local hotspot identification is performed through the temperature distribution gradient map according to the normal operating temperature range corresponding to each area to obtain local hotspot identification information.
[0084] It should be noted that, first, by monitoring the target group distribution cabinet, real-time operation monitoring information is obtained, and the temperature field prediction model is used to predict the temperature field. After the data is input into the model, the model extracts features from the input data, obtains features related to the operating conditions of the distribution cabinet and temperature monitoring features, and based on the extracted operating condition features, constitutes constraints in the prediction process, so that the prediction results of the model are more in line with the actual operating conditions. In the temperature field prediction process, the generator generates an initial temperature field prediction result based on the temperature monitoring features. The initial prediction result is then input into the discriminator for judgment, and the discriminator judges whether to accept the currently generated prediction result in combination with the set constraints. If the judgment result of the discriminator is acceptance, the temperature field prediction result is directly output as the final temperature field prediction information. If the judgment result of the discriminator is not accepted, the generator will continue to generate new prediction results until a prediction result that meets the constraints is generated, ensuring the accuracy and reliability of the prediction results. After obtaining the temperature field prediction information that meets the requirements, a temperature distribution gradient map of the target group distribution cabinet is generated based on the prediction result. The temperature distribution gradient map is used to divide the internal area of the power distribution cabinet into heat source area and non-heat source area. Next, by analyzing the temperature distribution gradient map, local hotspot areas where the temperature exceeds the normal operating range are identified, and local hotspot identification information is obtained to timely warn of potential failure risks and guide subsequent maintenance and management work.
[0085] Further, in a preferred embodiment of the present invention, the hot spot area inside the target power distribution cabinet is determined according to the local hot spot identification information, whether the hot spot area will have a thermal runaway condition is evaluated, and a control strategy is formulated to prevent thermal runaway, specifically including:
[0086] Acquire local hotspot identification information, extract hotspot area position characteristics and hotspot area temperature characteristics according to the local hotspot identification information, and obtain hotspot area characteristic information;
[0087] Acquire the temperature tolerance of the components in the hot spot area according to the characteristic information of the hot spot area, make a judgment based on the temperature characteristics corresponding to the hot spot area, and analyze whether the temperature of the target hot spot area exceeds the temperature tolerance;
[0088] If the temperature tolerance of the regional components is within the limit, the corresponding hot spot area is a temperature controllable area; if the temperature tolerance of the regional components is exceeded, thermal runaway analysis is performed on the target hot spot area;
[0089] Extracting temperature characteristics of the target hot spot area according to the characteristic information of the hot spot area, calculating the temperature change rate of the target hot spot area through the extracted temperature characteristics, analyzing the temperature change of the target hot spot area in the future time period based on the temperature change rate of the target area, and generating a temperature change trend graph;
[0090] Constructing a thermal runaway assessment model, extracting the temperature prediction values corresponding to the target hot spot area at each future time node according to the temperature change area map, inputting them into the thermal runaway assessment model for analysis, and obtaining thermal runaway assessment information;
[0091] The time node of the occurrence of thermal runaway is extracted according to the thermal runaway assessment information, the time deviation from the actual time is calculated, and a judgment is made with a preset threshold, and a corresponding control strategy is obtained according to the judgment result to prevent thermal runaway.
