Fault diagnosis method and system based on FTU of distribution network

Abstract

The present invention relates to the technical field of distribution network fault diagnosis, and in particular to a fault diagnosis method and system based on distribution network FTU. The method comprises the following steps: acquiring real-time FTU voltage monitoring data and distribution network FTU location data; performing time series window decomposition on the real-time FTU voltage monitoring data to obtain an FTU voltage sequence; performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data; performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; performing three-dimensional feature space mapping on the frequency components of multiple frequency bands to construct an FTU three-dimensional feature mapping diagram; performing dynamic voltage change analysis on the optimized multi-frequency domain voltage data to generate dynamic voltage fluctuation trend data; obtaining FTU spatial distribution data based on distribution network FTU location data; and constructing an FTU topology network model based on FTU spatial distribution data. The present invention realizes efficient and accurate distribution network fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault diagnosis, and in particular to a fault diagnosis method and system based on a distribution network FTU. Background Art

[0002] With the development of power systems and the improvement of their intelligence levels, distribution networks play a vital role in modern industrial and civil fields. However, due to the influence of various internal and external factors, such as changes in power load, equipment aging, short-circuit faults, etc., distribution networks will face various faults and abnormal conditions. If these faults and abnormal conditions cannot be discovered and diagnosed in time, they will lead to unstable operation of the power system and even cause accidents. Traditional distribution network fault detection methods often have the problems of low fault diagnosis efficiency and inaccuracy. Distribution network FTU (Fault Terminal Unit) as a terminal device in the distribution network can obtain power parameter information such as current and voltage in real time, and transmit this information to the upper system through the communication network for analysis and processing. Therefore, a method for distribution network fault diagnosis through distribution network FTU is needed. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a fault diagnosis method and system based on a distribution network FTU to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a fault diagnosis method based on a distribution network FTU, comprising the following steps:

[0005] Step S1: acquiring real-time FTU voltage monitoring data and distribution network FTU location data; performing time series window decomposition on the real-time FTU voltage monitoring data to obtain an FTU voltage sequence; performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data;

[0006] Step S2: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; performing three-dimensional feature space mapping on the frequency components of the multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU;

[0007] Step S3: Perform dynamic voltage change analysis on the optimized multi-frequency domain voltage data to generate dynamic voltage fluctuation trend data; obtain FTU spatial distribution data based on the FTU location data of the distribution network; and construct an FTU topology network model based on the FTU spatial distribution data;

[0008] Step S4: obtaining a topological coupling correlation strength value based on the dynamic voltage fluctuation trend data and the FTU topological network model; dynamically rendering and evolving the dynamic topological network model based on the topological coupling correlation strength value and the FTU three-dimensional feature map, and constructing an FTU nonlinear state space model;

[0009] Step S5: performing abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; performing time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data;

[0010] Step S6: Mark the faulty FTU node based on the FTU abnormal time series probability data; perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0011] The present invention can obtain the current status information of the distribution network by acquiring real-time FTU voltage monitoring data and location data, and provide basic data for fault detection. The time series window decomposition can segment the real-time FTU voltage data, extract the FTU voltage sequence, and provide continuous voltage data for subsequent analysis. The discrete wavelet multi-band decomposition can decompose the voltage sequence into components of different frequency bands, so that subsequent analysis can perform more refined processing and analysis on different frequency bands. Frequency component analysis can help identify the frequency characteristics of different frequency bands and reveal the frequency anomalies or resonance phenomena in the distribution network. The three-dimensional feature space mapping maps the frequency components of multiple frequency bands into three-dimensional space, and can correlate the characteristics of different frequency bands. , forming a more comprehensive feature description, constructing FTU three-dimensional feature mapping diagrams helps to visualize and understand the characteristic distribution of different frequency bands in the distribution network, and provides an intuitive reference for subsequent fault detection. Dynamic voltage change analysis can capture the instantaneous changes and fluctuations of voltage in the distribution network, which helps to find abnormal or unstable voltage. Generating dynamic voltage fluctuation trend data can help identify periodic or trend changes in voltage and provide more feature information for fault detection. FTU spatial distribution data and FTU topological network models can describe the location and connection relationship of different FTU devices in the distribution network, and provide a reference for the topological structure for subsequent fault detection. Based on dynamic voltage fluctuation trend data and FTU topological network models, the The topological coupling strength between different FTU nodes reveals the degree of mutual influence between nodes. Dynamic rendering evolution can combine the topological coupling strength value with the FTU three-dimensional feature map to form a dynamic topological network model, which can better reflect the state evolution and mutual relationship between nodes in the distribution network. Constructing the FTU nonlinear state space model helps to model and analyze the nonlinear behavior in the distribution network and provide a more accurate model basis for fault detection. Abnormal space feature analysis can extract abnormal features in the FTU nonlinear state space model to help identify abnormal behaviors or fault conditions in the distribution network. Time series probability association analysis can associate abnormal feature data with time series to determine the occurrence of abnormal features. The relationship between probability and timing provides a more reliable basis for fault detection. Generating FTU abnormal timing probability data can provide probability estimation and timing information of fault occurrence, which helps to optimize the accuracy and timeliness of fault detection. Marking faulty FTU nodes can determine the nodes where the fault occurs based on the FTU abnormal timing probability data, accurately locate the fault location, and improve the accuracy of fault diagnosis. Adaptive fault decision-making can take corresponding fault handling strategies based on the characteristics and context information of the faulty FTU nodes to improve the efficiency and reliability of fault handling. Constructing a distribution network fault decision-making model can integrate fault detection, location and processing processes to form a closed-loop system, providing a comprehensive solution for fault detection and maintenance of distribution networks.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: Acquire real-time FTU voltage monitoring data and distribution network FTU location data;

[0014] Step S12: performing time series window decomposition on the real-time FTU voltage monitoring data to obtain FTU voltage time series data;

[0015] Step S13: Serializing the FTU voltage timing data to obtain an FTU voltage sequence;

[0016] Step S14: performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain multi-frequency domain voltage data;

[0017] Step S15: performing high frequency noise elimination processing on the multi-frequency domain voltage data to obtain optimized multi-frequency domain voltage data.

[0018] The present invention provides voltage information of each node of the current distribution network through real-time FTU voltage monitoring data, which can reflect the operating status and load conditions of the power grid. The distribution network FTU position data provides the position information of each FTU device in the distribution network, which helps to determine the connection relationship and voltage transmission path between devices. The timing window decomposition processes the real-time FTU voltage monitoring data in segments to obtain continuous voltage timing data, so that subsequent analysis can be based on continuous data. The FTU voltage timing data reflects the changing trend and fluctuation of the voltage, which helps to judge and locate voltage anomalies in fault detection. The serialization processing converts the FTU voltage timing data into a sequence, so that the voltage data can be more conveniently analyzed. It is convenient to analyze and process. The FTU voltage sequence provides an orderly arrangement of voltage data, which is helpful to discover periodic changes, trend changes or other characteristics of voltage. Discrete wavelet multi-band decomposition decomposes the FTU voltage sequence into components of different frequency bands, which can separate and extract the frequency domain characteristics of voltage. Multi-frequency domain voltage data provides voltage information on different frequency bands, which is helpful to detect and analyze abnormal conditions in different frequency ranges, such as harmonics, interference, etc. High-frequency noise removal processing can remove high-frequency noise components in multi-frequency domain voltage data, improve data quality and accuracy, and optimize multi-frequency domain voltage data to eliminate noise interference that affects analysis and detection results, making subsequent fault detection more accurate and reliable.

