A digital twin method and device for substation based on large-scale point cloud

Abstract

The present invention provides a substation digital twin method and device based on large-scale point cloud, wherein the method comprises: step 1: obtaining target point cloud data of a target substation; step 2: based on large-scale point cloud rendering technology, loading target nodes according to target point cloud data; step 3: performing data twinning on the target point cloud data through the target node to obtain a process digital twin model; step 4: performing point cloud optimization on the process digital twin model to obtain a target digital twin model. The present invention provides a substation digital twin method and device based on large-scale point cloud, wherein the target point cloud data of the target substation is obtained, large-scale point cloud rendering technology is introduced, and target nodes are loaded. Data twinning is performed on the target point cloud data through the target node to obtain a process digital twin model, point cloud optimization is performed on the process digital twin model, and a target digital twin model is obtained, thereby reducing the deviation of the point cloud data and improving the accuracy of the target digital twin model.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and in particular to a substation digital twin method and device based on large-scale point cloud. Background Art

[0002] Large-scale point clouds refer to a 3D data set containing a large number of discrete points. For a substation, a large-scale point cloud can contain geometric information of various components and structures of the entire substation, such as generators, transformers, switchgear, and transmission lines.

[0003] Digital Twin is a concept that combines a physical entity with its digital representation. It creates a virtual, real-time, and operational digital model by acquiring, analyzing, and simulating the data of a physical entity in real time.

[0004] In digital twins, large-scale point clouds are usually acquired through laser scanning or other 3D data acquisition technologies. These technologies can quickly and accurately capture the geometric shape and surface features of the target object or scene and generate a large amount of discrete point data. These point cloud data can provide detailed information such as spatial position, shape, size and relative position relationship, providing a basis for establishing a digital twin model of the substation.

[0005] The invention patent with application number: CN202111363476.8 discloses a method and device for digital twinning of substations based on large-scale point clouds, wherein the method includes: effectively extracting key equipment data in the point cloud through point cloud preprocessing, cluster segmentation, straight line fitting and other technologies, so as to facilitate subsequent classification and storage. Through the tile-type inner layer storage method, point cloud data of different importance levels are stored according to different requirements, which reduces the size of point cloud data and facilitates subsequent rendering; through large-scale point cloud rendering technology, combined with actual conditions, nodes at the corresponding level that need to be loaded and rendered are found for loading, thereby improving the rendering efficiency of large-scale point clouds; through digital twin virtual-real mapping, data simulation, real-time synchronization and other technologies, more intuitive monitoring and analysis of substation equipment and comprehensive management of the entire life cycle are achieved; through robot simulation technology, the robot inspection process is synchronized to the digital twin virtual model of the substation, which makes it more convenient to monitor and manage inspection tasks.

[0006] However, after the above-mentioned existing technology creates a digital twin of the substation, the twin results are not verified, and there is a situation where the point cloud data is biased, which leads to inaccurate twin models.

[0007] In view of this, there is an urgent need for a substation digital twin method and device based on large-scale point cloud to at least solve the above-mentioned shortcomings. Summary of the invention

[0008] One of the purposes of the present invention is to provide a substation digital twin method based on large-scale point cloud, obtain the target point cloud data of the target substation, introduce large-scale point cloud rendering technology, and load the target node. Through the target node, the target point cloud data is twinned to obtain the process digital twin model, the process digital twin model is optimized by point cloud, and the target digital twin model is obtained, which reduces the deviation of the point cloud data and improves the accuracy of the target digital twin model.

[0009] An embodiment of the present invention provides a substation digital twin method based on large-scale point cloud, including:

[0010] Step 1: Obtain the target point cloud data of the target substation;

[0011] Step 2: Based on large-scale point cloud rendering technology, load the target node according to the target point cloud data;

[0012] Step 3: Perform data twinning on the target point cloud data through the target node to obtain the process digital twin model;

[0013] Step 4: Perform point cloud optimization on the process digital twin model to obtain the target digital twin model.

[0014] Preferably, step 2: based on large-scale point cloud rendering technology, loading the target node according to the target point cloud data, includes:

[0015] Obtaining structural attributes and functional attributes of the target substation;

[0016] Based on machine learning technology, the first target features of the substation components are learned according to the structural attributes and functional attributes;

[0017] According to the first target feature, the target point cloud data is segmented to obtain segmentation data;

[0018] Extracting a second target feature of the segmented data, the second target feature being: a volume, a shape and a curvature of a transformer component corresponding to the segmented data;

[0019] Determine a third target feature according to the relative position of the segmented data in the target point cloud data, the third target feature being: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data;

[0020] The target node is loaded according to the second target feature and the third target feature.

[0021] Preferably, step 3: performing data twinning on the target point cloud data through the target node to obtain a process digital twin model, including:

[0022] The fourth target feature and related parameters of the extracted target node are input into the digital twin parameter input model to obtain a process digital twin model. The digital twin parameter input model is: a model based on physical principles or a model based on data-driven.

[0023] Preferably, step 4: performing point cloud optimization on the process digital twin model to obtain a target digital twin model includes:

[0024] Simulate and analyze the established process digital twin model, simulate the power transmission process, electromagnetic field distribution and thermal coupling process of the target substation under different target working conditions, and obtain simulation results;

[0025] Obtain historical results corresponding to the target working conditions;

[0026] Point cloud optimization is performed based on simulation results and historical results to obtain the target digital twin model.

