Photovoltaic grid-connected power generation on-line monitoring and state evaluation system under transient interference

Abstract

The invention discloses a photovoltaic grid-connected power generation on-line monitoring and state evaluation system under transient interference, relates to the technical field of grid-connected power generation evaluation, and solves the technical problem that multiple nodes of photovoltaic grid-connected power generation under transient interference are difficult to accurately analyze due to neglect of mutual influence characteristics among regional nodes. Through integrated analysis of multi-source data such as a topological structure, an electrical distance, an impedance matrix, transient characteristics and the like, the limitation of a single data source is overcome, and region division and detection strategies can be dynamically adjusted along with the topological change of the system. Sub-regions are dynamically divided according to transient characteristics and power characteristics of initial region nodes; the nodes with similar transient characteristics and tight power coupling can be classified into the same sub-region, and the situation that local characteristics are covered by averaging due to the fact that the region is too large is avoided. The deep learning evaluation model is trained by integrating node and region data, so that the states of the nodes in the region can be analyzed and evaluated more accurately.

Description

Technical Field

[0001] The present invention belongs to the field of grid-connected power generation assessment, and in particular is an online monitoring and status assessment system for photovoltaic grid-connected power generation under transient interference. Background Art

[0002] Grid-connected photovoltaic power generation systems convert solar energy into direct current (DC) through the photoelectric effect of semiconductor materials, which is then converted by an inverter into alternating current (AC) with the same frequency and phase as the power grid, enabling grid-connected power transmission. The system primarily consists of photovoltaic modules, inverters, and grid access devices. Without energy storage, surplus power can be fed back to the grid. With the rapid increase in photovoltaic penetration, the transient interference problems faced by its grid-connected systems are becoming increasingly prominent, posing a severe challenge to the safe and stable operation of the power grid. Against this backdrop, online monitoring and status assessment systems for grid-connected photovoltaic power generation under transient interference have emerged, becoming a key technical support for ensuring the reliability of new power systems.

[0003] Existing online monitoring and evaluation technologies for photovoltaic grid-connected power generation usually only focus on a single indicator of power quality or operating efficiency, only process a single type of data, ignore the mutual influence characteristics between regional nodes, and cannot accurately analyze multiple nodes of photovoltaic grid-connected power generation under transient interference. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an online monitoring and status assessment system for photovoltaic grid-connected power generation under transient interference, which is used to solve the technical problem of ignoring the mutual influence characteristics between regional nodes and making it difficult to accurately analyze multiple nodes of photovoltaic grid-connected power generation under transient interference.

[0005] To solve the above problems, the first aspect of the present invention provides an online monitoring and status assessment system for photovoltaic grid-connected power generation under transient interference, comprising:

[0006] Data collection layer: PV grid-connected nodes are set up in the PV grid-connected power generation area, and high-precision synchronous measurement sensors are deployed at the PV grid-connected nodes to monitor the electrical and environmental data of the grid-connected points in real time;

[0007] Edge computing layer:

[0008] The photovoltaic grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology, and the photovoltaic grid-connected nodes are marked on the busbars of the topology. The electrical distance between nodes is quantified based on the data collected at the collection layer.

[0009] Taking the node impedance matrix as input, the spectral clustering algorithm is used to divide the initial area. According to the electromagnetic compatibility standard, the voltage sag events of each node in the initial area are detected, and the transient characteristics of each node are recorded.

[0010] The initial area is divided into sub-areas according to the transient characteristics and power characteristics of the nodes in the initial area;

[0011] For the divided sub-areas, a regional analysis model is constructed to evaluate the power quality, operating efficiency and environmental adaptability of the sub-areas when transient events occur.

[0012] Cloud analysis layer: Based on the transient characteristics of the node and the power quality, operating efficiency and environmental adaptability of the sub-area when transient events occur, an integrated multi-source data fusion algorithm and deep learning evaluation model are constructed to perform node status assessment.

[0013] Optionally, in an example of the above aspect, the data collection layer includes:

[0014] Electrical quantity acquisition unit: real-time monitoring of the three-phase voltage / current at the grid connection point, DC side voltage / current, and inverter output power;

[0015] Environmental sensing unit: monitors the light intensity and temperature at the grid connection point;

[0016] Transient event trigger unit: Set the trigger threshold through the hardware comparator to detect transient characteristics in real time; voltage sag / swell amplitude, frequency offset and harmonic distortion.

[0017] Optionally, in an example of the above aspect, the photovoltaic grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology, and photovoltaic grid-connected nodes are marked on the buses of the topology structure. Based on the data of the acquisition layer, the electrical distance between the nodes is quantified, including the following steps:

[0018] By equating the lines connecting the PV arrays and the grid connection points to feeders, and the PV branch junctions or grid connection points to buses, the PV grid-connected power generation area is equated to a multi-feeder-multi-busbar topology structure.