[0019] Preferably, the specific steps of step S2 are:

[0020] Step S21: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands;

[0021] Step S22: performing frequency component distribution analysis on the frequency components of the multiple frequency bands to obtain frequency domain distribution data;

[0022] Step S23: Perform three-dimensional feature space mapping on the frequency components of multiple frequency bands based on the FTU voltage time series data and the frequency domain distribution data to construct an FTU three-dimensional feature mapping diagram.

[0023] The present invention further decomposes the optimized multi-frequency domain voltage data through frequency component analysis, and extracts frequency components in different frequency bands. The frequency components of multiple frequency bands reflect the fluctuation and characteristics of voltage in different frequency ranges, which is helpful to judge and analyze the abnormality in a specific frequency range in fault detection. The frequency component distribution analysis performs statistical and distribution analysis on the frequency components of multiple frequency bands to obtain frequency domain distribution data. The frequency domain distribution data provides information such as the occurrence frequency and amplitude distribution of different frequency components, which is helpful to discover the abnormal distribution and change trend of frequency components, and further guide fault detection. The three-dimensional feature space mapping combines the FTU voltage time series data and the frequency domain distribution data, and maps the frequency components of multiple frequency bands into a three-dimensional feature space. The FTU three-dimensional feature mapping diagram presents the frequency components in different frequency bands in the form of points or areas in space, which is helpful to visualize the analysis and understanding of the time domain and frequency domain characteristics of the voltage, and quickly discover abnormalities and fault signals.

[0024] Preferably, the specific steps of step S23 are:

[0025] Obtaining first coordinate axis data based on FTU voltage timing data;

[0026] Obtaining second coordinate axis data based on frequency domain distribution data;

[0027] Obtaining third coordinate axis data based on frequency components of multiple frequency bands;

[0028] Construct a three-dimensional coordinate system;

[0029] Based on the first coordinate axis data, the second coordinate axis data and the third coordinate axis data, the three-dimensional coordinate system is mapped into a three-dimensional feature space to construct a three-dimensional feature mapping diagram of the FTU.

[0030] The present invention can extract the time domain characteristics of the voltage, such as mean, variance, peak-to-peak value, etc., according to the FTU voltage time series data. The first coordinate axis data is constructed based on these time domain characteristics and is used to represent the characteristic information of the voltage signal in the time domain. The frequency domain distribution data reflects the component distribution of the voltage signal in different frequency ranges, such as the occurrence frequency and amplitude of the frequency components. The second coordinate axis data is constructed based on these frequency domain distribution information and is used to represent the characteristic information of the voltage signal in the frequency domain. The frequency components of multiple frequency bands provide the fluctuation and characteristics of the voltage signal in different frequency ranges. The third coordinate axis data is constructed based on these frequency components and is used to represent the characteristic information of the voltage signal in the frequency domain. To represent the characteristic information of the voltage signal in the frequency domain, the three-dimensional coordinate system is composed of the first, second and third coordinate axes, which is used to integrate the time domain characteristics, frequency domain distribution and frequency component characteristics of the voltage signal in one space. The three-dimensional coordinate system provides a comprehensive feature space, which is helpful to visualize and analyze information of different feature dimensions. The three-dimensional feature space mapping maps the first, second and third coordinate axis data based on the FTU voltage timing data, frequency domain distribution data and frequency component into a three-dimensional space. The FTU three-dimensional feature mapping diagram shows the comprehensive characteristics of different voltage signals in the form of points or areas, which is helpful for visual analysis and detection of voltage anomalies and fault conditions.

[0031] Preferably, the specific steps of step S3 are:

[0032] Step S31: performing dynamic voltage change analysis on the optimized multi-frequency domain voltage data to obtain FTU dynamic voltage change data;

[0033] Step S32: Perform voltage fluctuation trend analysis on the FTU dynamic voltage change data to generate dynamic voltage fluctuation trend data;

[0034] Step S33: performing FTU spatial distribution analysis on the FTU location data of the distribution network to obtain FTU spatial distribution data;

[0035] Step S34: performing FTU topology structure analysis on the FTU spatial distribution data to generate FTU topology structure data; and constructing an FTU topology network model based on the FTU topology structure data.

[0036] The present invention monitors and analyzes the optimized multi-frequency domain voltage data in real time through dynamic voltage change analysis to obtain the dynamic change of voltage signal. The FTU dynamic voltage change data reflects the fluctuation and change trend of voltage signal over time, which is helpful for identifying and analyzing abnormal voltage behavior in fault detection. The voltage fluctuation trend analysis is based on the FTU dynamic voltage change data, and the trend analysis and statistics of the fluctuation of voltage signal are performed. The dynamic voltage fluctuation trend data provides information such as the amplitude and frequency of voltage fluctuation, which is helpful for discovering the periodicity, trend and abnormality of voltage fluctuation, and provides a reference basis for fault detection. The FTU spatial distribution analysis is based on the location data of each FTU in the distribution network, and the spatial distribution of FTU is analyzed and counted. The FTU spatial distribution data provides the location of FTU. The information such as the location coordinates and regional distribution is helpful to understand the spatial layout of FTU and discover the existing fault clustering areas or abnormal situations. The FTU topology structure analysis is based on the FTU spatial distribution data to analyze and model the connection relationship and topological structure between FTUs. The FTU topology structure data describes the topological connection relationship between FTUs, such as line connection, proximity relationship, etc., which helps to understand the topological structure of the distribution network and provides a basis for fault detection and network analysis. The FTU topology network model is a network representation built based on the FTU topology structure data, which is used to simulate and analyze the relationship and communication between FTUs in the distribution network. The FTU topology network model provides a comprehensive perspective, which can help understand the information transmission and interaction between FTUs in the distribution network and provide support for fault detection and network management.

[0037] Preferably, the specific steps of step S4 are:

[0038] Step S41: performing regional load fluctuation calculation on the dynamic voltage fluctuation trend data to obtain regional load fluctuation data;

[0039] Step S42: performing a topological coupling strength analysis on the FTU topological network model based on the regional dynamic load fluctuation data to obtain a topological coupling association strength value;

[0040] Step S43: performing dynamic topology network adjustment on the FTU topology network model based on the topology coupling association strength value to obtain a dynamic topology network model;

[0041] Step S44: dynamically render and evolve the dynamic topology network model according to the FTU three-dimensional feature map to construct an FTU nonlinear state space model.

[0042] The present invention analyzes the load fluctuations in different regions through regional load fluctuation calculation based on dynamic voltage fluctuation trend data. The regional load fluctuation data provides information such as the fluctuation amplitude and frequency of the loads in different regions, which is helpful for detecting load anomalies and faults. The topological coupling strength analysis is based on the regional dynamic load fluctuation data and the FTU topological network model to evaluate the strength of the topological coupling relationship between different FTUs. The topological coupling association strength value provides the topological coupling degree between different FTUs, which is helpful for discovering topological relationship anomalies and faults. The dynamic topological network adjustment is based on the topological coupling association strength value, and the FTU topological network model is adjusted according to the change of the topological relationship. The dynamic topological network model reflects the real-time topological relationship between FTUs, which is helpful for real-time monitoring and management of topological changes in the distribution network. According to the FTU three-dimensional feature mapping diagram, the dynamic topological network model is dynamically rendered and evolved to generate an FTU nonlinear state space model. The FTU nonlinear state space model can more intuitively represent the state and change of the dynamic topological network model, which is helpful for fault detection and analysis of the nonlinear behavior of the distribution network.