[0027] Preferably, simulating the power transmission process, electromagnetic field distribution and thermal coupling process of the target substation under different target working conditions to obtain simulation results includes:

[0028] Determine the different target operating conditions that need to be simulated, including: rated load, partial load, overload and fault;

[0029] According to the target working conditions, set the corresponding power supply voltage, load power and system fault;

[0030] Based on the power system simulation software, the process digital twin model is simulated and calculated according to the set target operating conditions. The simulation calculations include: power flow calculation, short-circuit calculation, steady-state and transient stability analysis;

[0031] Obtaining the calculation results of the simulation calculation;

[0032] According to the calculation results, calculate the electromagnetic field distribution;

[0033] Evaluate electromagnetic compatibility (EMC) and electromagnetic radiation based on electromagnetic field distribution based on finite element analysis (FEA) and / or finite difference time domain method (FDTD);

[0034] Combine the calculation results with the thermal simulation to perform thermal coupling simulation and obtain thermal coupling simulation results;

[0035] Determine the thermal stability of the target substation based on the thermal coupling simulation results;

[0036] The calculation results, electromagnetic compatibility, electromagnetic radiation and thermal stability of the target substation are taken together as simulation results.

[0037] An embodiment of the present invention provides a substation digital twin method based on large-scale point cloud, further comprising:

[0038] Step 5: Compare and monitor the twin data of the target digital twin model and the actual data of the target substation, monitor the operating status of the target substation in real time, and perform fault detection;

[0039] Among them, the twin data of the target digital twin model and the actual data of the target substation are compared and monitored, including:

[0040] Obtaining comparison results of data comparison;

[0041] Get the abnormal level interval set;

[0042] Determine that the comparison result falls within an abnormal level interval in the abnormal level interval set;

[0043] Obtaining a target marked range of the first simulated power transformation component corresponding to the comparison result;

[0044] According to the marking rules of the abnormal level range, the comparison results are marked within the corresponding target marking range;

[0045] Wherein, obtaining the target marking range of the first simulated power transformation component corresponding to the comparison result includes:

[0046] Acquire a second simulated power conversion component adjacent to the first simulated power conversion component, and calculate a component distance between the first simulated power conversion component and the second simulated power conversion component;

[0047] The target marking range is determined according to the component distance and the simulated display range of the first simulated power transformation component.

[0048] Preferably, obtaining an abnormal level interval set includes:

[0049] Obtain target abnormal data;

[0050] Analyze target abnormal data and obtain abnormal comparison data;

[0051] Based on the preset clustering rules, the abnormal comparison data is clustered according to the abnormal probabilities of the abnormal items corresponding to the abnormal comparison data to obtain multiple abnormal comparison data clusters;

[0052] Determine the abnormal level interval sequence of the abnormal comparison data clustering cluster according to the size order of the median of the abnormal probability corresponding to the abnormal comparison data clustering cluster;

[0053] The abnormal comparison data in the abnormal comparison data cluster corresponding to each abnormal level interval serial number are continuousized to obtain the abnormal level interval corresponding to each abnormal level interval serial number;

[0054] Integrate the abnormal level intervals to obtain an abnormal level interval set;

[0055] Among them, obtaining target abnormal data includes:

[0056] Acquire the first abnormal data in the history of the target substation, and at the same time, acquire the second abnormal data of the target substation based on the big data technology, the second abnormal data being: the abnormal data of the substation with similar attributes to the target substation acquired by the big data;

[0057] The first abnormal data and the second abnormal data are used together as target abnormal data.

[0058] Preferably, real-time monitoring of the operating status of the target substation and fault detection are performed, including:

[0059] Determine the substation components that need to be fault-detected and use them as target components;

[0060] Obtaining a comparison sub-result of the twin data of the target component and the actual data of the target component for data comparison;

[0061] According to the comparison results, determine the detection equipment and obtain the location of the detection equipment;

[0062] Acquire a first position of a target component corresponding to a substation;

[0063] Obtain the second position of the maintenance personnel;

[0064] Based on the regional distribution map of the target substation, a plurality of travel routes are planned according to the detection equipment location, the first location, and the second location;

[0065] Determine the travel cost of each maintenance worker according to the regional attributes of the travel area of ​​the travel route;

[0066] The maintenance personnel with the lowest travel cost are selected as target maintenance personnel, and the target maintenance personnel are notified to carry detection equipment to perform corresponding fault detection;

[0067] The substation components that need to be detected for faults include:

[0068] Obtain the first display duration of the comparison result in real time;

[0069] Obtaining a first threshold duration of the color of the comparison result, and at the same time, obtaining a second threshold duration of the color of the comparison result;

[0070] If the sum of the first display durations of the comparison results is greater than or equal to the preset second threshold duration or any first display duration is greater than or equal to the corresponding first threshold duration, the substation component compared with the corresponding comparison result is used as the substation component that needs to be detected for faults;

[0071] Among them, multiple travel routes are planned, including:

[0072] Obtain the first inertia weight, the maximum number of iterations and the second inertia weight of the particle swarm algorithm to search for the route;

[0073] Based on the target distribution generator, the improved inertia weight is determined according to the first inertia weight, the maximum number of iterations and the second inertia weight. The determination model is as follows:

[0074]

[0075] Among them, γ is the improved inertia weight, γ min is the second inertia weight, γ max is the first inertia weight, exp(*) is the exponential function with e as the base, count is the current number of iterations, count max is the maximum number of iterations, betarnd is the target distribution generator;

[0076] The travel route is planned based on the particle swarm algorithm with improved inertia weight.