[0019] Photovoltaic grid-connected nodes, including:

[0020] PV injection node: located on the AC side of the busbar and inverter;

[0021] Grid connection node: located at the connection between the busbar and the upper-level grid;

[0022] Load node: located at the connection point of the busbar power structure;

[0023] Mark the PV injection nodes, grid connection nodes, and grid connection nodes at the busbars of the topology structure and number the nodes according to their types;

[0024] Based on the data from the acquisition layer, the node impedance matrix is ​​used to quantify the electrical distance between nodes:

[0025]

[0026] Where Dij is the electrical distance between node i and node j, Zii is the self-impedance of node i, Zjj is the self-impedance of node j, and Zij is the mutual impedance between nodes i and j.

[0027] Optionally, in an example of the above aspect, taking the node impedance matrix as input, using a spectral clustering algorithm to divide the initial region includes the following steps:

[0028] The node admittance matrix Ybus is formed, and then Zbus=Ybus-1 is obtained by matrix inversion. The node impedance matrix Zbus is established through the complex impedance between nodes;

[0029] Calculate the similarity between nodes using the Gaussian kernel function:

[0030]

[0031] Where Wij is the similarity between nodes i and j, |Zij| is the mutual impedance between nodes i and j, and σ is the similarity attenuation coefficient;

[0032] According to the similarity Wij between nodes, construct the similarity matrix W;

[0033] Calculate the Laplace matrix L=DW, where D is the degree matrix and the diagonal elements of the degree matrix are equal to the degree of the corresponding node;

[0034] Calculate the eigenvectors corresponding to the first k smallest eigenvalues ​​of L to form the eigenvector matrix U;

[0035] The K-means clustering algorithm is used to divide the eigenvector corresponding to each node of the eigenvector matrix U into k clusters to complete the initial region division.

[0036] Optionally, in an example of the above aspect, detecting voltage sag events at each node in the initial area and recording transient characteristics of each node according to an electromagnetic compatibility standard include the following steps:

[0037] Detect voltage sag events at each node in the initial area according to electromagnetic compatibility standards;

[0038] Extract transient energy distribution through WPT wavelet packet transform method:

[0039]

[0040] Where dj,k is the coefficient of the kth subband of the jth node, Ej is the concentration of the transient energy of node j in the frequency domain, k∈(1,2,…,N), and N is the total number of subbands;

[0041] The concentration of the transient energy of the nodes in the frequency domain is statistically analyzed and recorded in the transient characteristics of each node.

[0042] Optionally, in an example of the above aspect, dividing the initial area into sub-areas according to transient characteristics and power characteristics of nodes in the initial area includes the following steps:

[0043] According to the transient energy distribution of regional nodes and the correlation of transient characteristics between the nodes in the initial region, the nodes in the initial region are divided into sub-branches;

[0044] The historical output data of each sub-branch is collected, and the DTW dynamic time warping distance is calculated. With the DTW distance as input, hierarchical clustering is used to merge branches with similar power characteristics into sub-regions.

[0045] Optionally, in an example of the above aspect, secondary division of the internal nodes of the initial region into sub-branches is performed based on the transient energy distribution of the regional nodes and the correlation of transient characteristics between the nodes of the initial region, including the following steps:

[0046] Perform wavelet decomposition on the node transient signal and set the maximum number of decomposition layers. In each layer, calculate the energy value of the layer in each frequency band, and use the energy value of each frequency band as the element to construct the node energy feature vector.

[0047] The vector correlation is calculated by calculating the ratio of the covariance of the energy eigenvectors of two nodes to the product of the standard deviation:

[0048] , where ρij is the vector correlation between nodes, Cov(Eni,Enj) is the covariance of the energy feature vector Eni of node i and the energy feature vector Enj of node j, σEni and σEnj are the standard deviations between the energy feature vector Eni and the energy feature vector Enj of node;

[0049] By calculating the cosine similarity of the energy feature vectors of two nodes, the correlation between the vectors and the cosine similarity is weighted averaged to obtain the correlation between the nodes.

[0050] Through the clustering algorithm, each node is regarded as an independent cluster;

[0051] Calculate the correlation between all clusters based on the correlation calculation results between nodes, merge the two clusters with the smallest correlation, and repeat this step until all nodes are merged into k clusters;

[0052] Each cluster corresponds to a sub-branch, and the nodes in the cluster are members of the sub-branch, forming a secondary divided sub-branch.