[0043] Preferably, the specific steps of step S5 are:

[0044] Step S51: performing abnormal space feature analysis on the FTU nonlinear state space model to obtain abnormal feature data;

[0045] Step S52: performing time series recognition on the abnormal feature data to mark the abnormal feature time series data;

[0046] Step S53: performing abnormal probability distribution statistics on the abnormal feature data, thereby obtaining abnormal feature probability distribution data;

[0047] Step S54: Perform time series probability association analysis on the abnormal feature probability distribution data based on the abnormal feature time series data, thereby generating FTU abnormal time series probability data.

[0048] The present invention analyzes and extracts abnormal spatial features in the model based on the FTU nonlinear state space model through abnormal spatial feature analysis. Abnormal feature data contains abnormal feature information in the FTU nonlinear state space model, which is helpful for fault detection and identification of abnormal behavior. Time series recognition analyzes the abnormal feature data to identify the time series pattern and abnormal features therein. Marking abnormal feature time series data can help identify abnormal behavior and abnormal moments in the FTU nonlinear state space model. Abnormal probability distribution statistics are based on abnormal feature data to perform statistics and analysis on the probability distribution of abnormal features. Abnormal feature probability distribution data provides probability information of the occurrence of abnormal features, which helps to determine the severity and nature of the abnormality. Time series probability association analysis is based on abnormal feature time series data and abnormal feature probability distribution data to analyze the time series correlation and probability distribution changes between them. FTU abnormal time series probability data provides the evolution and change of abnormal features in time, which helps to monitor and analyze the abnormal state of FTU in real time.

[0049] Preferably, the specific steps of step S51 are:

[0050] Perform FTU operation simulation on the FTU nonlinear state space model to obtain the FTU operation simulation trajectory;

[0051] Based on a preset normal FTU state trajectory database, trajectory deviation identification is performed on the FTU operation simulation trajectory to obtain a deviation trajectory;

[0052] Based on the deviation trajectory, the abnormal area of ​​the FTU nonlinear state space model is marked to obtain the abnormal area of ​​the FTU state;

[0053] Reconstruct the local state of the abnormal area of ​​FTU state to obtain the local model of the abnormal area;

[0054] Perform residual calculation on the local model of the abnormal area to generate local deviation error;

[0055] Perform PCA decomposition on the local deviation error to obtain a local anomaly vector;

[0056] Perform abnormal spatial feature analysis on the local abnormal vector to obtain abnormal feature data.

[0057] The present invention simulates the operation trajectory of the FTU under normal conditions based on a nonlinear state space model through FTU operation simulation. The FTU operation simulation trajectory provides the expected behavior and state change information of the FTU under normal conditions, and provides a benchmark for subsequent abnormality detection and analysis. The deviation identification is based on a preset normal FTU state trajectory database, and the difference between the FTU operation simulation trajectory and the normal trajectory is compared. The deviation trajectory identifies the part of the FTU operation simulation trajectory that is significantly deviated from the normal situation, which is helpful to locate potential abnormal areas. The abnormal area marking is based on the deviation trajectory. The area corresponding to the deviation trajectory in the FTU nonlinear state space model is marked as an abnormal area. The FTU state abnormal area provides specific area information where the abnormality occurs in the operation simulation, which is helpful for further abnormality analysis and fault detection. The local state reconstruction is based on the FTU state abnormal area , reconstruct the local state model in the abnormal area. The local model of the abnormal area provides local state information in the abnormal area, which helps to deeply understand and analyze the nature and cause of the abnormality. The residual calculation calculates the local deviation error by comparing the difference between the local model of the abnormal area and the actual observation data. The local deviation error quantifies the degree of difference between the model and the observation data in the abnormal area, which helps to judge the degree and type of the abnormality. The PCA decomposition process converts the local deviation error into the form of principal component analysis and extracts the local anomaly vector. The local anomaly vector provides key feature information in the abnormal area, which helps to detect anomalies and analyze abnormal features. The abnormal spatial feature analysis is based on the local anomaly vector to analyze and extract the abnormal features. The abnormal feature data contains the key feature information in the local anomaly vector, which is used for fault detection and abnormal behavior analysis of the distribution network.

[0058] Preferably, the specific steps of step S6 are:

[0059] Step S61: performing fault diagnosis analysis on the FTU abnormal timing probability data to generate FTU fault diagnosis data;

[0060] Step S62: locating the faulty node in the FTU topology network model based on the FTU fault diagnosis data, and marking the faulty FTU node;

[0061] Step S63: Perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0062] The present invention uses fault diagnosis analysis based on FTU abnormal time series probability data to determine whether an FTU has a fault by analyzing abnormal characteristics and time series data. The FTU fault diagnosis data provides the result of fault detection, identifies the FTU with a fault and related information, and provides a basis for subsequent fault processing and repair. The fault node is located based on the FTU fault diagnosis data and the topological network model of the distribution network to determine the location of the fault. The marking of the faulty FTU node provides information on the fault location, helps to accurately locate the specific location of the fault, and is helpful for subsequent troubleshooting and repair work. The adaptive fault decision is based on the diagnosis result of the faulty FTU node and the status information of the distribution network. The corresponding fault processing plan is formulated according to the specific situation. The distribution network fault decision model is constructed to provide response strategies and decision support under different fault conditions, which is helpful to quickly restore and ensure the normal operation of the distribution network.

[0063] In this specification, a fault diagnosis system based on a distribution network FTU is provided, which is used to execute the fault diagnosis method based on the distribution network FTU as described above, including:

[0064] The multi-band decomposition module is used to obtain real-time FTU voltage monitoring data and distribution network FTU location data; perform time series window decomposition on the real-time FTU voltage monitoring data to obtain the FTU voltage sequence; perform discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data;

[0065] The three-dimensional feature module is used to perform frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; perform three-dimensional feature space mapping on the frequency components of multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU;

[0066] The topology network module is used to analyze the dynamic voltage changes of the optimized multi-frequency domain voltage data and generate dynamic voltage fluctuation trend data; obtain the FTU spatial distribution data based on the FTU location data of the distribution network; and construct the FTU topology network model based on the FTU spatial distribution data;

[0067] Dynamic rendering module, used to obtain the topological coupling correlation strength value based on dynamic voltage fluctuation trend data and FTU topological network model; dynamically render and evolve the dynamic topological network model based on the topological coupling correlation strength value and FTU three-dimensional feature map, and construct the FTU nonlinear state space model;

[0068] The abnormal feature module is used to perform abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; perform time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data;

[0069] The fault decision module is used to mark the faulty FTU node based on the FTU abnormal time series probability data; and to make an adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0070] The present invention obtains FTU voltage monitoring data and distribution network FTU position data in real time to provide necessary input data for subsequent fault detection. Through time series window decomposition, the real-time FTU voltage monitoring data is converted into a continuous FTU voltage sequence, which is convenient for subsequent frequency component analysis and dynamic voltage change analysis. Discrete wavelet multi-band decomposition is used to obtain optimized multi-frequency domain voltage data, and more abundant frequency domain information is extracted to provide more accurate features for fault detection. Through frequency component analysis, frequency components of different frequency bands in the optimized multi-frequency domain voltage data are extracted to obtain the frequency characteristics of the voltage signal. Three-dimensional feature space mapping is performed based on the frequency components. By mapping the frequency components to the three-dimensional space, an FTU three-dimensional feature mapping diagram is constructed, so that the voltage characteristics are easier to visualize and analyze. Through dynamic voltage change analysis, fluctuation trend information in the optimized multi-frequency domain voltage data is extracted to obtain the dynamic characteristics of the voltage signal. The distribution network FTU position data is used to obtain the distribution of the FTU in space, which provides basic data for subsequent topological association analysis. An FTU topological network model is constructed based on the FTU spatial distribution data, and a topological relationship between FTU nodes is established to provide topological information for fault detection. The pressure fluctuation trend data and FTU topology network model are used to calculate the topology coupling correlation strength value, evaluate the topological connection degree between FTU nodes, and use the topology coupling correlation strength value and FTU three-dimensional feature map to dynamically render and evolve the dynamic topology network model, visualize the topological state of the node, and build the FTU nonlinear state space model to facilitate fault detection and analysis. By performing abnormal spatial feature analysis on the FTU nonlinear state space model, abnormal feature data that does not conform to the normal situation is identified, which is helpful for fault detection and location. The time series probability association analysis of the abnormal feature data can determine the correlation degree and probability distribution between the abnormal features, and generate FTU abnormal time series probability data for further fault diagnosis and decision-making. The fault decision module marks the faulty FTU node based on the FTU abnormal time series probability data, and makes adaptive fault decisions to build a distribution network fault decision model. According to the FTU abnormal time series probability data, the FTU node with fault is determined and marked. Based on the marked faulty FTU node and the status information of the distribution network, a corresponding fault handling plan is formulated to build a distribution network fault decision model, which helps to quickly restore and ensure the normal operation of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A schematic flow chart of the steps of a fault diagnosis method based on a distribution network FTU of the present invention;