[0092] It should be noted that after predictive analysis by the temperature field prediction model, the local hot spot area in the target group distribution cabinet is identified, and the temperature tolerance of each component in the hot spot area is queried based on the local hot spot area. By comparing with the current temperature characteristics of the hot spot area, it is evaluated whether the temperature of the target hot spot area exceeds the tolerance range of the components. If the temperature does not exceed, the hot spot area is considered to be a temperature controllable area, and no additional heat dissipation measures are required for the time being. On the contrary, if the temperature exceeds, the system will conduct an in-depth thermal runaway analysis on the target hot spot area. First, the temperature characteristics of the target hot spot area are extracted, and the temperature change rate is calculated based on this, which reflects the speed of temperature change in the hot spot area and is an important basis for predicting future temperature changes. A temperature change trend graph is generated based on the temperature change rate to intuitively display the temperature change of the target hot spot area in the future. A thermal runaway assessment model is further constructed, and the temperature prediction values of each future time node extracted from the temperature change trend graph are used as input. After analysis, the thermal runaway assessment information is output to assess the possibility, severity and possible time nodes of thermal runaway. Among them, for the thermal runaway assessment model, a convolutional neural network (CNN) or a recurrent neural network (RNN) can be used to construct it. By obtaining historical thermal runaway instances to build a training data set, the model is deeply learned and trained, and cross-validation or grid search is used to optimize the model's hyperparameters, so as to obtain a thermal runaway assessment model that meets the expectations. Finally, based on the thermal runaway assessment information, especially the time node when the thermal runaway occurs, the deviation between this time point and the actual time is calculated. By comparing this deviation with the preset threshold, it is determined whether there is sufficient time for normal regulation before the thermal runaway event occurs, such as increasing the operating power of the heat dissipation equipment, etc. If there is no sufficient time for regulation, the target distribution cabinet needs to be stopped until the temperature drops to a safe range and then restarted to ensure the safe and stable operation of the distribution cabinet.
[0093] Further, in a preferred embodiment of the present invention, the operation characteristics of the internal components of the target group distribution cabinet are obtained according to the real-time operation monitoring information, the abnormal operation data of the internal components of the distribution cabinet are detected by using the isolation forest algorithm, and the abnormal operation status is analyzed to obtain the abnormal operation status analysis information, which specifically includes:
[0094] Acquire real-time operation monitoring information of the target group distribution cabinet, extract internal component operation characteristics of the target group distribution cabinet during daily operation according to the real-time operation monitoring information, and obtain internal component operation characteristic information;
[0095] An anomaly detection model is constructed based on the isolation forest algorithm, the operating characteristic information of the internal components is processed in a time series manner to form a time series sequence, and the time series sequence is input into the anomaly detection model for abnormal operation detection;
[0096] The RANSAC algorithm is introduced to randomly select a number of sample data according to the constructed time series, perform linear fitting through the selected sample data, and calculate the difference between the actual observed value and the fitted predicted value according to the linear fitting to generate residual features;
[0097] Extract features from the constructed time series, build a new feature matrix based on the residual features, determine the total number of trees in the isolated forest and build a feature space based on the new feature matrix;
[0098] Construct a random binary search tree to detect low-density and isolated data points in the feature space, which are defined as isolated points. Calculate the path length of each isolated point in the corresponding tree as the abnormal score, and compare it with the preset threshold to obtain abnormal operation detection information;
[0099] Performing feature extraction on the abnormal operation detection information, extracting abnormal operation features of internal components of the target grouped power distribution cabinet, and obtaining a normal operation parameter benchmark based on the extracted abnormal operation features;
[0100] Calculate the deviation between the abnormal operation characteristics and the normal operation parameter benchmark, perform abnormal operation status analysis and analyze the periodicity of abnormal operation phenomena to obtain abnormal operation status analysis information.
[0101] It should be noted that during the operation of the group distribution cabinet, due to the highly integrated characteristics, the components in the cabinet often work in a high temperature environment. When the ambient temperature fluctuates, the operating temperature of the components in the cabinet will also fluctuate. Even under the control of the heat dissipation equipment, the component life consumption will still be accelerated. In this scenario, the operating life prediction of the components in the cabinet can be evaluated from the operating environment and operating abnormalities. By detecting the daily operating data of each component in the cabinet, the isolated forest algorithm is used to detect abnormal data in the daily operating data, and the abnormal conditions when abnormal data appear are further analyzed, so as to judge the operating status of the target distribution cabinet components. At the same time, the periodicity of the abnormality is considered, that is, the frequency and period of abnormal operating data, to provide a data basis for subsequent life prediction.