[0077] An embodiment of the present invention provides a substation digital twin device based on large-scale point cloud, including:

[0078] A target point cloud data acquisition subsystem is used to acquire target point cloud data of a target substation;

[0079] The target node loading subsystem is used to load the target node according to the target point cloud data based on the large-scale point cloud rendering technology;

[0080] The process digital twin model acquisition subsystem is used to perform data twinning on the target point cloud data through the target node to obtain the process digital twin model;

[0081] The target digital twin model acquisition subsystem is used to perform point cloud optimization on the process digital twin model and obtain the target digital twin model.

[0082] Preferably, the target node loading subsystem includes:

[0083] An attribute acquisition module is used to acquire the structural attributes and functional attributes of the target substation;

[0084] A first target feature extraction module, used for learning the first target feature of the power transformation component according to the structural attributes and the functional attributes based on the machine learning technology;

[0085] A data segmentation module, used to segment the target point cloud data according to the first target feature to obtain segmentation data;

[0086] A second target feature extraction module, used to extract the second target feature of the segmented data, the second target feature being: the volume, shape and curvature of the transformer component corresponding to the segmented data;

[0087] A third target feature determination module is used to determine a third target feature according to the relative position of the segmented data in the target point cloud data, where the third target feature is: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data;

[0088] The target node loading module is used to load the target node according to the second target feature and the third target feature.

[0089] The beneficial effects of the present invention are:

[0090] The present invention obtains the target point cloud data of the target substation, introduces large-scale point cloud rendering technology, and loads the target node. Through the target node, the target point cloud data is subjected to data twinning to obtain the process digital twin model, and the process digital twin model is subjected to point cloud optimization to obtain the target digital twin model, thereby reducing the deviation of the point cloud data and improving the accuracy of the target digital twin model.

[0091] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the present application documents.

[0092] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0094] Figure 1 Schematic diagram of a substation digital twin method based on large-scale point cloud in an embodiment of the present invention;

[0095] Figure 2 Schematic diagram of a substation digital twin device based on large-scale point cloud in an embodiment of the present invention. DETAILED DESCRIPTION

[0096] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0097] The embodiment of the present invention provides a substation digital twin method based on large-scale point cloud, such as Figure 1 As shown, including:

[0098] Step 1: Obtain the target point cloud data of the target substation; the target substation is: the substation that needs to be digitally twinned, and the digital twinning of the substation is to optimize the operation, improve efficiency, fault diagnosis, predictive maintenance, safety and risk management of the substation; the target point cloud data is: the point cloud data of the substation components, equipment or structures, such as: the point cloud data corresponding to the transformer, switchgear, generator and transmission line; when obtaining the target point cloud data, according to the point cloud data acquisition method, obtain the pre-processed point cloud data of the target substation, the point cloud data acquisition method includes: laser scanning, 3D photogrammetry, structured light scanning and radar scanning, the pre-processed point cloud data is: the point cloud data directly obtained by the point cloud data acquisition method; perform target data processing on the pre-processed point cloud data to obtain the target point cloud data, the target data processing includes: filtering, sampling, denoising, point cloud pre-processing, clustering segmentation and straight line fitting;

[0099] Step 2: Based on the large-scale point cloud rendering technology, load the target node according to the target point cloud data; the large-scale point cloud rendering technology refers to the technical methods and algorithms used to process and render large-scale point cloud data sets; when loading the target node, select the target node to be loaded and rendered in the point cloud data according to the identification or other attributes of the target node. When selecting, you can filter and select by the label, ID or other attributes of the point cloud data;

[0100] Step 3: Data twinning of the target point cloud data is performed through the target node to obtain a process digital twin model; wherein the rendering engine will apply rendering algorithms and effects according to the geometric information and attributes of the target node to generate realistic rendering results, including: point cloud coloring, lighting model and shadow generation, etc.;

[0101] Step 4: Perform point cloud optimization on the process digital twin model to obtain the target digital twin model. Point cloud optimization can improve the quality, fidelity and efficiency of point cloud data, improve the performance and effect of subsequent point cloud processing and application, and improve the accuracy of the target digital twin model. Point cloud optimization methods include: denoising, sampling and dimensionality reduction, point cloud registration, point cloud reconstruction, point cloud filtering, point cloud repair and point cloud enhancement.

[0102] The working principle and beneficial effects of the above technical solution are:

[0103] This application obtains the target point cloud data of the target substation, introduces large-scale point cloud rendering technology, and loads the target node. Through the target node, the target point cloud data is twinned to obtain the process digital twin model, and the process digital twin model is optimized to obtain the target digital twin model, which reduces the deviation of the point cloud data and improves the accuracy of the target digital twin model.