[0053] Optionally, in an example of the above aspect, historical output data of each sub-branch is collected, DTW dynamic time warping distance is calculated, and hierarchical clustering is used to merge branches with similar power characteristics into sub-regions using the DTW distance as input, including the following steps:

[0054] By collecting the historical output data of each photovoltaic branch, the transient characteristics of each photovoltaic branch are obtained according to the historical output data;

[0055] Map the transient characteristics to the [-1,1] interval using the Min-Max normalization algorithm;

[0056] Through the time series X=(x1,x2,…,xn) and Y=(y1,y2,…,ym) of the transient characteristics between nodes, the power fluctuation correlation analysis is performed and the DTW dynamic time warping distance is calculated:

[0057]

[0058] Where DTW(X,Y) is the DTW distance between nodes, X,Y are the power time series of the two branches, d(xi,yj) is the Euclidean distance between nodes, n is the total number of time series of transient features of node i, and m is the total number of time series of transient features of node j;

[0059] Calculate the DTW distance between all branches to form a symmetric distance matrix;

[0060] By using the hierarchical clustering method, each branch is initialized as an independent cluster;

[0061] Perform iterative merging to find the two clusters Cp and Cq with the smallest current distance, merge Cp and Cq, update the distance matrix, and use the average link strategy to calculate the new cluster distance;

[0062] Setting termination conditions includes: presetting the number of clusters and the maximum inter-cluster distance threshold (in this embodiment, set to 0.2-0.5);

[0063] Set electrical constraint verification: perform electrical distance check. The electrical distance between nodes in the merged sub-regions is less than 0.3 pu.

[0064] Optionally, in an example of the above aspect, for the divided sub-regions, a regional analysis model is constructed to evaluate the power quality, operating efficiency, and environmental adaptability of the sub-regions when a transient event occurs, including the following steps:

[0065] By building a regional analysis model to evaluate the power quality of sub-regions:

[0066] By analyzing harmonic distortion and combining the sag depth, duration and occurrence frequency, the power quality coefficient of the sub-area when the transient event occurs is calculated;

[0067] By building a regional analysis model to evaluate the operating efficiency of sub-regions:

[0068] The operating efficiency of the sub-regional power grid is calculated by constructing a BCC model, and the obtained efficiency value is used as the operating efficiency coefficient of the sub-region when a transient event occurs;

[0069] By constructing a regional analysis model to evaluate the environmental adaptability of sub-regions:

[0070] By analyzing the impact of extreme climate on the power grid, the environmental adaptability coefficient of the sub-region when transient events occur is calculated.

[0071] Optionally, in an example of the above aspect, based on the transient characteristics of the node and the power quality, operating efficiency, and environmental adaptability of the sub-area when the transient event occurs, an integrated multi-source data fusion algorithm and a deep learning evaluation model are constructed to perform node status evaluation, including the following steps:

[0072] Obtain transient characteristic data from the node's historical output data when a pause event occurs and for a preset time period before the pause event occurs, and calculate the power quality, operating efficiency, and environmental adaptability of the corresponding transient event;

[0073] Based on whether a node fails in a preset time period before and after a pause event, labels for potential fault data or normal data are added based on the corresponding transient characteristic data, as well as the power quality, operating efficiency, and environmental adaptability when the transient event occurs.

[0074] Build an integrated multi-source data fusion algorithm and deep learning evaluation model, dynamically calculate the weights of power quality, operating efficiency, and environmental adaptability when transient events occur based on the entropy weight method, and perform weighted fusion;

[0075] The transient feature data and the weighted fusion data are input into the LSTM-Attention model for training. The model outputs node data as potential fault data or normal data.

[0076] The power quality, operating efficiency, and environmental adaptability data of the corresponding sub-area of ​​the most recently monitored node at the time of a transient event are weighted and fused, and then input into the trained LSTM-Attention model together with the transient feature data to identify the data as potential fault data or normal data.

[0077] Compared with the prior art, the present invention has the following beneficial effects:

[0078] This method overcomes the limitations of a single data source by integrating and analyzing multiple data sources, including topology, electrical distances, impedance matrices, and transient characteristics, providing a more comprehensive perspective for system status assessment. This allows for dynamic adjustment of zoning and detection strategies as system topology changes, whereas traditional methods may require remodeling or parameter tuning.

[0079] The present invention dynamically divides sub-regions based on the transient and power characteristics of the nodes in the initial region. In a photovoltaic grid-connected system, nodes with similar transient characteristics and close power coupling can be grouped into the same sub-region, avoiding the averaging of local characteristics due to overly large regions.

[0080] The present invention comprehensively analyzes the transient characteristic data of the nodes, as well as the power quality, operating efficiency and environmental adaptability data of the area corresponding to the nodes. By integrating the node and area data, a deep learning evaluation model is trained to facilitate more accurate analysis and evaluation of the status of the nodes in the area. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0082] Figure 1 Schematic diagram of the system framework of the present invention;

[0083] Figure 2 This is a schematic diagram of the data acquisition layer framework of the present invention. DETAILED DESCRIPTION

[0084] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0085] See also Figure 1-Figure 2 The first embodiment of the present invention provides an online monitoring and status assessment system for photovoltaic grid-connected power generation under transient interference, comprising:

[0086] Data collection layer: PV grid-connected nodes are set up in the PV grid-connected power generation area, and high-precision synchronous measurement sensors are deployed at the PV grid-connected nodes to monitor the electrical and environmental data of the grid-connected points in real time;

[0087] Edge computing layer:

[0088] The photovoltaic grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology, and the photovoltaic grid-connected nodes are marked on the busbars of the topology. The electrical distance between nodes is quantified based on the data collected at the collection layer.