[0102] Furthermore, in a preferred embodiment of the present invention, the component life prediction model is constructed using a graph neural network, the operating environment monitoring information of the target group distribution cabinet is obtained, and the remaining life of the internal components of the target group distribution cabinet is predicted in combination with the abnormal operating status analysis information, and maintenance warning is performed, specifically including:
[0103] Based on big data retrieval, historical evaluation instances of distribution cabinet components with different life spans are obtained, life evaluation features are extracted for each historical evaluation instance, operating status features and operating environment features of the distribution cabinet components corresponding to each historical evaluation instance are obtained, and life evaluation feature information is obtained;
[0104] Based on the component type, life assessment characteristics and life degree, entity triples are formed, a life assessment knowledge graph is formed according to the entity triples, the life assessment knowledge graph is represented by a graph, the graph representation of the life assessment knowledge graph is learned based on a graph neural network, and a component life prediction model is constructed;
[0105] Acquire the operating environment monitoring information and abnormal operating condition analysis information of the target group distribution cabinet, and input them into the component life prediction model to perform life assessment on the internal components of the target group distribution cabinet;
[0106] Generate a target node and a node feature matrix of the target node according to the input data, perform first-order sampling on the target node to obtain neighbor nodes and form an adjacency matrix, obtain neighbors of the neighbor nodes through the adjacency matrix, use the attention mechanism to obtain attention coefficients of neighbors of the neighbor nodes, aggregate the adjacency matrix, and update the feature vectors of the neighbor nodes;
[0107] The updated adjacency matrix is obtained based on the feature vectors of the updated neighbor nodes, the attention coefficients of the neighbor nodes are obtained through the shared attention parameters, the node feature matrix of the target node is updated, and the final node feature matrix of the target node is obtained by stacking multiple layers of graph convolutional layers;
[0108] The life prediction results of the internal components of the target grouped distribution cabinet are generated according to the final node feature matrix of the target node to obtain the life prediction result information; the life prediction result information is judged with the preset threshold to analyze whether the internal components of the target grouped distribution cabinet need to be repaired, and a maintenance warning prompt is issued.
[0109] It should be noted that when predicting the life of the internal components of the group distribution cabinet, the prediction is made from the operating environment and abnormal operation. Abnormal operation represents the performance of the internal components. When the life declines, the corresponding performance will also change, that is, more frequent abnormal operation data will appear. The operating environment is a key factor affecting the change of life. The components of the distribution cabinet under different operating environments have different lifespans. For example, the internal components that work in a high temperature environment for a long time have a lower service life than the internal components that work in a controllable temperature environment. The component life prediction model is constructed using a graph neural network, and the component life prediction model is trained by constructing a life assessment knowledge graph consisting of entity triples of component type, life assessment features and life degree. The operating environment monitoring information and abnormal operation status analysis information are input into the component life prediction model. First, the target node and its feature matrix are generated, and the neighbor nodes are obtained through first-order sampling to construct an adjacency matrix. Subsequently, the attention mechanism is used to calculate the attention coefficients of the neighbor nodes and their neighbors, and the adjacency matrix is aggregated to update the feature vectors of the neighbor nodes. Through the stacking of multiple layers of graph convolutional layers, the deep feature information of the target node and its neighbor nodes is gradually extracted and integrated. Finally, according to the final node feature matrix of the target node, the life prediction results of the internal components of the target group distribution cabinet are generated and compared with the preset threshold to determine whether each component needs maintenance, and a maintenance warning prompt is issued accordingly. This improves the comprehensiveness and intelligence of the monitoring of the internal components of the group distribution cabinet and ensures the safe operation of the distribution cabinet.