[0104] In one embodiment, step 2: based on large-scale point cloud rendering technology, loading the target node according to the target point cloud data includes:

[0105] Acquire the structural attributes and functional attributes of the target substation; wherein the structural attributes are: connection information between substation components in the target substation; and the functional attributes are: functional information of the substation components;

[0106] Based on machine learning technology, the first target feature of the substation component is learned according to the structural attributes and functional attributes; wherein the first target feature is, for example, which substation component is connected to which substation component, and for example, which substation component has the function of insulation;

[0107] According to the first target feature, the target point cloud data is segmented to obtain segmented data; wherein, when the target point cloud data is segmented according to the first target feature, the point cloud data is segmented into different components according to the structure and function of the substation, such as transformers, switchgear, and insulators, etc., and a method based on geometric features or a machine learning algorithm can be used to perform point cloud segmentation;

[0108] Extracting a second target feature of the segmented data, the second target feature being: a volume, a shape and a curvature of a transformer component corresponding to the segmented data;

[0109] Determine a third target feature according to the relative position of the segmented data in the target point cloud data, the third target feature being: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data;

[0110] The target node is loaded according to the second target feature and the third target feature.

[0111] The working principle and beneficial effects of the above technical solution are:

[0112] This application introduces machine learning technology to learn the first target feature based on the acquired structural attributes and functional attributes of the target substation, and segment the target point cloud data based on the first target feature to obtain the segmented data, extract the second target feature and the third target feature of the segmented data, and load the data at the target node to make the loading process more suitable.

[0113] In one embodiment, step 3: performing data twinning on the target point cloud data through the target node to obtain a process digital twin model includes:

[0114] The fourth target feature and related parameters of the extracted target node are input into the digital twin parameter input model to obtain the process digital twin model. The digital twin parameter input model is a model based on physical principles or a model driven by data. Among them, the fourth target feature is: the geometric parameters and connection relationship parameters corresponding to the target node loaded; the related parameters are: material parameters.

[0115] The working principle and beneficial effects of the above technical solution are:

[0116] The present application inputs the fourth target feature and related parameters of the target node into a model based on physical principles or a data-driven model to obtain a process digital twin model, thereby improving the accuracy of twin model construction.

[0117] In one embodiment, step 4: performing point cloud optimization on the process digital twin model to obtain a target digital twin model includes:

[0118] Simulate and analyze the established process digital twin model, simulate the power transmission process, electromagnetic field distribution and thermal coupling process of the target substation under different target working conditions, and obtain simulation results;

[0119] Obtaining historical results corresponding to the target working condition; wherein the historical results are: the operating status of the target substation under the target working condition;

[0120] Point cloud optimization is performed based on simulation results and historical results to obtain the target digital twin model. When point cloud optimization is performed based on simulation results and historical results, if the deviation between the simulation results and historical results is greater than a preset threshold, the cause of the deviation is determined based on the attribute information corresponding to the deviation position (such as the type, connection relationship, and material of the substation component). If the cause of the deviation is related to the point cloud error, optimization is performed.

[0121] The working principle and beneficial effects of the above technical solution are:

[0122] This application simulates and analyzes the power transmission process, electromagnetic field distribution and thermal coupling process under different target working conditions, obtains simulation results, and compares the simulation results with the historical results corresponding to the target working conditions. Point cloud optimization is performed based on the comparison results to obtain the target digital twin model. The point cloud optimization process is more appropriate and reasonable.

[0123] In one embodiment, the power transmission process, electromagnetic field distribution and thermal coupling process of the target substation under different target working conditions are simulated to obtain simulation results, including:

[0124] Determine the different target operating conditions that need to be simulated, including: rated load, partial load, overload and fault;

[0125] According to the target working conditions, set the corresponding power supply voltage, load power and system fault;

[0126] Based on the power system simulation software, the process digital twin model is simulated and calculated according to the set target operating conditions. The simulation calculations include: power flow calculation, short-circuit calculation, steady-state and transient stability analysis;

[0127] Obtaining the calculation results of the simulation calculation; wherein the calculation results include: information such as voltage, current and power;

[0128] Calculate the electromagnetic field distribution according to the calculation results; when calculating the electromagnetic field distribution, calculate the electromagnetic field distribution around each device and the transmission line according to the simulation results;

[0129] Based on finite element analysis FEA and / or finite difference time domain method FDTD, electromagnetic compatibility (EMC) and electromagnetic radiation are evaluated according to the electromagnetic field distribution; FEA (Finish Element Analysis) discretizes complex practical problems into a finite number of simplified units, and obtains approximate solutions to the problems through numerical calculation methods; FDTD (Finish Difference Time Domain) is a method for numerically solving electromagnetic wave equations, which is often used in time domain simulation and analysis of electromagnetic fields; EMC (Electromagnetic Compatibility) refers to the ability of electronic equipment, systems or systems to operate normally in an electromagnetic environment without generating or being subject to unacceptable interference to their performance;

[0130] Combining the calculation results with the thermal simulation, performing thermal coupling simulation, and obtaining thermal coupling simulation results; wherein, combining the calculation results with the thermal simulation can be achieved by adding a thermal model in the power system simulation software or inputting the calculation results into the thermal simulation software;

[0131] Determine the thermal stability of the target substation based on the thermal coupling simulation results;

[0132] The calculation results, electromagnetic compatibility, electromagnetic radiation and thermal stability of the target substation are taken together as simulation results.