[0089] Taking the node impedance matrix as input, the spectral clustering algorithm is used to divide the initial area. According to the electromagnetic compatibility standard, the voltage sag events of each node in the initial area are detected, and the transient characteristics of each node are recorded.

[0090] The initial area is divided into sub-areas according to the transient characteristics and power characteristics of the nodes in the initial area;

[0091] For the divided sub-areas, a regional analysis model is constructed to evaluate the power quality, operating efficiency and environmental adaptability of the sub-areas when transient events occur.

[0092] Cloud analysis layer: Based on the transient characteristics of the node and the power quality, operating efficiency and environmental adaptability of the sub-area when transient events occur, an integrated multi-source data fusion algorithm and deep learning evaluation model are constructed to perform node status assessment.

[0093] Specifically, this embodiment uses equivalent modeling to abstract the complex photovoltaic grid-connected system into a structured network, clarifying the connection between each feeder and busbar, and providing a clear physical framework for subsequent analysis. Based on acquisition layer data, the electrical distance between nodes is calculated using an impedance matrix or related algorithms. This overcomes the limitations of traditional methods that rely solely on geographic distance or simple topological relationships, and more accurately reflects the electrical coupling strength between nodes.

[0094] Based on electromagnetic compatibility standards (such as IEC 61000-4-11), voltage sag thresholds, duration, and other parameters are clearly defined to ensure standardized and comparable test results. Existing technologies may miss detections or misidentify them due to inconsistent standards or crude detection methods. While detecting voltage sags, the transient characteristics of the node are recorded, providing a basis for in-depth analysis of the sag's cause. Traditional methods may only focus on the sag's amplitude and duration.

[0095] Integrating and analyzing multi-source data, including topology, electrical distances, impedance matrices, and transient characteristics, overcomes the limitations of a single data source and provides a more comprehensive perspective for system status assessment. This facilitates dynamic adjustment of zoning and detection strategies as system topology changes, whereas traditional methods may require remodeling or parameter tuning.

[0096] While existing technologies often rely on static topology or fixed geographic boundaries to divide regions, this method dynamically divides subregions based on the transient and power characteristics of the nodes in the initial region. For example, in a photovoltaic grid-connected system, nodes with similar transient characteristics and close power coupling can be grouped into the same subregion, preventing local characteristics from being masked by averaging due to large regions.

[0097] While existing technologies typically focus solely on power quality or operational efficiency, the regional analysis model constructed by this method simultaneously evaluates power quality, operational efficiency, and environmental adaptability. For example, within an industrial park power grid, it can identify sub-regions with poor power quality but high operational efficiency during transient events, providing a balance point for optimized scheduling.

[0098] Existing technologies mostly process a single type of data and ignore the mutual influence characteristics between regional nodes. This method comprehensively analyzes the transient characteristic data of the nodes, as well as the power quality, operating efficiency and environmental adaptability data of the node's corresponding area. By integrating node and regional data, a deep learning evaluation model is trained to facilitate more accurate analysis and evaluation of the status of nodes in the region.

[0099] In one embodiment of the present invention, the data acquisition layer includes:

[0100] Electrical quantity acquisition unit: real-time monitoring of the three-phase voltage / current, DC side voltage / current and inverter output power at the grid connection point, supporting the IEC 61850 standard communication protocol;

[0101] Environmental sensing unit: integrated light intensity sensor and temperature sensor (resolution 0.1°C) to monitor the light intensity and temperature at the grid connection point;

[0102] Transient event trigger unit: The trigger threshold is set by the hardware comparator to detect transient characteristics in real time; voltage sag / swell amplitude, frequency offset and harmonic distortion. The trigger threshold is configurable and can be set to start recording when the voltage drops to 85%Un.

[0103] In one embodiment of the present invention, a photovoltaic grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology, and photovoltaic grid-connected nodes are marked on the buses of the topology. Based on the data from the acquisition layer, the electrical distance between the nodes is quantified, including the following steps:

[0104] By equating the lines connecting the PV arrays and the grid connection points to feeders, and the PV branch junctions or grid connection points to buses, the PV grid-connected power generation area is equated to a multi-feeder-multi-busbar topology structure.

[0105] Photovoltaic grid-connected nodes, including:

[0106] PV injection node: located on the AC side of the busbar and inverter, directly reflects the location where PV power is injected into the grid;

[0107] Grid connection node: located at the connection between the busbar and the upper-level grid, used to monitor grid-connected power, voltage, frequency and other parameters;

[0108] Load node: Located at the connection point of the busbar power structure, load nodes need to be marked to distinguish between power generation and power consumption;

[0109] Mark the PV injection nodes, grid connection nodes, and grid connection nodes at the busbars of the topology structure and number the nodes according to their types;

[0110] Based on the data from the acquisition layer, the node impedance matrix is ​​used to quantify the electrical distance between nodes:

[0111]

[0112] Where Dij is the electrical distance between nodes i and j, Zii is the self-impedance of node i, the diagonal element of the impedance matrix, Zjj is the self-impedance of node j, and Zij is the mutual impedance between nodes i and j, the off-diagonal element of the impedance matrix.