[0110] Figure 2 A flow chart of intelligent monitoring of target group distribution cabinets provided by an embodiment of the present invention;
[0111] like Figure 2 As shown, the present invention provides a target group distribution cabinet intelligent monitoring flow chart, including:
[0112] S202, monitoring the target group distribution cabinet to obtain real-time operation monitoring information, extracting operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, and inputting them into a temperature field prediction model to predict the temperature field in the cabinet;
[0113] S204, constructing a temperature distribution gradient map of the target grouped power distribution cabinets based on the temperature field results, identifying local hot spots, and conducting a thermal runaway assessment on the hot spot areas to determine whether a thermal runaway event is likely to occur, and conducting early warning and control;
[0114] S206, extracting operation characteristic data of internal components of the target group distribution cabinet according to the real-time operation monitoring information, inputting the data into an abnormality detection model to detect abnormal operation data, and analyzing the abnormal operation status to obtain abnormal operation status analysis information;
[0115] S208, obtaining the operating environment monitoring information of the target group distribution cabinet, performing internal component life prediction in combination with the abnormal operating condition analysis information, and obtaining life prediction result information through a component life prediction model;
[0116] S210, the life prediction result information is compared with a preset threshold value to analyze whether each component inside the target group distribution cabinet needs to be repaired, and a maintenance warning prompt is issued.
[0117] In addition, in the intelligent monitoring method of group distribution cabinets based on highly integrated power modules provided by the present invention, the following steps are also included:
[0118] Obtaining historical maintenance instances of components in different operating environments, performing feature extraction on each historical maintenance instance of the component, obtaining maintenance frequency features of each historical maintenance instance of the component, and obtaining first feature information;
[0119] Perform maintenance cycle analysis based on the first feature information, analyze the maintenance rules of each component under each operating environment, and obtain maintenance rule analysis information; construct maintenance cycle portraits of each type of component under different operating environments based on the first feature information and the maintenance rule analysis information;
[0120] A component maintenance database is constructed based on the constructed maintenance cycle portrait, and the operating environment monitoring information and component maintenance record information of the target distribution cabinet are obtained. According to the operating environment monitoring information, the component maintenance database is searched to obtain the corresponding component maintenance portrait under the current operating environment;
[0121] According to the component maintenance profile under the current operating environment, combined with the component maintenance record, it is queried whether each component inside the target group distribution cabinet is in the maintenance cycle at the current moment, and the information of the components to be repaired is obtained;
[0122] The corresponding operating status monitoring information and operating environment monitoring information are obtained according to the information of the component to be repaired, abnormal operating status detection and component life prediction are performed, and whether the target component to be repaired needs maintenance is evaluated according to the component life prediction result, and a maintenance warning prompt is issued.
[0123] It should be noted that when performing intelligent maintenance inspection on the target group distribution cabinet, the detection cycle and frequency are usually preset. Although this method can know the life status of the internal components of the target group distribution cabinet in detail, frequent comprehensive inspection of the internal components of the distribution cabinet has the problem of high cumbersomeness and excessive resource consumption. The main factor affecting the life of the internal components of the distribution cabinet is the operating environment. Usually, the operating environment of the distribution cabinet is periodic. Under the periodic environment, the life change of the internal components of the distribution cabinet also has a periodic change law. By analyzing the maintenance rules of each component under different operating environments, the intelligent maintenance inspection is optimized. Through the maintenance rules of each component, the corresponding maintenance cycle portrait is constructed. When performing maintenance inspection, the similarity of the environment is used to further obtain the maintenance cycle portrait under the environment, so as to judge whether there are components that need maintenance at the current moment in combination with the maintenance record, so as to detect and analyze the corresponding components in the corresponding time period, reduce the cumbersomeness of each intelligent maintenance inspection, reduce resource consumption, and improve the accuracy of maintenance inspection.
[0124] Figure 3 An embodiment of the present invention provides an intelligent monitoring system 3 for a group distribution cabinet based on a highly integrated power module, the system comprising: a memory 31, a processor 32, the memory 31 containing an intelligent monitoring method program for a group distribution cabinet based on a highly integrated power module, the intelligent monitoring method program for a group distribution cabinet based on a highly integrated power module being executed by the processor 32 to implement the following steps:
[0125] Construct a CFD model of a power distribution cabinet, preset several operating conditions of a group of power distribution cabinets, simulate the temperature inside the cabinet and form a simulation data set of the temperature inside the cabinet, construct a temperature field prediction model through cGAN and use the simulation data set of the temperature inside the cabinet for model training;
[0126] Acquire real-time operation monitoring information, acquire operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, use the temperature field prediction model to predict the temperature field inside the power distribution cabinet, identify local hot spots according to the prediction results, and obtain local hot spot identification information;
[0127] Determine the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluate whether the hot spot area will have thermal runaway, and formulate a control strategy to prevent thermal runaway;
[0128] Acquire the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, use the isolation forest algorithm to detect abnormal operation data of the internal components of the distribution cabinet, and analyze the abnormal operation status to obtain abnormal operation status analysis information;
[0129] A component life prediction model is constructed using graph neural networks to obtain the operating environment monitoring information of the target group distribution cabinet. The remaining life of the internal components of the target group distribution cabinet is predicted based on the abnormal operating status analysis information, and maintenance warnings are issued.