[0133] The working principle and beneficial effects of the above technical solution are:

[0134] This application introduces power system simulation software to perform simulation calculations according to different target operating conditions that need to be simulated. According to the calculation results, the electromagnetic field distribution is calculated and the electromagnetic compatibility EMC and electromagnetic radiation are analyzed. The calculation results are combined with thermal simulation to obtain thermal coupling simulation results, and the thermal stability of the target substation is determined based on the thermal coupling simulation results. The calculation results, electromagnetic compatibility, electromagnetic radiation and thermal stability of the target substation are used together as simulation results, which improves the comprehensiveness of the simulation results and improves the suitability of subsequent simulation results and historical results.

[0135] The embodiment of the present invention provides a substation digital twin method based on large-scale point cloud, further comprising:

[0136] Step 5: Compare and monitor the twin data of the target digital twin model and the actual data of the target substation, monitor the operating status of the target substation in real time, and perform fault detection; where the twin data refers to the model data captured within the preset target time period before the target digital twin model generates the corresponding data when the substation is abnormal;

[0137] Among them, the twin data of the target digital twin model and the actual data of the target substation are compared and monitored, including:

[0138] Obtaining a comparison result of data comparison; wherein data comparison is: comparing data items of the same type in the twin data and the actual data; the comparison result is: the numerical difference of the data item;

[0139] Get the abnormal level interval set;

[0140] Determine that the comparison result falls within an abnormal level interval in the abnormal level interval set; wherein the abnormal level interval is: a value range of the comparison result corresponding to each abnormal level interval;

[0141] Obtain a target annotation range of the first simulated substation component corresponding to the comparison result; wherein the first simulated substation component is: a model of the substation component in the target digital twin model; the target annotation range is: an annotation position interval of the prompt word of the first simulated substation component in the target digital twin model;

[0142] According to the marking rules of the abnormal level range, the comparison results are marked within the corresponding target marking range; wherein the marking rules are: which color is marked for which abnormal level;

[0143] Wherein, obtaining the target marking range of the first simulated power transformation component corresponding to the comparison result includes:

[0144] Acquire a second simulated substation component adjacent to the first simulated substation component, and calculate a component distance between the first simulated substation component and the second simulated substation component; wherein the second simulated substation component is: a model of a substation component adjacent to the first simulated substation component in a target digital twin model; and the component distance is: a distance between the first simulated substation component and the second simulated substation component in the target digital twin model;

[0145] The target marking range is determined according to the component distance and the simulated display range of the first simulated transformer component, wherein the target marking range is: the centroid of the area that determines the simulated display range of the first simulated transformer component is the center of the circle, the shortest component distance is the radius of the circular area, and the area outside the simulated display range in the circular area.

[0146] The working principle and beneficial effects of the above technical solution are:

[0147] The present application determines that the comparison result falls within the abnormal level interval set according to the comparison result of the acquired data and the abnormal level interval set. According to the marking rule of the abnormal level interval, the comparison result is marked within the acquired target marking range. When obtaining the target marking range of the first simulated substation component corresponding to the comparison result, the component distance between the first simulated substation component and the second simulated substation component is obtained, and the target marking range is determined according to the component distance and the simulated display range of the first simulated substation component. The determination of the target marking range is more reasonable.

[0148] In one embodiment, obtaining an abnormal level interval set includes:

[0149] Obtain target abnormal data;

[0150] Analyze the target abnormal data and obtain abnormal comparison data; wherein the abnormal comparison data is: the comparison data to be compared when the abnormality occurs;

[0151] Based on the preset clustering rules, the abnormal comparison data are clustered according to the abnormal probabilities of the abnormal items corresponding to the abnormal comparison data to obtain multiple abnormal comparison data clustering clusters; wherein the clustering rule is: what is the proportion of the amount of abnormal comparison data allocated to each clustering cluster to the total amount of abnormal comparison data, for example: based on the order of abnormal probability from small to large, the abnormal comparison data are clustered according to the proportion of 7:2:1;

[0152] Determine the abnormal level interval sequence of the abnormal comparison data clustering cluster according to the size order of the median of the abnormal probability corresponding to the abnormal comparison data clustering cluster;

[0153] The abnormal comparison data in the abnormal comparison data cluster corresponding to each abnormal level interval serial number are continuousized to obtain the abnormal level interval corresponding to each abnormal level interval serial number;

[0154] Integrate the abnormal level intervals to obtain an abnormal level interval set;

[0155] Among them, obtaining target abnormal data includes:

[0156] Acquire the first abnormal data in the history of the target substation, and at the same time, acquire the second abnormal data of the target substation based on big data technology; wherein the first abnormal data is: the abnormal data generated in the history of the target substation; the second abnormal data is: the abnormal data of the substation with similar attributes to the target substation acquired by big data;

[0157] The first abnormal data and the second abnormal data are used together as target abnormal data.