[0113] In one embodiment of the present invention, a spectral clustering algorithm is used to divide the initial region using a node impedance matrix as input, including the following steps:

[0114] The node admittance matrix Ybus is formed, and then Zbus=Ybus-1 is obtained by matrix inversion. The node impedance matrix Zbus is established through the complex impedance between nodes;

[0115] Calculate the similarity between nodes using the Gaussian kernel function:

[0116]

[0117] Where Wij is the similarity between nodes i and j, |Zij| is the mutual impedance between nodes i and j, and σ is the similarity attenuation coefficient;

[0118] According to the similarity Wij between nodes, construct the similarity matrix W;

[0119] Calculate the Laplace matrix L=DW, where D is the degree matrix. The diagonal elements of the degree matrix are equal to the degree of the corresponding node, reflecting the connection strength of the node with other nodes;

[0120] Calculate the eigenvectors corresponding to the first k smallest eigenvalues ​​of L to form the eigenvector matrix U;

[0121] The K-means clustering algorithm is used to divide the eigenvector corresponding to each node of the eigenvector matrix U into k clusters to complete the initial region division.

[0122] In one embodiment of the present invention, detecting voltage sag events at each node in an initial area and recording transient characteristics of each node according to electromagnetic compatibility standards includes the following steps:

[0123] Detect voltage sag events at each node in the initial area according to electromagnetic compatibility standards;

[0124] Extract transient energy distribution through WPT wavelet packet transform method:

[0125]

[0126] Where dj,k is the coefficient of the kth subband of the jth node, Ej is the concentration of the transient energy of node j in the frequency domain, k∈(1,2,…,N), and N is the total number of subbands;

[0127] The concentration of the transient energy of the nodes in the frequency domain is statistically analyzed and recorded in the transient characteristics of each node.

[0128] In this embodiment, a detailed model of a photovoltaic power station is built in PSCAD / EMTDC, and the IEC 61000-4-34 electromagnetic compatibility standard is injected. According to the voltage sag event defined therein, with a depth of 20% and a duration of 100ms, the transient voltage / current waveforms of each node are recorded.

[0129] The WPT wavelet packet transform method is used to extract transient energy distribution.

[0130] In one embodiment of the present invention, the initial area is divided into sub-areas according to the transient characteristics and power characteristics of the nodes in the initial area, including the following steps:

[0131] According to the transient energy distribution of regional nodes and the correlation of transient characteristics between the nodes in the initial region, the nodes in the initial region are divided into sub-branches;

[0132] The historical output data of each sub-branch is collected, and the DTW dynamic time warping distance is calculated. With the DTW distance as input, hierarchical clustering is used to merge branches with similar power characteristics into sub-regions.

[0133] In one embodiment of the present invention, based on the transient energy distribution of the regional nodes and the correlation of the transient characteristics between the nodes in the initial region, the nodes in the initial region are divided into sub-branches, including the following steps:

[0134] Perform wavelet decomposition on the node transient signal and set the maximum number of decomposition layers. In each layer, calculate the energy value of the layer in each frequency band, and use the energy value of each frequency band as the element to construct the node energy feature vector.

[0135] The vector correlation is calculated by calculating the ratio of the covariance of the energy eigenvectors of two nodes to the product of the standard deviation:

[0136] , where ρij is the vector correlation between nodes, Cov(Eni,Enj) is the covariance of the energy feature vector Eni of node i and the energy feature vector Enj of node j, σEni and σEnj are the standard deviations between the energy feature vector Eni and the energy feature vector Enj of node;

[0137] By calculating the cosine similarity of the energy feature vectors of two nodes, the correlation between the vectors and the cosine similarity is weighted averaged to obtain the correlation between the nodes.

[0138] Through the clustering algorithm, each node is regarded as an independent cluster;

[0139] Calculate the correlation between all clusters based on the correlation calculation results between nodes, merge the two clusters with the smallest correlation, and repeat this step until all nodes are merged into k clusters;

[0140] Each cluster corresponds to a sub-branch, and the nodes in the cluster are members of the sub-branch, forming a secondary divided sub-branch.