[0072] Figure 2 Detailed implementation flow chart of step S1;

[0073] Figure 3 Detailed implementation flow chart of step S2;

[0074] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0075] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0076] The present application example provides a fault diagnosis method and system based on the distribution network FTU. The execution subject of the fault diagnosis method and system based on the distribution network FTU includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0077] See also Figures 1 to 4 The present invention provides a fault diagnosis method based on a distribution network FTU, and the fault diagnosis method based on a distribution network FTU comprises the following steps:

[0078] Step S1: acquiring real-time FTU voltage monitoring data and distribution network FTU location data; performing time series window decomposition on the real-time FTU voltage monitoring data to obtain an FTU voltage sequence; performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data;

[0079] Step S2: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; performing three-dimensional feature space mapping on the frequency components of the multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU;

[0080] Step S3: Perform dynamic voltage change analysis on the optimized multi-frequency domain voltage data to generate dynamic voltage fluctuation trend data; obtain FTU spatial distribution data based on the FTU location data of the distribution network; and construct an FTU topology network model based on the FTU spatial distribution data;

[0081] Step S4: obtaining a topological coupling correlation strength value based on the dynamic voltage fluctuation trend data and the FTU topological network model; dynamically rendering and evolving the dynamic topological network model based on the topological coupling correlation strength value and the FTU three-dimensional feature map, and constructing an FTU nonlinear state space model;

[0082] Step S5: performing abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; performing time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data;

[0083] Step S6: Mark the faulty FTU node based on the FTU abnormal time series probability data; perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0084] The present invention can obtain the current status information of the distribution network by acquiring real-time FTU voltage monitoring data and location data, and provide basic data for fault detection. The time series window decomposition can segment the real-time FTU voltage data, extract the FTU voltage sequence, and provide continuous voltage data for subsequent analysis. The discrete wavelet multi-band decomposition can decompose the voltage sequence into components of different frequency bands, so that subsequent analysis can perform more refined processing and analysis on different frequency bands. Frequency component analysis can help identify the frequency characteristics of different frequency bands and reveal the frequency anomalies or resonance phenomena in the distribution network. The three-dimensional feature space mapping maps the frequency components of multiple frequency bands into three-dimensional space, and can correlate the characteristics of different frequency bands. , forming a more comprehensive feature description, constructing FTU three-dimensional feature mapping diagrams helps to visualize and understand the characteristic distribution of different frequency bands in the distribution network, and provides an intuitive reference for subsequent fault detection. Dynamic voltage change analysis can capture the instantaneous changes and fluctuations of voltage in the distribution network, which helps to find abnormal or unstable voltage. Generating dynamic voltage fluctuation trend data can help identify periodic or trend changes in voltage and provide more feature information for fault detection. FTU spatial distribution data and FTU topological network models can describe the location and connection relationship of different FTU devices in the distribution network, and provide a reference for the topological structure for subsequent fault detection. Based on dynamic voltage fluctuation trend data and FTU topological network models, the The topological coupling strength between different FTU nodes reveals the degree of mutual influence between nodes. Dynamic rendering evolution can combine the topological coupling strength value with the FTU three-dimensional feature map to form a dynamic topological network model, which can better reflect the state evolution and mutual relationship between nodes in the distribution network. Constructing the FTU nonlinear state space model helps to model and analyze the nonlinear behavior in the distribution network and provide a more accurate model basis for fault detection. Abnormal space feature analysis can extract abnormal features in the FTU nonlinear state space model to help identify abnormal behaviors or fault conditions in the distribution network. Time series probability association analysis can associate abnormal feature data with time series to determine the occurrence of abnormal features. The relationship between probability and timing provides a more reliable basis for fault detection. Generating FTU abnormal timing probability data can provide probability estimation and timing information of fault occurrence, which helps to optimize the accuracy and timeliness of fault detection. Marking faulty FTU nodes can determine the nodes where the fault occurs based on the FTU abnormal timing probability data, accurately locate the fault location, and improve the accuracy of fault diagnosis. Adaptive fault decision-making can take corresponding fault handling strategies based on the characteristics and context information of the faulty FTU nodes to improve the efficiency and reliability of fault handling. Constructing a distribution network fault decision-making model can integrate fault detection, location and processing processes to form a closed-loop system, providing a comprehensive solution for fault detection and maintenance of distribution networks.

[0085] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of a method for fault diagnosis based on a distribution network FTU of the present invention. In this example, the method for fault diagnosis based on a distribution network FTU includes the following steps:

[0086] Step S1: acquiring real-time FTU voltage monitoring data and distribution network FTU location data; performing time series window decomposition on the real-time FTU voltage monitoring data to obtain an FTU voltage sequence; performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data;

[0087] In this embodiment, after obtaining authorization, real-time FTU voltage monitoring data and FTU location data are obtained from the distribution network. Sensors, monitoring equipment or data acquisition systems can be used to obtain these data to ensure the accuracy and real-time nature of the data. An appropriate time window size is selected, such as 1 second or less, and the time series data is divided into continuous windows. The data in each window can be regarded as a time series sample. Discrete Wavelet Transform (DWT) is a method for decomposing a signal into different frequency bands. Appropriate wavelet basis functions and decomposition levels are selected to perform wavelet decomposition on the voltage data. Based on the selected wavelet basis functions and decomposition levels, discrete wavelet multi-band decomposition is performed on the FTU voltage data in each time series window. Optimized multi-frequency domain voltage data is extracted from the results of the discrete wavelet multi-band decomposition. Wavelet coefficients of specific frequency bands can be selected and reconstructed to obtain optimized multi-frequency domain voltage data.

[0088] Step S2: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; performing three-dimensional feature space mapping on the frequency components of the multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU;

[0089] In this embodiment, the original multi-frequency domain voltage data is subjected to an inverse fluctuation transformation, such as a fast Fourier transform (FFT), to extract the frequency component signal corresponding to each frequency segment, and the characteristic indicators of each frequency segment when the FTU is operating normally and abnormally, such as average value, peak value, power, etc., are selected to form a one-dimensional feature vector for each frequency segment. The feature vectors of each frequency segment are connected in series into a long feature vector to construct a multi-dimensional feature representation of the FTU at that time point. In a three-dimensional coordinate system, each dimensional component of the feature vector is assigned to the coordinate axis, and the feature point is projected into the three-dimensional space as a point. Each feature point is listed in chronological order and connected into multiple curves to form a dynamic simulation trajectory diagram of the FTU in the three-dimensional space. The mapping diagram is analyzed for regularities, abnormal points or segments are identified, and abnormal segmentation marks, such as intervals, step points, etc., are defined. By mining the inherent laws and structures in the three-dimensional mapping diagram, the corresponding position and morphological change laws of the FTU under different operating conditions are identified.