[0158] The working principle and beneficial effects of the above technical solution are:

[0159] This application obtains the first abnormal data in the history of the target substation and the second abnormal data of the target substation obtained from the big data as the target abnormal data, which improves the comprehensiveness of the target abnormal data. The target abnormal data is parsed to obtain abnormal comparison data. At the same time, clustering rules are introduced to cluster the abnormal comparison data according to the abnormal probability of the abnormal items corresponding to the abnormal comparison data to obtain abnormal comparison data clustering clusters. According to the order of the medians of the abnormal probabilities corresponding to the abnormal comparison data clustering clusters, the abnormal level interval serial number corresponding to the abnormal comparison data clustering cluster is determined, and the abnormal comparison data clustering cluster numerical value is continuous to obtain the abnormal level interval range corresponding to the abnormal level interval serial number, and the abnormal level interval is integrated to obtain the abnormal level interval set, and the construction of the abnormal level interval set is more accurate.

[0160] In one embodiment, real-time monitoring of the operating status of the target substation and fault detection include:

[0161] Determine the substation components that need to be fault-detected and use them as target components;

[0162] Obtaining a comparison sub-result of data comparison between the twin data of the target component and the actual data of the target component; wherein the comparison sub-result is: a numerical difference between data items of the same data type of the twin data of the target component and the actual data;

[0163] According to the comparison sub-result, determine the detection device and obtain the location of the detection device; wherein the detection device is determined according to the fault cause represented by the comparison sub-result;

[0164] Acquire a first position of the target component corresponding to the substation; wherein the first position is: the position of the target component in the substation;

[0165] Obtaining a second position of the maintenance personnel; wherein the second position is: a real-time position of the maintenance personnel;

[0166] Based on the regional distribution map of the target substation, multiple travel routes are planned according to the detection device location, the first location, and the second location; wherein the regional distribution map is: a map of the target substation; the travel route is: a route planned based on the path planning technology from the second location through the detection device location and to the first location;

[0167] According to the regional attributes of the passing area of ​​the passing route, the passing cost of each maintenance personnel is determined; the regional attributes are: the area range and the configuration information of the substation components in the area; the passing cost is: the normalized result of the time required to pass through the passing area and the resources required for the protective equipment;

[0168] The maintenance personnel with the lowest travel cost are selected as target maintenance personnel, and the target maintenance personnel are notified to carry detection equipment to perform corresponding fault detection;

[0169] The substation components that need to be detected for faults include:

[0170] Obtaining the first display duration of the comparison result in real time; wherein the first display duration is: the presentation time of the comparison result in the target digital twin model;

[0171] Obtaining a first threshold duration of the color of the comparison result, and at the same time, obtaining a second threshold duration of the color of the comparison result; wherein the first threshold duration of the color of the comparison result and the second threshold duration of the color of the comparison result are manually preset, and the first threshold duration is less than the second threshold duration;

[0172] If the sum of the first display durations of the comparison results is greater than or equal to the preset second threshold duration or any first display duration is greater than or equal to the corresponding first threshold duration, the substation component compared with the corresponding comparison result is used as the substation component that needs to be detected for faults;

[0173] Among them, multiple travel routes are planned, including:

[0174] Obtain the first inertia weight, maximum number of iterations and second inertia weight of the particle swarm algorithm for searching the route; wherein the first inertia weight is: the initial inertia weight of the particle swarm algorithm for searching the route; the maximum number of iterations is: the number of iterations when the particle swarm algorithm searches the route until convergence; the second inertia weight is: the inertia weight of searching the route at the maximum number of iterations;

[0175] Based on the target distribution generator, the improved inertia weight is determined according to the first inertia weight, the maximum number of iterations and the second inertia weight. The determination model is as follows:

[0176]

[0177] Among them, γ is the improved inertia weight, γ min is the second inertia weight, γ max is the first inertia weight, exp(*) is the exponential function with e as the base, count is the current number of iterations, count max is the maximum number of iterations, betarnd is the target distribution generator, which is the betarnd function called in the matlab simulation software;

[0178] The travel route is planned based on the particle swarm algorithm with improved inertia weight.

[0179] The working principle and beneficial effects of the above technical solution are:

[0180] The present application determines the target component that needs to be detected for faults. When determining, the first display time of the comparison result and the first threshold time and the second threshold time of the color marked corresponding to the comparison result are obtained in real time. If the sum of the time and value of the first display time of the comparison result is greater than or equal to the preset second threshold time, it means that the cumulative time for the corresponding substation component to prompt an abnormality is long and requires maintenance; in addition, if any first display time is greater than or equal to the corresponding first threshold time, it means that the duration of the corresponding substation component prompting an abnormality is long and also requires maintenance, thereby improving the comprehensiveness and suitability of the target component determination.

[0181] After determining the target component, determine the detection equipment based on the comparison results of the target component, such as what kind of detection instrument is needed. Introduce the regional distribution map of the target substation, plan the route from the second location through the detection equipment location and go to the first location. When planning the travel route, introduce the particle swarm algorithm with reorganized inertia weights to improve the planning ability when planning the travel route and improve the accuracy of the travel route. Determine the travel cost of the maintenance personnel based on the regional attributes of the travel area of ​​the travel route. Determine the target maintenance personnel with the lowest travel cost, and notify the target maintenance personnel to carry the detection equipment for corresponding fault detection, thereby improving the suitability of maintenance.