[0130] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.
[0131] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0132] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.
[0133] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.
[0134] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.
[0135] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. An intelligent monitoring method for group distribution cabinets based on highly integrated power modules, characterized in that: include: Construct a CFD model of a power distribution cabinet, preset several operating conditions of a group of power distribution cabinets, simulate the temperature inside the cabinet and form a simulation data set of the temperature inside the cabinet, construct a temperature field prediction model through cGAN and use the simulation data set of the temperature inside the cabinet for model training; Acquire real-time operation monitoring information, acquire operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, use the temperature field prediction model to predict the temperature field inside the power distribution cabinet, identify local hot spots according to the prediction results, and obtain local hot spot identification information; Determine the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluate whether the hot spot area will have thermal runaway, and formulate a control strategy to prevent thermal runaway; Acquire the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, use the isolation forest algorithm to detect abnormal operation data of the internal components of the distribution cabinet, and analyze the abnormal operation status to obtain abnormal operation status analysis information; Use graph neural network to build component life prediction model, obtain the operating environment monitoring information of the target group distribution cabinet, combine the abnormal operating status analysis information to predict the remaining life of the internal components of the target group distribution cabinet, and carry out maintenance warning; The construction of the power distribution cabinet CFD model, presetting a number of grouped power distribution cabinet operating conditions, performing cabinet temperature simulation and forming a cabinet temperature simulation data set, constructing a temperature field prediction model through cGAN and using the cabinet temperature simulation data set for model training specifically includes: Obtain the distribution cabinet structure information of the target group distribution cabinet, use the three-dimensional modeling technology to construct a geometric model of the target group distribution cabinet according to the distribution cabinet structure information, and use data retrieval to obtain and define the material properties of the target group distribution cabinet; Based on the structure information of the distribution cabinet, the structural characteristics of the target group distribution cabinet are extracted, including the structural characteristics of the components and the heat dissipation structure characteristics, and the structural characteristic information of the distribution cabinet is obtained. The heating components are analyzed in combination with the equipment specification of the target group distribution cabinet, and the heat source definition and heat source type definition are performed to obtain the heat source analysis information in the cabinet; Using CFD software combined with the geometric model of the target group distribution cabinet to build a CFD model of the distribution cabinet, and meshing the CFD model of the distribution cabinet, and defining the heat source area in the CFD model of the distribution cabinet according to the heat source analysis information in the cabinet; Based on the structural characteristic information of the power distribution cabinet, the heat dissipation structure characteristics are obtained to set the fluid boundary conditions, including natural convection boundary conditions and forced convection boundary conditions, and the radiation boundary conditions are set according to the heat source analysis information and the heat source area in the cabinet; Preset several group distribution cabinet operating conditions, set the global initial temperature, use the k-ε turbulence model and the steady-state solver to simulate the cabinet temperature of the target group distribution cabinet under different operating conditions, and form a cabinet temperature simulation data set; A temperature field prediction model is constructed based on cGAN, a training data set is constructed according to the cabinet temperature simulation data set, and the operating condition attributes corresponding to each cabinet temperature simulation data are used as condition constraints to train the temperature field prediction model to obtain a temperature field prediction model that meets expectations; The method of acquiring operating condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, predicting the temperature field inside the power distribution cabinet using the temperature field prediction model, and identifying local hot spots according to the prediction results specifically includes: Monitor the target group distribution cabinet to obtain real-time operation monitoring information, perform data preprocessing on the real-time operation monitoring information, and input the preprocessed real-time operation monitoring information into the temperature field prediction model for analysis; Extract features from the input real-time operation monitoring data, obtain the operating condition features and temperature monitoring features of the target group distribution cabinets, and form constraint conditions based on the operating condition features; The generator generates the initial temperature field prediction result according to the temperature monitoring characteristics, inputs the generated initial temperature field prediction result into the discriminator for judgment, and determines whether to accept the current prediction result in combination with the constraint conditions; If the judgment result of the discriminator is acceptance, the corresponding temperature field prediction result is output; if the judgment result of the discriminator is rejection, the generator continues to generate the next prediction result until the constraint condition is met and the final temperature field prediction result is output to obtain temperature field prediction information; Generate a temperature distribution gradient map of the target grouped power distribution cabinet according to the temperature field prediction information, divide the internal area of the target grouped power distribution cabinet into a heat source area and a non-heat source area, and set a normal operating temperature range; Local hotspot identification is performed through the temperature distribution gradient map according to the normal operating temperature range corresponding to each area to obtain local hotspot identification information.