[0090] Step S3: Perform dynamic voltage change analysis on the optimized multi-frequency domain voltage data to generate dynamic voltage fluctuation trend data; obtain FTU spatial distribution data based on the FTU location data of the distribution network; and construct an FTU topology network model based on the FTU spatial distribution data;

[0091] In this embodiment, the optimized voltage data sequence of each frequency band is analyzed for time domain fluctuation characteristics, such as extracting characteristic parameters such as voltage fluctuation rate and wave state coefficient, comparing the changing rules of frequency band characteristic parameters within the time window, generating dynamic change trend data of frequency band voltage of each FTU device, and drawing the precise location distribution data of FTU in the distribution network area through tools such as GIS according to the spatial displacement under different time windows. According to the spatial location data of FTU and the topological structure of distribution lines, a distribution network topology network diagram marked with FTU location identification is constructed, and a network diagram representation method is adopted to represent the FTU position and connection relationship with nodes and edges. Each FTU corresponds to a node, and additional node attributes such as FTU number and type, and edge attributes such as line type and length are added. The distance between FTU nodes and logical distance and other supplementary network attributes are further counted to provide an FTU logical topology model, and the dynamically changing data is continuously updated to synchronize FTU node attributes to construct a distribution network dynamic topology model.

[0092] Step S4: obtaining a topological coupling correlation strength value based on the dynamic voltage fluctuation trend data and the FTU topological network model; dynamically rendering and evolving the dynamic topological network model based on the topological coupling correlation strength value and the FTU three-dimensional feature map, and constructing an FTU nonlinear state space model;

[0093] In this embodiment, a dynamic time window is used for each FTU node, and the correlation coefficient matrix between its frequency band voltage and the frequency band voltage of the adjacent node is counted. Based on the correlation coefficient matrix, an algorithm model for measuring the topological coupling strength, such as angular similarity, is proposed. The topological coupling strength value between each pair of FTU nodes is counted and dynamically updated according to time, and the coupling strength value is attached to the topological network model to present the dynamic correlation strength between FTU nodes. For each FTU node, the positioning point of its three-dimensional feature space is used for dynamic spatial rendering. Combined with the coupling strength dynamics between nodes, the nonlinear evolutionary motion law of the rendering point group is realized. The non-balanced cluster aggregation process of the FTU network in the three-dimensional feature space is displayed in the form of animation, and the spatial state transformation dynamics are continuously updated to construct a nonlinear dynamic state space model of the FTU node group of the distribution network.

[0094] Step S5: performing abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; performing time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data;

[0095] In this embodiment, the FTU operation state is modeled as a nonlinear state space model, including a state equation and an observation equation. Nonlinear filtering technology, such as an extended Kalman filter (EKF) or a módulo-free particle filter (UPF), is used to perform state estimation on the model. Based on the state estimation results, the abnormal spatial characteristics of the FTU operation process, such as the distribution characteristics and correlation of abnormal state variables, are analyzed. A time series analysis method, such as an autoregressive moving average (ARMA) model and a hidden Markov model (HMM), is used to establish a time series probability model of the abnormal feature vector. The time series probability association relationship between the abnormal features is obtained through learning and inference of the model parameters.

[0096] Step S6: Mark the faulty FTU node based on the FTU abnormal time series probability data; perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0097] In this embodiment, according to the specific situation, a suitable threshold is set to determine whether the FTU node fails. The threshold can be obtained based on experience or statistical analysis. For FTU nodes whose abnormal probability exceeds the threshold, they are marked to indicate that the node has failed. For the FTU nodes marked as faulty, their characteristics are analyzed, including the changing trend of the abnormal timing probability, the status of the surrounding nodes, etc. According to the characteristics of the faulty FTU node and the status information of the distribution network, a corresponding fault handling plan is formulated, which includes switching to a backup power supply, adjusting load distribution, restarting the faulty equipment, etc. The fault handling plan is recorded and associated with the distribution network status information to form a fault decision model of the distribution network. This model can be used for adaptive decision-making in similar fault situations.

[0098] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0099] Step S11: Acquire real-time FTU voltage monitoring data and distribution network FTU location data;

[0100] Step S12: performing time series window decomposition on the real-time FTU voltage monitoring data to obtain FTU voltage time series data;

[0101] Step S13: Serializing the FTU voltage timing data to obtain an FTU voltage sequence;

[0102] Step S14: performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain multi-frequency domain voltage data;

[0103] Step S15: performing high frequency noise elimination processing on the multi-frequency domain voltage data to obtain optimized multi-frequency domain voltage data.

[0104] In this embodiment, sensors or monitoring equipment are used to obtain real-time voltage monitoring data of FTUs. These data include information such as voltage values ​​and timestamps. The location information of each FTU in the distribution network is collected, including coordinates or numbers, which can be obtained through on-site surveys or other means. The window size used for time series window decomposition can be determined according to actual needs and data characteristics. The real-time FTU voltage monitoring data is slidingly decomposed according to the set window size to obtain a series of time series windows. Each window contains a continuous section of voltage data. The voltage value is extracted from each time series window to form a voltage sequence. For each data point in the voltage sequence, Time marking is performed to retain the corresponding time information, and a suitable discrete wavelet transform method, such as Daubechies wavelet, Haar wavelet, etc., is selected. The FTU voltage sequence is transformed by applying the selected discrete wavelet transform method to obtain the corresponding frequency domain coefficients. The frequency domain coefficients are decomposed into multiple frequency bands according to requirements. The decomposition can be performed by setting thresholds or using specific wavelet filters. The multi-frequency domain voltage data is analyzed and processed to identify the high-frequency noise components therein. According to the noise characteristics, filtering or other signal processing techniques are used to remove the high-frequency noise from the multi-frequency domain voltage data to obtain the optimized multi-frequency domain voltage data.

[0105] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0106] Step S21: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands;

[0107] Step S22: performing frequency component distribution analysis on the frequency components of the multiple frequency bands to obtain frequency domain distribution data;

[0108] Step S23: Perform three-dimensional feature space mapping on the frequency components of multiple frequency bands based on the FTU voltage time series data and the frequency domain distribution data to construct an FTU three-dimensional feature mapping diagram.

[0109] In this embodiment, Fourier transform is applied to the optimized multi-frequency domain voltage data to convert the time domain signal into a frequency domain signal, and multiple frequency components of interest are extracted from the frequency domain signal. Specific frequency components can be selected according to application requirements and domain knowledge. For the frequency components of each frequency segment, their distribution in the optimized multi-frequency domain voltage data is statistically analyzed. Histograms, probability density functions and other methods can be used for calculation. Key statistical indicators such as mean, variance, peak value, etc. are extracted from the frequency distribution results to represent the distribution characteristics of the frequency domain. The FTU voltage time series data and the frequency domain distribution data are used as features to construct a three-dimensional feature space, in which the first dimension represents time, the second dimension represents or frequency domain distribution characteristics, and the third dimension represents the energy of each frequency segment. The frequency components of each frequency segment and their corresponding time domain characteristics or frequency domain distribution characteristics are mapped in the feature space to form a three-dimensional feature mapping diagram.