[0141] In one embodiment of the present invention, historical output data of each sub-branch is collected, the DTW dynamic time warping distance is calculated, and hierarchical clustering is used to merge branches with similar power characteristics into sub-regions using the DTW distance as input, including the following steps:

[0142] By collecting the historical output data of each photovoltaic branch, the transient characteristics of each photovoltaic branch are obtained according to the historical output data;

[0143] Map the transient characteristics to the [-1,1] interval using the Min-Max normalization algorithm;

[0144] Through the time series X=(x1,x2,…,xn) and Y=(y1,y2,…,ym) of the transient characteristics between nodes, the power fluctuation correlation analysis is performed and the DTW dynamic time warping distance is calculated:

[0145]

[0146] Where DTW(X,Y) is the DTW distance between nodes, X,Y are the power time series of the two branches, d(xi,yj) is the Euclidean distance between nodes, n is the total number of time series of transient features of node i, and m is the total number of time series of transient features of node j;

[0147] Calculate the DTW distance between all branches to form a symmetric distance matrix;

[0148] By using the hierarchical clustering method, each branch is initialized as an independent cluster;

[0149] Perform iterative merging to find the two clusters Cp and Cq with the smallest current distance, merge Cp and Cq, update the distance matrix, and use the average link strategy to calculate the new cluster distance;

[0150] Setting termination conditions includes: presetting the number of clusters and the maximum inter-cluster distance threshold (in this embodiment, set to 0.2-0.5);

[0151] Set electrical constraint verification: Perform electrical distance check. The electrical distance between nodes in the merged sub-regions is less than 0.3 pu, ensuring a strong coupling relationship.

[0152] In one embodiment of the present invention, for the divided sub-regions, a regional analysis model is constructed to evaluate the power quality, operating efficiency, and environmental adaptability of the sub-regions when a transient event occurs, including the following steps:

[0153] By building a regional analysis model to evaluate the power quality of sub-regions:

[0154] By analyzing the harmonic distortion and combining the sag depth, duration and occurrence frequency, the power quality coefficient of the sub-area when the transient event occurs is calculated:

[0155]

[0156] Where TE is the power quality coefficient of the sub-area, Uh is the effective value of the harmonic voltage of the h-th node in the sub-area, and U0 is the effective value of the fundamental voltage of the h-th node in the sub-area; h∈(1,2,…,H), H is the total number of nodes in the sub-area, wh is the dynamic weight, calculated by the entropy weight method; Vch is the residual voltage during the sag of the h-th node in the sub-area, V0h is the rated voltage of the h-th node in the sub-area; th is the mean duration of the voltage sag at the h-th node in the sub-area, and T is the statistical period; λh is the correction coefficient for the sag frequency of the h-th node in the sub-area, which is fitted by the Poisson distribution; a1 and a2 are the corresponding coefficients respectively.

[0157] By building a regional analysis model to evaluate the operating efficiency of sub-regions:

[0158] The operating efficiency of the sub-regional power grid is calculated by constructing a BCC model, and the obtained efficiency value is used as the operating efficiency coefficient of the sub-region when a transient event occurs;

[0159]

[0160]

[0161]

[0162]

[0163] Among them, θ is the efficiency value (0≤θ≤1), is the efficiency value of the c-th input of all nodes in the sub-region, is the r-th output of all nodes in the sub-region, is a non-Archimedean infinitesimal; xch is the cth input of the hth node in the subregion; yrh is the rth output of the hth node in the subregion; sc - and sr + is a slack variable, representing input redundancy and output deficiency, c∈(1,2,…,g), r∈(1,2,…,s), g is the total number of input items of all nodes in the sub-region, s is the total number of output items of all nodes in the sub-region;

[0164] By constructing a regional analysis model to evaluate the environmental adaptability of sub-regions:

[0165] By analyzing the impact of extreme climate on the power grid, the environmental adaptability coefficient of the sub-region when transient events occur is calculated:

[0166]

[0167] Where Re is the environmental adaptability coefficient of the sub-region, Pb(h,t) is the probability of equipment failure at the h-th node in the sub-region in the t-th time monitoring interval, and f(h,t) is the equipment spatial distribution density function of the h-th node in the sub-region in the t-th time monitoring interval, generated by kernel density estimation, t∈(1,2,…,W), W is the total number of time monitoring intervals;

[0168] The probability of equipment failure is calculated using the following formula:

[0169]

[0170] Where Pb is the probability of equipment failure, β0, β1, and β2 are calculation parameters. In this embodiment, β0, β1, and β2 are set to 0.2, 0.4, and 0.4, respectively.

[0171] In one embodiment of the present invention, based on the transient characteristics of the node and the power quality, operating efficiency, and environmental adaptability of the sub-region when the transient event occurs, an integrated multi-source data fusion algorithm and a deep learning evaluation model are constructed to perform node status evaluation, including the following steps:

[0172] Obtain transient characteristic data from the node's historical output data when a pause event occurs and for a preset time period before the pause event occurs, and calculate the power quality, operating efficiency, and environmental adaptability of the corresponding transient event;

[0173] Based on whether the node sends a fault signal in the preset time period before and after the pause event, the corresponding transient characteristic data, as well as the power quality, operating efficiency and environmental adaptability when the transient event occurs, are labeled as potential fault data or normal data;

[0174] Build an integrated multi-source data fusion algorithm and deep learning evaluation model, dynamically calculate the weights of power quality, operating efficiency, and environmental adaptability when transient events occur based on the entropy weight method, and perform weighted fusion;

[0175] The transient feature data and the weighted fusion data are input into the LSTM-Attention model for training. The model outputs node data as potential fault data or normal data.