2. According to claim 1, a method for intelligent monitoring of group distribution cabinets based on highly integrated power modules is characterized in that: Determining the hotspot area inside the target power distribution cabinet according to the local hotspot identification information, evaluating whether the hotspot area will have a thermal runaway condition, and formulating a control strategy to prevent thermal runaway specifically includes: Acquire local hotspot identification information, extract hotspot area position characteristics and hotspot area temperature characteristics according to the local hotspot identification information, and obtain hotspot area characteristic information; Acquire the temperature tolerance of the components in the hot spot area according to the characteristic information of the hot spot area, make a judgment based on the temperature characteristics corresponding to the hot spot area, and analyze whether the temperature of the target hot spot area exceeds the temperature tolerance; If the temperature tolerance of the regional components is within the limit, the corresponding hot spot area is a temperature controllable area; if the temperature tolerance of the regional components is exceeded, thermal runaway analysis is performed on the target hot spot area; Extracting temperature characteristics of the target hot spot area according to the characteristic information of the hot spot area, calculating the temperature change rate of the target hot spot area through the extracted temperature characteristics, analyzing the temperature change of the target hot spot area in the future time period based on the temperature change rate of the target area, and generating a temperature change trend graph; Constructing a thermal runaway assessment model, extracting the temperature prediction values corresponding to the target hot spot area at each future time node according to the temperature change trend diagram, inputting the predicted values into the thermal runaway assessment model for analysis, and obtaining thermal runaway assessment information; The time node of the occurrence of thermal runaway is extracted according to the thermal runaway assessment information, the time deviation from the actual time is calculated, and a judgment is made with a preset threshold, and a corresponding control strategy is obtained according to the judgment result to prevent thermal runaway.
3. The intelligent monitoring method for group distribution cabinets based on highly integrated power modules according to claim 1 is characterized in that: The step of obtaining the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, detecting the abnormal operation data of the internal components of the distribution cabinet by using the isolation forest algorithm, and analyzing the abnormal operation status to obtain the abnormal operation status analysis information specifically includes: Acquire real-time operation monitoring information of the target group distribution cabinet, extract internal component operation characteristics of the target group distribution cabinet during daily operation according to the real-time operation monitoring information, and obtain internal component operation characteristic information; An anomaly detection model is constructed based on the isolation forest algorithm, the operating characteristic information of the internal components is processed in a time series manner to form a time series sequence, and the time series sequence is input into the anomaly detection model for abnormal operation detection; The RANSAC algorithm is introduced to randomly select a number of sample data according to the constructed time series, perform linear fitting through the selected sample data, and calculate the difference between the actual observed value and the fitted predicted value according to the linear fitting to generate residual features; Extract features from the constructed time series, build a new feature matrix based on the residual features, determine the total number of trees in the isolated forest and build a feature space based on the new feature matrix; Construct a random binary search tree to detect low-density and isolated data points in the feature space, which are defined as isolated points. Calculate the path length of each isolated point in the corresponding tree as the abnormal score, and compare it with the preset threshold to obtain abnormal operation detection information; Performing feature extraction on the abnormal operation detection information, extracting abnormal operation features of internal components of the target grouped power distribution cabinet, and obtaining a normal operation parameter benchmark based on the extracted abnormal operation features; Calculate the deviation between the abnormal operation characteristics and the normal operation parameter benchmark, perform abnormal operation status analysis and analyze the periodicity of abnormal operation phenomena to obtain abnormal operation status analysis information.