[0110] In this embodiment, the specific steps of step S23 are:

[0111] Obtaining first coordinate axis data based on FTU voltage timing data;

[0112] Obtaining second coordinate axis data based on frequency domain distribution data;

[0113] Obtaining third coordinate axis data based on frequency components of multiple frequency bands;

[0114] Construct a three-dimensional coordinate system;

[0115] Based on the first coordinate axis data, the second coordinate axis data and the third coordinate axis data, the three-dimensional coordinate system is mapped into a three-dimensional feature space to construct a three-dimensional feature mapping diagram of the FTU.

[0116] In this embodiment, the first coordinate axis data is used as the X-axis, the second coordinate axis data is used as the Y-axis, and the third coordinate axis data is used as the Z-axis to construct a three-dimensional coordinate system. The first coordinate axis data of each sample point is used as the X-axis coordinate, the second coordinate axis data is used as the Y-axis coordinate, and the third coordinate axis data is used as the Z-axis coordinate. The sample points are mapped to corresponding positions on the three-dimensional coordinate system, and a scatter plot or other visualization methods can be used to represent the three-dimensional feature map.

[0117] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0118] Step S31: performing dynamic voltage change analysis on the optimized multi-frequency domain voltage data to obtain FTU dynamic voltage change data;

[0119] Step S32: Perform voltage fluctuation trend analysis on the FTU dynamic voltage change data to generate dynamic voltage fluctuation trend data;

[0120] Step S33: performing FTU spatial distribution analysis on the FTU location data of the distribution network to obtain FTU spatial distribution data;

[0121] Step S34: performing FTU topology structure analysis on the FTU spatial distribution data to generate FTU topology structure data; and constructing an FTU topology network model based on the FTU topology structure data.

[0122] In this embodiment, key moments or time periods are extracted from the optimized multi-frequency domain voltage data to capture the voltage changes. For each extracted moment or time period, the voltage change is calculated, which can be an absolute value or a relative value. The extracted moment or time period is used as the horizontal coordinate, and the corresponding voltage change is used as the vertical coordinate to construct FTU dynamic voltage change data. The FTU dynamic voltage change data is smoothed, and the smoothed data can be analyzed for fluctuation trend by using sliding window averaging, exponential weighted averaging, and other methods, such as calculating fluctuation range, fluctuation frequency, and other indicators to describe the voltage fluctuation characteristics. The smoothed data is used as the horizontal coordinate, and the corresponding fluctuation trend indicator is used as the vertical coordinate to construct dynamic voltage fluctuation trend data, and the spatial distribution analysis of FTU position data is performed. Spatial density statistics, cluster analysis and other methods can be used to understand the distribution of FTUs in the distribution network, map the FTU location data to the spatial map of the distribution network, represent the position corresponding to each FTU as a data point, and construct FTU spatial distribution data. Based on the FTU spatial distribution data, analyze the connectivity between each FTU, that is, determine whether there is an electrical connection between them. Perform topological structure analysis based on the connection relationship between FTUs, such as identifying structural features such as trunk lines, branches, and ring networks. Convert the results of the topological structure analysis into data representation. Adjacency matrices, edge lists and other forms can be used to represent the connection relationship between FTUs. Based on the FTU topological structure data, construct a topological network model of FTU, where FTU represents nodes and power connections represent edges.

[0123] In this embodiment, step S4 includes the following steps:

[0124] Step S41: performing regional load fluctuation calculation on the dynamic voltage fluctuation trend data to obtain regional load fluctuation data;

[0125] Step S42: performing a topological coupling strength analysis on the FTU topological network model based on the regional dynamic load fluctuation data to obtain a topological coupling association strength value;

[0126] Step S43: performing dynamic topology network adjustment on the FTU topology network model based on the topology coupling association strength value to obtain a dynamic topology network model;

[0127] Step S44: dynamically render and evolve the dynamic topology network model according to the FTU three-dimensional feature map to construct an FTU nonlinear state space model.

[0128] In this embodiment, the distribution network is divided into different areas according to actual conditions, and can be divided according to factors such as load characteristics and power connection. The dynamic voltage fluctuation trend data in each area is subjected to load fluctuation calculation, and the load fluctuation situation in the area can be described by calculating indicators such as fluctuation range, fluctuation frequency, and fluctuation index. The load fluctuation calculation results of each area are sorted into an appropriate data format, for example, using the area identifier as the horizontal coordinate and the load fluctuation index as the vertical coordinate to construct regional load fluctuation data. Based on the regional load fluctuation data, a topological coupling strength analysis is performed on the FTU topological network model. The topological coupling correlation strength between different areas can be evaluated by calculating indicators such as load fluctuation correlation and resonance index between areas. The results of the topological coupling strength analysis are sorted into an appropriate data format, for example, using the connection relationship between FTUs as the horizontal and vertical coordinates and the topological coupling correlation strength value as the weight to construct a topological coupling network model. Combined association strength value data is used to dynamically adjust the FTU topology network model based on the topology coupling association strength value. The connection relationship between FTUs can be adjusted according to the size of the topology coupling association strength value, and the connection weight can be increased or decreased to reflect the dynamic changes of the power system. The adjusted FTU topology network model is represented as a new data structure, in which FTU represents a node and the adjusted connection relationship represents an edge. A dynamic topology network model is constructed, and the dynamic topology network model is combined with the FTU three-dimensional feature map to perform dynamic rendering evolution. According to the changes of nodes and edges in the dynamic topology network model, the position of the FTU node in the three-dimensional space is updated to reflect the dynamic changes of the system. Based on the dynamic topology network model and the updated FTU three-dimensional feature map, a nonlinear state space model of FTU is constructed. This model can describe the interaction and state evolution between FTU nodes for further system analysis and simulation research.

[0129] In this embodiment, the specific steps of step S5 are:

[0130] Step S51: performing abnormal space feature analysis on the FTU nonlinear state space model to obtain abnormal feature data;

[0131] Step S52: performing time series recognition on the abnormal feature data to mark the abnormal feature time series data;

[0132] Step S53: performing abnormal probability distribution statistics on the abnormal feature data, thereby obtaining abnormal feature probability distribution data;

[0133] Step S54: Perform time series probability association analysis on the abnormal feature probability distribution data based on the abnormal feature time series data, thereby generating FTU abnormal time series probability data.

[0134] In this embodiment, the abnormal spatial feature analysis is performed on the FTU nonlinear state space model. The abnormal features of the nodes in the space can be described by calculating the distance, similarity, clustering and other indicators between the nodes. According to the results of the abnormal spatial feature analysis, data representing the abnormal features are extracted and sorted. These data may include abnormal feature values ​​and abnormal feature vectors of the nodes. Time series recognition is performed on the abnormal feature data. Various time series analysis methods, such as ARIMA model and exponential smoothing method, can be used to identify the time series patterns and trends in the abnormal feature data. According to the results of time series recognition, the abnormal time series data in the abnormal feature data are marked. Binary marking (normal / abnormal) or more detailed marking methods (such as using different abnormality degree indicators) can be used. To perform abnormal probability distribution statistics on abnormal feature data, statistical analysis methods such as histograms and kernel density estimation can be used to calculate the probability distribution of each abnormal feature. The statistical results of the probability distribution of abnormal features can be organized into an appropriate data format. The abnormal feature can be used as the horizontal axis and the probability distribution as the vertical axis to construct the abnormal feature probability distribution data. Based on the abnormal feature time series data, a time series probability association analysis can be performed on the abnormal feature probability distribution data. Time series analysis methods such as correlation analysis and cointegration analysis can be used to study the correlation between the abnormal feature time series data and the abnormal feature probability distribution. According to the results of the time series probability association analysis, the abnormal time series probability data of FTU is generated. These data can be used to represent the probability of abnormalities in FTU at different time points.