[0182] The embodiment of the present invention provides a substation digital twin device based on large-scale point cloud, such as Figure 2 As shown, including:

[0183] The target point cloud data acquisition subsystem 1 is used to acquire the target point cloud data of the target substation;

[0184] The target node loading subsystem 2 is used to load the target node according to the target point cloud data based on the large-scale point cloud rendering technology;

[0185] The process digital twin model acquisition subsystem 3 is used to perform data twinning on the target point cloud data through the target node to obtain the process digital twin model;

[0186] The target digital twin model acquisition subsystem 4 is used to perform point cloud optimization on the process digital twin model to obtain the target digital twin model.

[0187] In one embodiment, the target node loads a subsystem, including:

[0188] An attribute acquisition module is used to acquire the structural attributes and functional attributes of the target substation;

[0189] A first target feature extraction module, used for learning the first target feature of the power transformation component according to the structural attributes and the functional attributes based on the machine learning technology;

[0190] A data segmentation module, used to segment the target point cloud data according to the first target feature to obtain segmentation data;

[0191] A second target feature extraction module, used to extract the second target feature of the segmented data, the second target feature being: the volume, shape and curvature of the transformer component corresponding to the segmented data;

[0192] A third target feature determination module is used to determine a third target feature according to the relative position of the segmented data in the target point cloud data, where the third target feature is: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data;

[0193] The target node loading module is used to load the target node according to the second target feature and the third target feature.

[0194] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A substation digital twin method based on large-scale point cloud, characterized in that: include: Step 1: Obtain the target point cloud data of the target substation; Step 2: Based on large-scale point cloud rendering technology, load the target node according to the target point cloud data; Step 3: Perform data twinning on the target point cloud data through the target node to obtain the process digital twin model; Step 4: Optimize the point cloud of the process digital twin model to obtain the target digital twin model; Step 5: Compare and monitor the twin data of the target digital twin model and the actual data of the target substation, monitor the operating status of the target substation in real time, and perform fault detection; Among them, the twin data of the target digital twin model and the actual data of the target substation are compared and monitored, including: Obtaining comparison results of data comparison; Get the abnormal level interval set; Determine that the comparison result falls within an abnormal level interval in the abnormal level interval set; Obtaining a target marked range of the first simulated power transformation component corresponding to the comparison result; According to the marking rules of the abnormal level range, the comparison results are marked within the corresponding target marking range; Wherein, obtaining the target marking range of the first simulated power transformation component corresponding to the comparison result includes: Acquire a second simulated power conversion component adjacent to the first simulated power conversion component, and calculate a component distance between the first simulated power conversion component and the second simulated power conversion component; The target marking range is determined according to the component distance and the simulated display range of the first simulated power transformation component.

2. A substation digital twin method based on large-scale point cloud as claimed in claim 1, characterized in that: Step 2: Based on large-scale point cloud rendering technology, load the target node according to the target point cloud data, including: Obtaining structural attributes and functional attributes of the target substation; Based on machine learning technology, first target features of the substation components are learned according to structural attributes and functional attributes, and the first target features include: connection relationships between substation components and functions of substation components; According to the first target feature, the target point cloud data is segmented to obtain segmented data, where the segmented data is: point cloud data corresponding to different components obtained by segmenting the target point cloud data according to the structure and function of the substation; Extracting a second target feature of the segmented data, the second target feature being: a volume, a shape and a curvature of a transformer component corresponding to the segmented data; Determine a third target feature according to the relative position of the segmented data in the target point cloud data, the third target feature being: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data; The target node is loaded according to the second target feature and the third target feature.

3. A substation digital twin method based on large-scale point cloud as claimed in claim 1, characterized in that: Step 3: Use the target node to perform data twinning on the target point cloud data and obtain the process digital twin model, including: The fourth target feature and related parameters of the extracted target node are input into the digital twin parameter input model to obtain a process digital twin model. The digital twin parameter input model is: a model based on physical principles or a model based on data-driven; the fourth target feature is: the geometric parameters and connection relationship parameters loaded corresponding to the target node; the related parameters are: material parameters.

4. A substation digital twin method based on large-scale point cloud as claimed in claim 1, characterized in that: Step 4: Optimize the point cloud of the process digital twin model to obtain the target digital twin model, including: Simulate and analyze the established process digital twin model to simulate the power transmission process, electromagnetic field distribution, and thermal coupling process of the target substation under different target working conditions and obtain simulation results; Obtain historical results corresponding to the target working conditions; Point cloud optimization is performed based on simulation results and historical results to obtain the target digital twin model.

5. A substation digital twin method based on large-scale point cloud as claimed in claim 4, characterized in that: Simulate the power transmission process, electromagnetic field distribution and thermal coupling process of the target substation under different target working conditions and obtain simulation results, including: Determine the different target operating conditions that need to be simulated, including: rated load, partial load, overload and fault; According to the target working conditions, set the corresponding power supply voltage, load power and system fault; Based on the power system simulation software, the process digital twin model is simulated and calculated according to the set target operating conditions. The simulation calculations include: power flow calculation, short-circuit calculation, steady-state and transient stability analysis; Obtaining the calculation results of the simulation calculation; According to the calculation results, calculate the electromagnetic field distribution; Evaluate electromagnetic compatibility (EMC) and electromagnetic radiation based on electromagnetic field distribution based on finite element analysis (FEA) and / or finite difference time domain method (FDTD); Combine the calculation results with the thermal simulation to perform thermal coupling simulation and obtain thermal coupling simulation results; Determine the thermal stability of the target substation based on the thermal coupling simulation results; The calculation results, electromagnetic compatibility, electromagnetic radiation and thermal stability of the target substation are taken together as simulation results.