4. The intelligent monitoring method for group distribution cabinets based on highly integrated power modules according to claim 1 is characterized in that: The component life prediction model is constructed by using a graph neural network, the operating environment monitoring information of the target group distribution cabinet is obtained, and the remaining life of the internal components of the target group distribution cabinet is predicted in combination with the abnormal operating status analysis information, and maintenance warning is performed, specifically including: Based on big data retrieval, historical evaluation instances of distribution cabinet components with different life spans are obtained, life evaluation features are extracted for each historical evaluation instance, operating status features and operating environment features of the distribution cabinet components corresponding to each historical evaluation instance are obtained, and life evaluation feature information is obtained; Based on the component type, life assessment characteristics and life degree, entity triples are formed, a life assessment knowledge graph is formed according to the entity triples, the life assessment knowledge graph is represented by a graph, the graph representation of the life assessment knowledge graph is learned based on a graph neural network, and a component life prediction model is constructed; Acquire the operating environment monitoring information and abnormal operating condition analysis information of the target group distribution cabinet, and input them into the component life prediction model to perform life assessment on the internal components of the target group distribution cabinet; Generate a target node and a node feature matrix of the target node according to the input data, perform first-order sampling on the target node to obtain neighbor nodes and form an adjacency matrix, obtain neighbors of the neighbor nodes through the adjacency matrix, use the attention mechanism to obtain attention coefficients of neighbors of the neighbor nodes, aggregate the adjacency matrix, and update the feature vectors of the neighbor nodes; The updated adjacency matrix is obtained based on the feature vectors of the updated neighbor nodes, the attention coefficients of the neighbor nodes are obtained through the shared attention parameters, the node feature matrix of the target node is updated, and the final node feature matrix of the target node is obtained by stacking multiple layers of graph convolutional layers; The life prediction results of the internal components of the target grouped distribution cabinet are generated according to the final node feature matrix of the target node to obtain the life prediction result information; the life prediction result information is judged with the preset threshold to analyze whether the internal components of the target grouped distribution cabinet need to be repaired, and a maintenance warning prompt is issued.
5. An intelligent monitoring system for group distribution cabinets based on highly integrated power modules, characterized in that: The system includes: a memory and a processor, wherein the memory contains a program of an intelligent monitoring method for group distribution cabinets based on a highly integrated power module, and when the program of an intelligent monitoring method for group distribution cabinets based on a highly integrated power module is executed by the processor, the following steps are implemented: Construct a CFD model of a power distribution cabinet, preset several operating conditions of a group of power distribution cabinets, simulate the temperature inside the cabinet and form a simulation data set of the temperature inside the cabinet, construct a temperature field prediction model through cGAN and use the simulation data set of the temperature inside the cabinet for model training; Acquire real-time operation monitoring information, acquire operation condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, use the temperature field prediction model to predict the temperature field inside the power distribution cabinet, identify local hot spots according to the prediction results, and obtain local hot spot identification information; Determine the hot spot area inside the target power distribution cabinet according to the local hot spot identification information, evaluate whether the hot spot area will have thermal runaway, and formulate a control strategy to prevent thermal runaway; Acquire the operation characteristics of the internal components of the target group distribution cabinet according to the real-time operation monitoring information, use the isolation forest algorithm to detect abnormal operation data of the internal components of the distribution cabinet, and analyze the abnormal operation status to obtain abnormal operation status analysis information; Use graph neural network to build component life prediction model, obtain the operating environment monitoring information of the target group distribution cabinet, combine the abnormal operating status analysis information to predict the remaining life of the internal components of the target group distribution cabinet, and carry out maintenance warning; The construction of the power distribution cabinet CFD model, presetting a number of grouped power distribution cabinet operating conditions, performing cabinet temperature simulation and