[0135] In this embodiment, the specific steps of step S51 are:

[0136] Perform FTU operation simulation on the FTU nonlinear state space model to obtain the FTU operation simulation trajectory;

[0137] Based on a preset normal FTU state trajectory database, trajectory deviation identification is performed on the FTU operation simulation trajectory to obtain a deviation trajectory;

[0138] Based on the deviation trajectory, the abnormal area of ​​the FTU nonlinear state space model is marked to obtain the abnormal area of ​​the FTU state;

[0139] Reconstruct the local state of the abnormal area of ​​FTU state to obtain the local model of the abnormal area;

[0140] Perform residual calculation on the local model of the abnormal area to generate local deviation error;

[0141] Perform PCA decomposition on the local deviation error to obtain a local anomaly vector;

[0142] Perform abnormal spatial feature analysis on the local abnormal vector to obtain abnormal feature data.

[0143] In this embodiment, a nonlinear state space model is used to simulate the operation of the FTU. The state value of each time point in the simulation trajectory is calculated step by step through the state update equation and observation equation of the model. The state value of each time point in the simulation trajectory is recorded to form an FTU operation simulation trajectory. The FTU operation simulation trajectory is compared with a normal FTU state trajectory database. Various trajectory similarity measurement methods, such as Euclidean distance, dynamic time warping, etc., can be used to evaluate the degree of deviation between the simulation trajectory and the normal trajectory. The FTU nonlinear state space model is marked for abnormal areas. Various methods, such as threshold setting based on statistical analysis, abnormal area division based on clustering, etc., can be used to mark the abnormal areas in the model corresponding to the deviated trajectory. According to the results of the abnormal area marking, the state abnormal areas of the FTU are obtained. These areas represent the abnormal conditions related to the deviated trajectory in the model. The local state of each abnormal area is measured. Reconstruction. Various methods can be used, such as state estimation based on sub-models, state reconstruction based on data interpolation, etc., to reconstruct the local state in the abnormal area. According to the results of local state reconstruction, local models of the abnormal area are obtained. These local models represent the state changes in the abnormal area. Local state reconstruction is performed on each abnormal area. Various methods can be used, such as state estimation based on sub-models, state reconstruction based on data interpolation, etc., to reconstruct the local state in the abnormal area. According to the results of local state reconstruction, local models of the abnormal area are obtained. These local models represent the state changes in the abnormal area. Residual calculation is performed on each local model. The actual state value in the abnormal area is compared with the state value predicted by the local model to obtain the residual, which represents the difference between the local model and the actual data. According to the results of residual calculation, local deviation errors are generated. These error data reflect the degree of deviation in the abnormal area.

[0144] In this embodiment, the specific steps of step S6 are:

[0145] Step S61: performing fault diagnosis analysis on the FTU abnormal timing probability data to generate FTU fault diagnosis data;

[0146] Step S62: locating the faulty node in the FTU topology network model based on the FTU fault diagnosis data, and marking the faulty FTU node;

[0147] Step S63: Perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0148] In this embodiment, meaningful features are extracted from the abnormal time series probability data, such as the time when the abnormality occurs, the duration of the abnormality, the amplitude of the abnormality, etc., and a fault diagnosis algorithm, such as rule-based diagnosis, machine learning algorithm, statistical method, etc., is applied to analyze and process the features to generate FTU fault diagnosis data. According to the results of the fault diagnosis analysis, FTU fault diagnosis data is generated. These data include information such as fault type, fault occurrence time, and fault location, which are used for subsequent fault node location and fault decision. The fault type in the FTU fault diagnosis data is matched with the fault mode in the topology network model to determine the node type with fault, and the node location is applied. Algorithms, such as those based on topology analysis, electrical feature analysis, and data-driven methods, are used to further locate the node types obtained by matching the fault types to determine the specific faulty nodes. Fault feature analysis is performed on each faulty FTU node, which may include analysis of the node status, electrical features, and status of surrounding nodes. Fault decision algorithms, such as rule-based decision making, optimization methods, and machine learning algorithms are applied to analyze and process the fault features to generate corresponding fault decision results. Based on the results of the fault decision, a distribution network fault decision model is constructed, which can be used to quickly respond to fault events and provide guidance and suggestions to support fault repair and operation recovery.

[0149] In this embodiment, a fault diagnosis system based on a distribution network FTU is provided, which is used to execute the fault diagnosis method based on the distribution network FTU as described above, including:

[0150] The multi-band decomposition module is used to obtain real-time FTU voltage monitoring data and distribution network FTU location data; perform time series window decomposition on the real-time FTU voltage monitoring data to obtain the FTU voltage sequence; perform discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data;

[0151] The three-dimensional feature module is used to perform frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; perform three-dimensional feature space mapping on the frequency components of multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU;

[0152] The topology network module is used to analyze the dynamic voltage changes of the optimized multi-frequency domain voltage data and generate dynamic voltage fluctuation trend data; obtain the FTU spatial distribution data based on the FTU location data of the distribution network; and construct the FTU topology network model based on the FTU spatial distribution data;

[0153] Dynamic rendering module, used to obtain the topological coupling correlation strength value based on dynamic voltage fluctuation trend data and FTU topological network model; dynamically render and evolve the dynamic topological network model based on the topological coupling correlation strength value and FTU three-dimensional feature map, and construct the FTU nonlinear state space model;

[0154] The abnormal feature module is used to perform abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; perform time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data;

[0155] The fault decision module is used to mark the faulty FTU node based on the FTU abnormal time series probability data; and to make an adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

[0156] The present invention obtains FTU voltage monitoring data and distribution network FTU position data in real time to provide necessary input data for subsequent fault detection. Through time series window decomposition, the real-time FTU voltage monitoring data is converted into a continuous FTU voltage sequence, which is convenient for subsequent frequency component analysis and dynamic voltage change analysis. Discrete wavelet multi-band decomposition is used to obtain optimized multi-frequency domain voltage data, and more abundant frequency domain information is extracted to provide more accurate features for fault detection. Through frequency component analysis, frequency components of different frequency bands in the optimized multi-frequency domain voltage data are extracted to obtain the frequency characteristics of the voltage signal. Three-dimensional feature space mapping is performed based on the frequency components. By mapping the frequency components to the three-dimensional space, an FTU three-dimensional feature mapping diagram is constructed, so that the voltage characteristics are easier to visualize and analyze. Through dynamic voltage change analysis, fluctuation trend information in the optimized multi-frequency domain voltage data is extracted to obtain the dynamic characteristics of the voltage signal. The distribution network FTU position data is used to obtain the distribution of the FTU in space, which provides basic data for subsequent topological association analysis. An FTU topological network model is constructed based on the FTU spatial distribution data, and a topological relationship between FTU nodes is established to provide topological information for fault detection. The pressure fluctuation trend data and FTU topology network model are used to calculate the topology coupling correlation strength value, evaluate the topological connection degree between FTU nodes, and use the topology coupling correlation strength value and FTU three-dimensional feature map to dynamically render and evolve the dynamic topology network model, visualize the topological state of the node, and build the FTU nonlinear state space model to facilitate fault detection and analysis. By performing abnormal spatial feature analysis on the FTU nonlinear state space model, abnormal feature data that does not conform to the normal situation is identified, which is helpful for fault detection and location. The time series probability association analysis of the abnormal feature data can determine the correlation degree and probability distribution between the abnormal features, and generate FTU abnormal time series probability data for further fault diagnosis and decision-making. The fault decision module marks the faulty FTU node based on the FTU abnormal time series probability data, and makes adaptive fault decisions to build a distribution network fault decision model. According to the FTU abnormal time series probability data, the FTU node with fault is determined and marked. Based on the marked faulty FTU node and the status information of the distribution network, a corresponding fault handling plan is formulated to build a distribution network fault decision model, which helps to quickly restore and ensure the normal operation of the distribution network.