6. A substation digital twin method based on large-scale point cloud as claimed in claim 1, characterized in that: Get the exception level interval set, including: Obtain target abnormal data; Analyze target abnormal data and obtain abnormal comparison data; Based on the preset clustering rules, the abnormal comparison data is clustered according to the abnormal probabilities of the abnormal items corresponding to the abnormal comparison data to obtain multiple abnormal comparison data clusters; Determine the abnormal level interval sequence of the abnormal comparison data clustering cluster according to the size order of the median of the abnormal probability corresponding to the abnormal comparison data clustering cluster; The abnormal comparison data in the abnormal comparison data cluster corresponding to each abnormal level interval serial number are continuousized to obtain the abnormal level interval corresponding to each abnormal level interval serial number; Integrate the abnormal level intervals to obtain an abnormal level interval set; Among them, obtaining target abnormal data includes: Acquire the first abnormal data in the history of the target substation, and at the same time, acquire the second abnormal data of the target substation based on the big data technology, the second abnormal data being: the abnormal data of the substation with similar attributes to the target substation acquired by the big data; The first abnormal data and the second abnormal data are used together as target abnormal data.

7. A substation digital twin method based on large-scale point cloud as claimed in claim 1, characterized in that: Real-time monitoring of the operating status of the target substation and fault detection, including: Determine the substation components that need to be fault-detected and use them as target components; Obtaining a comparison sub-result of the twin data of the target component and the actual data of the target component for data comparison; According to the comparison results, determine the detection equipment and obtain the location of the detection equipment; Acquire a first position of a target component corresponding to a substation; Obtain the second position of the maintenance personnel; Based on the regional distribution map of the target substation, a plurality of travel routes are planned according to the detection equipment location, the first location, and the second location; Determine the travel cost of each maintenance worker according to the regional attributes of the travel area of ​​the travel route; The maintenance personnel with the lowest travel cost are selected as target maintenance personnel, and the target maintenance personnel are notified to carry detection equipment to perform corresponding fault detection; Among them, multiple travel routes are planned, including: Obtain the first inertia weight, the maximum number of iterations and the second inertia weight of the particle swarm algorithm to search for the route; Based on the target distribution generator, the improved inertia weight is determined according to the first inertia weight, the maximum number of iterations and the second inertia weight. The determination model is as follows: Among them, γ is the improved inertia weight, γ min is the second inertia weight, γ max is the first inertia weight, exp(*) is the exponential function with e as the base, count is the current number of iterations, count max is the maximum number of iterations, betarnd is the target distribution generator; The travel route is planned based on the particle swarm algorithm with improved inertia weight.

8. A substation digital twin device based on large-scale point cloud, characterized in that: include: A target point cloud data acquisition subsystem is used to acquire target point cloud data of a target substation; The target node loading subsystem is used to load the target node according to the target point cloud data based on the large-scale point cloud rendering technology; The process digital twin model acquisition subsystem is used to perform data twinning on the target point cloud data through the target node to obtain the process digital twin model; The target digital twin model acquisition subsystem is used to optimize the point cloud of the process digital twin model and obtain the target digital twin model; The substation digital twin device also performs the following operations: Compare and monitor the twin data of the target digital twin model and the actual data of the target substation, monitor the operating status of the target substation in real time, and perform fault detection; Among them, the twin data of the target digital twin model and the actual data of the target substation are compared and monitored, including: Obtaining comparison results of data comparison; Get the abnormal level interval set; Determine that the comparison result falls within an abnormal level interval in the abnormal level interval set; Obtaining a target marked range of the first simulated power transformation component corresponding to the comparison result; According to the marking rules of the abnormal level range, the comparison results are marked within the corresponding target marking range; Wherein, obtaining the target marking range of the first simulated power transformation component corresponding to the comparison result includes: Acquire a second simulated power conversion component adjacent to the first simulated power conversion component, and calculate a component distance between the first simulated power conversion component and the second simulated power conversion component; The target marking range is determined according to the component distance and the simulated display range of the first simulated power transformation component.

9. A substation digital twin device based on large-scale point cloud as claimed in claim 8, characterized in that: The target node loads subsystems, including: An attribute acquisition module is used to acquire the structural attributes and functional attributes of the target substation; A first target feature extraction module is used to learn the first target feature of the substation component based on the structural attribute and the functional attribute based on the machine learning technology, and the first target feature includes: the connection relationship between the substation components and the function of the substation components; A data segmentation module is used to segment the target point cloud data according to the first target feature to obtain segmentation data, where the segmentation data is: point cloud data corresponding to different components obtained by segmenting the target point cloud data according to the structure and function of the substation; A second target feature extraction module, used to extract the second target feature of the segmented data, the second target feature being: the volume, shape and curvature of the transformer component corresponding to the segmented data; A third target feature determination module is used to determine a third target feature according to the relative position of the segmented data in the target point cloud data, where the third target feature is: a connection relationship and an adjacency relationship between the substation components corresponding to the segmented data; The target node loading module is used to load the target node according to the second target feature and the third target feature.

Images