forming a cabinet temperature simulation data set, constructing a temperature field prediction model through cGAN and using the cabinet temperature simulation data set for model training specifically includes: Obtain the distribution cabinet structure information of the target group distribution cabinet, use the three-dimensional modeling technology to construct a geometric model of the target group distribution cabinet according to the distribution cabinet structure information, and use data retrieval to obtain and define the material properties of the target group distribution cabinet; Based on the structure information of the distribution cabinet, the structural characteristics of the target group distribution cabinet are extracted, including the structural characteristics of the components and the heat dissipation structure characteristics, and the structural characteristic information of the distribution cabinet is obtained. The heating components are analyzed in combination with the equipment specification of the target group distribution cabinet, and the heat source definition and heat source type definition are performed to obtain the heat source analysis information in the cabinet; Using CFD software combined with the geometric model of the target group distribution cabinet to build a CFD model of the distribution cabinet, and meshing the CFD model of the distribution cabinet, and defining the heat source area in the CFD model of the distribution cabinet according to the heat source analysis information in the cabinet; Based on the structural characteristic information of the power distribution cabinet, the heat dissipation structure characteristics are obtained to set the fluid boundary conditions, including natural convection boundary conditions and forced convection boundary conditions, and the radiation boundary conditions are set according to the heat source analysis information and the heat source area in the cabinet; Preset several group distribution cabinet operating conditions, set the global initial temperature, use the k-ε turbulence model and the steady-state solver to simulate the cabinet temperature of the target group distribution cabinet under different operating conditions, and form a cabinet temperature simulation data set; A temperature field prediction model is constructed based on cGAN, a training data set is constructed according to the cabinet temperature simulation data set, and the operating condition attributes corresponding to each cabinet temperature simulation data are used as condition constraints to train the temperature field prediction model to obtain a temperature field prediction model that meets expectations; The method of acquiring operating condition characteristics and temperature monitoring characteristics based on the real-time operation monitoring information, predicting the temperature field inside the power distribution cabinet using the temperature field prediction model, and identifying local hot spots according to the prediction results specifically includes: Monitor the target group distribution cabinet to obtain real-time operation monitoring information, perform data preprocessing on the real-time operation monitoring information, and input the preprocessed real-time operation monitoring information into the temperature field prediction model for analysis; Extract features from the input real-time operation monitoring data, obtain the operating condition features and temperature monitoring features of the target group distribution cabinets, and form constraint conditions based on the operating condition features; The generator generates the initial temperature field prediction result according to the temperature monitoring characteristics, inputs the generated initial temperature field prediction result into the discriminator for judgment, and determines whether to accept the current prediction result in combination with the constraint conditions; If the judgment result of the discriminator is acceptance, the corresponding temperature field prediction result is output; if the judgment result of the discriminator is rejection, the generator continues to generate the next prediction result until the constraint condition is met and the final temperature field prediction result is output to obtain temperature field prediction information; Generate a temperature distribution gradient map of the target grouped power distribution cabinet according to the temperature field prediction information, divide the internal area of the target grouped power distribution cabinet into a heat source area and a non-heat source area, and set a normal operating temperature range; Local hotspot identification is performed through the temperature distribution gradient map according to the normal operating temperature range corresponding to each area to obtain local hotspot identification information.
Citation Information
Patent Citations
Method for identifying overheating fault of high-voltage power distribution cabinet
CN114611566A
Air conditioner operation and maintenance system and method of air cooling data center and electronic equipment
CN116963451A
Control strategy optimization method and system for industrial and commercial park light storage and charging micro-grid
CN117040028A
Direct-current arc simulation method and apparatus, and computer device and storage medium
WO2023226113A1
Cited By
Positioning equipment for water wall tube welding
CN121042665A