[0157] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0158] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A fault diagnosis method based on distribution network FTU, characterized in that: The following steps are involved: Step S1: Acquire real-time FTU voltage monitoring data and distribution network FTU location data; Perform time series window decomposition on real-time FTU voltage monitoring data to obtain FTU voltage sequence; perform discrete wavelet multi-band decomposition on FTU voltage sequence to obtain optimized multi-frequency domain voltage data; Step S2: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; performing three-dimensional feature space mapping on the frequency components of the multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU; Step S3: Perform dynamic voltage change analysis on the optimized multi-frequency domain voltage data to generate dynamic voltage fluctuation trend data; obtain FTU spatial distribution data based on the FTU location data of the distribution network; and construct an FTU topology network model based on the FTU spatial distribution data. The specific steps of step S3 are: Step S31: performing dynamic voltage change analysis on the optimized multi-frequency domain voltage data to obtain FTU dynamic voltage change data; Step S32: Perform voltage fluctuation trend analysis on the FTU dynamic voltage change data to generate dynamic voltage fluctuation trend data; Step S33: performing FTU spatial distribution analysis on the FTU location data of the distribution network to obtain FTU spatial distribution data; Step S34: performing FTU topology structure analysis on the FTU spatial distribution data to generate FTU topology structure data; and constructing an FTU topology network model based on the FTU topology structure data; Step S4: obtaining a topological coupling correlation strength value based on the dynamic voltage fluctuation trend data and the FTU topology network model; Based on the topological coupling correlation strength value and the FTU three-dimensional feature map, the dynamic topological network model is dynamically rendered and evolved to construct the FTU nonlinear state space model; the specific steps of step S4 are: Step S41: performing regional load fluctuation calculation on the dynamic voltage fluctuation trend data to obtain regional load fluctuation data; Step S42: performing a topological coupling strength analysis on the FTU topological network model based on the regional dynamic load fluctuation data to obtain a topological coupling association strength value; Step S43: performing dynamic topology network adjustment on the FTU topology network model based on the topology coupling association strength value to obtain a dynamic topology network model; Step S44: dynamically rendering and evolving the dynamic topology network model according to the FTU three-dimensional feature map to construct an FTU nonlinear state space model; Step S5: performing abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; performing time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data; Step S6: Marking the faulty FTU node based on the FTU abnormal timing probability data; Adaptive fault decision is made for the faulty FTU node to build a distribution network fault decision model.

2. The fault diagnosis method based on the distribution network FTU according to claim 1 is characterized in that: The specific steps of step S1 are: Step S11: Acquire real-time FTU voltage monitoring data and distribution network FTU location data; Step S12: performing time series window decomposition on the real-time FTU voltage monitoring data to obtain FTU voltage time series data; Step S13: Serializing the FTU voltage timing data to obtain an FTU voltage sequence; Step S14: performing discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain multi-frequency domain voltage data; Step S15: performing high frequency noise elimination processing on the multi-frequency domain voltage data to obtain optimized multi-frequency domain voltage data.

3. The fault diagnosis method based on distribution network FTU according to claim 1 is characterized in that: The specific steps of step S2 are: Step S21: performing frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; Step S22: performing frequency component distribution analysis on the frequency components of the multiple frequency bands to obtain frequency domain distribution data; Step S23: Perform three-dimensional feature space mapping on the frequency components of multiple frequency bands based on the FTU voltage time series data and the frequency domain distribution data to construct an FTU three-dimensional feature mapping diagram.

4. The fault diagnosis method based on the distribution network FTU according to claim 3 is characterized in that: The specific steps of step S23 are: Obtaining first coordinate axis data based on FTU voltage timing data; Obtaining second coordinate axis data based on frequency domain distribution data; Obtaining third coordinate axis data based on frequency components of multiple frequency bands; Construct a three-dimensional coordinate system; Based on the first coordinate axis data, the second coordinate axis data and the third coordinate axis data, the three-dimensional coordinate system is mapped into a three-dimensional feature space to construct a three-dimensional feature mapping diagram of the FTU.

5. The fault diagnosis method based on distribution network FTU according to claim 1 is characterized in that: The specific steps of step S5 are: Step S51: performing abnormal space feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; Step S52: performing time series recognition on the abnormal feature data to mark the abnormal feature time series data; Step S53: performing abnormal probability distribution statistics on the abnormal feature data, thereby obtaining abnormal feature probability distribution data; Step S54: Perform time series probability association analysis on the abnormal feature probability distribution data based on the abnormal feature time series data, thereby generating FTU abnormal time series probability data.

6. The fault diagnosis method based on the distribution network FTU according to claim 5 is characterized in that: The specific steps of step S51 are: Perform FTU operation simulation on the FTU nonlinear state space model to obtain the FTU operation simulation trajectory; Based on a preset normal FTU state trajectory database, trajectory deviation identification is performed on the FTU operation simulation trajectory to obtain a deviation trajectory; Based on the deviation trajectory, the abnormal area of ​​the FTU nonlinear state space model is marked to obtain the abnormal area of ​​the FTU state; Reconstruct the local state of the abnormal area of ​​FTU state to obtain the local model of the abnormal area; Perform residual calculation on the local model of the abnormal area to generate local deviation error; Perform PCA decomposition on the local deviation error to obtain a local anomaly vector; Perform abnormal spatial feature analysis on the local abnormal vector to obtain abnormal feature data.

7. The fault diagnosis method based on distribution network FTU according to claim 1 is characterized in that: The specific steps of step S6 are: Step S61: performing fault diagnosis analysis on the FTU abnormal timing probability data to generate FTU fault diagnosis data; Step S62: locating the faulty node in the FTU topology network model based on the FTU fault diagnosis data, and marking the faulty FTU node; Step S63: Perform adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

8. A fault diagnosis system based on distribution network FTU, characterized in that: Used to execute the fault diagnosis method based on the distribution network FTU as claimed in claim 1, comprising: The multi-band decomposition module is used to obtain real-time FTU voltage monitoring data and distribution network FTU location data; perform time series window decomposition on the real-time FTU voltage monitoring data to obtain the FTU voltage sequence; perform discrete wavelet multi-band decomposition on the FTU voltage sequence to obtain optimized multi-frequency domain voltage data; The three-dimensional feature module is used to perform frequency component analysis on the optimized multi-frequency domain voltage data to obtain frequency components of multiple frequency bands; perform three-dimensional feature space mapping on the frequency components of multiple frequency bands to construct a three-dimensional feature mapping diagram of the FTU; The topology network module is used to analyze the dynamic voltage changes of the optimized multi-frequency domain voltage data and generate dynamic voltage fluctuation trend data; obtain the FTU spatial distribution data based on the FTU location data of the distribution network; and construct the FTU topology network model based on the FTU spatial distribution data; Dynamic rendering module, used to obtain the topological coupling correlation strength value based on dynamic voltage fluctuation trend data and FTU topological network model; dynamically render and evolve the dynamic topological network model based on the topological coupling correlation strength value and FTU three-dimensional feature map, and construct the FTU nonlinear state space model; The abnormal feature module is used to perform abnormal spatial feature analysis on the FTU nonlinear state space model to obtain abnormal feature data; perform time series probability association analysis on the abnormal feature data to generate FTU abnormal time series probability data; The fault decision module is used to mark the faulty FTU node based on the FTU abnormal time series probability data; and to make an adaptive fault decision on the faulty FTU node to build a distribution network fault decision model.

Images