[0176] The power quality, operating efficiency, and environmental adaptability data of the corresponding sub-area of ​​the most recently monitored node at the time of a transient event are weighted and fused, and then input into the trained LSTM-Attention model together with the transient feature data to identify the data as potential fault data or normal data.

[0177] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference, characterized in that: include: Data collection layer: PV grid-connected nodes are set up in the PV grid-connected power generation area, and high-precision synchronous measurement sensors are deployed at the PV grid-connected nodes to monitor the electrical and environmental data of the grid-connected points in real time; Edge computing layer: The photovoltaic grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology, and the photovoltaic grid-connected nodes are marked on the busbars of the topology. The electrical distance between nodes is quantified based on the data collected at the collection layer. Taking the node impedance matrix as input, the spectral clustering algorithm is used to divide the initial area. According to the electromagnetic compatibility standard, the voltage sag events of each node in the initial area are detected, and the transient characteristics of each node are recorded. The initial area is divided into sub-areas according to the transient characteristics and power characteristics of the nodes in the initial area; For the divided sub-areas, a regional analysis model is constructed to evaluate the power quality, operating efficiency and environmental adaptability of the sub-areas when transient events occur. Cloud analysis layer: Based on the transient characteristics of the node and the power quality, operating efficiency and environmental adaptability of the sub-area when transient events occur, an integrated multi-source data fusion algorithm and deep learning evaluation model are constructed to perform node status assessment.

2. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1 is characterized in that: The data collection layer includes: Electrical quantity acquisition unit: real-time monitoring of the three-phase voltage / current at the grid connection point, DC side voltage / current, and inverter output power; Environmental sensing unit: monitors the light intensity and temperature at the grid connection point; Transient event trigger unit: Set the trigger threshold through the hardware comparator to detect transient characteristics in real time; voltage sag / swell amplitude, frequency offset and harmonic distortion.

3. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1 is characterized in that: The PV grid-connected power generation area is equivalent to a multi-feeder-multi-busbar topology. The PV grid-connected nodes are marked on the busbars of the topology. Based on the data collected at the acquisition layer, the electrical distance between the nodes is quantified. The following steps are included: By equating the lines connecting the PV arrays and the grid connection points to feeders, and the PV branch junctions or grid connection points to buses, the PV grid-connected power generation area is equated to a multi-feeder-multi-busbar topology structure. Photovoltaic grid-connected nodes, including: PV injection node: located on the AC side of the busbar and inverter; Grid connection node: located at the connection between the busbar and the upper-level grid; Load node: located at the connection point of the busbar power structure; Mark the PV injection nodes, grid connection nodes, and grid connection nodes at the busbars of the topology structure and number the nodes according to their types; Based on the data from the acquisition layer, the node impedance matrix is ​​used to quantify the electrical distance between nodes: ； Where Dij is the electrical distance between node i and node j, Zii is the self-impedance of node i, Zjj is the self-impedance of node j, and Zij is the mutual impedance between nodes i and j.

4. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1 is characterized in that: Taking the node impedance matrix as input, the spectral clustering algorithm is used to divide the initial region, which includes the following steps: The node admittance matrix Ybus is formed, and then Zbus=Ybus-1 is obtained by matrix inversion. The node impedance matrix Zbus is established through the complex impedance between nodes; Calculate the similarity between nodes using the Gaussian kernel function: ； Where Wij is the similarity between nodes i and j, |Zij| is the mutual impedance between nodes i and j, and σ is the similarity attenuation coefficient; According to the similarity Wij between nodes, construct the similarity matrix W; Calculate the Laplace matrix L=DW, where D is the degree matrix and the diagonal elements of the degree matrix are equal to the degree of the corresponding node; Calculate the eigenvectors corresponding to the first k smallest eigenvalues ​​of L to form the eigenvector matrix U; The K-means clustering algorithm is used to divide the eigenvector corresponding to each node of the eigenvector matrix U into k clusters to complete the initial region division.

5. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1 is characterized in that: According to the electromagnetic compatibility standard, the voltage sag events at each node in the initial area are detected and the transient characteristics of each node are recorded, including the following steps: Detect voltage sag events at each node in the initial area according to electromagnetic compatibility standards; Extract transient energy distribution through WPT wavelet packet transform method: ； Where dj,k is the coefficient of the kth subband of the jth node, Ej is the concentration of the transient energy of node j in the frequency domain, k∈(1,2,…,N), and N is the total number of subbands; The concentration of the transient energy of the nodes in the frequency domain is statistically analyzed and recorded in the transient characteristics of each node.

6. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1 is characterized in that: Based on the transient characteristics and power characteristics of the nodes in the initial area, the initial area is divided into sub-areas, including the following steps: According to the transient energy distribution of regional nodes and the correlation of transient characteristics between the nodes in the initial region, the nodes in the initial region are divided into sub-branches; The historical output data of each sub-branch is collected, and the DTW dynamic time warping distance is calculated. With the DTW distance as input, hierarchical clustering is used to merge branches with similar power characteristics into sub-regions.

7. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 6 is characterized in that: According to the transient energy distribution of regional nodes and the correlation of transient characteristics between nodes in the initial region, the internal nodes of the initial region are divided into sub-branches, including the following steps: Perform wavelet decomposition on the node transient signal and set the maximum number of decomposition layers. In each layer, calculate the energy value of the layer in each frequency band, and use the energy value of each frequency band as the element to construct the node energy feature vector. The vector correlation is calculated by calculating the ratio of the covariance of the energy eigenvectors of two nodes to the product of the standard deviation: , where ρij is the vector correlation between nodes, Cov(Eni,Enj) is the covariance of the energy feature vector Eni of node i and the energy feature vector Enj of node j, σEni and σEnj are the standard deviations between the energy feature vector Eni and the energy feature vector Enj of node; By calculating the cosine similarity of the energy feature vectors of two nodes, the correlation between the vectors and the cosine similarity is weighted averaged to obtain the correlation between the nodes. Through the clustering algorithm, each node is regarded as an independent cluster; Calculate the correlation between all clusters based on the correlation calculation results between nodes, merge the two clusters with the smallest correlation, and repeat this step until all nodes are merged into k clusters; Each cluster corresponds to a sub-branch, and the nodes in the cluster are members of the sub-branch, forming a secondary divided sub-branch.

8. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 6 is characterized in that: Collect historical output data of each sub-branch, calculate the DTW dynamic time warping distance, use the DTW distance as input, and use hierarchical clustering to merge branches with similar power characteristics into sub-regions. The following steps are included: By collecting the historical output data of each photovoltaic branch, the transient characteristics of each photovoltaic branch are obtained according to the historical output data; Map the transient characteristics to the [-1,1] interval using the Min-Max normalization algorithm; Through the time series X=(x1,x2,…,xn) and Y=(y1,y2,…,ym) of the transient characteristics between nodes, the power fluctuation correlation analysis is performed and the DTW dynamic time warping distance is calculated: ； Where DTW(X,Y) is the DTW distance between nodes, X,Y are the power time series of the two branches, d(xi,yj) is the Euclidean distance between nodes, n is the total number of time series of transient features of node i, and m is the total number of time series of transient features of node j; Calculate the DTW distance between all branches to form a symmetric distance matrix; By using the hierarchical clustering method, each branch is initialized as an independent cluster; Perform iterative merging to find the two clusters Cp and Cq with the smallest current distance, merge Cp and Cq, update the distance matrix, and use the average link strategy to calculate the new cluster distance; Setting termination conditions includes: presetting the number of clusters and the maximum inter-cluster distance threshold; Set electrical constraint verification: perform electrical distance check. The electrical distance between nodes in the merged sub-regions is less than 0.3 pu.

9. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1, characterized in that: For the divided sub-areas, a regional analysis model is constructed to evaluate the power quality, operating efficiency, and environmental adaptability of the sub-areas when transient events occur. The following steps are included: By building a regional analysis model to evaluate the power quality of sub-regions: By analyzing the harmonic distortion and combining the sag depth, duration and occurrence frequency, the power quality coefficient of the sub-area when the transient event occurs is calculated: ； Where TE is the power quality coefficient of the sub-area, Uh is the effective value of the harmonic voltage at the h-th node in the sub-area, and U0 is the effective value of the fundamental voltage at the h-th node in the sub-area; h∈(1,2,…,H), H is the total number of nodes in the sub-area, and wh is the dynamic weight; Vch is the residual voltage during the sag at the h-th node in the sub-area, and V0h is the rated voltage at the h-th node in the sub-area; th is the mean duration of the voltage sag at the h-th node in the sub-area, and T is the statistical period; λh is the correction coefficient for the sag occurrence frequency at the h-th node in the sub-area, which is fitted by the Poisson distribution; a1 and a2 are the corresponding coefficients respectively. By building a regional analysis model to evaluate the operating efficiency of sub-regions: The operating efficiency of the sub-regional power grid is calculated by constructing a BCC model, and the obtained efficiency value is used as the operating efficiency coefficient of the sub-region when a transient event occurs; ； ； ； ； Among them, θ is the efficiency value (0≤θ≤1), is the efficiency value of the c-th input of all nodes in the sub-region, is the r-th output of all nodes in the sub-region, is a non-Archimedean infinitesimal; xch is the cth input of the hth node in the subregion; yrh is the rth output of the hth node in the subregion; sc - and sr + is a slack variable, representing input redundancy and output deficiency, c∈(1,2,…,g), r∈(1,2,…,s), g is the total number of input items of all nodes in the sub-region, s is the total number of output items of all nodes in the sub-region; By constructing a regional analysis model to evaluate the environmental adaptability of sub-regions: By analyzing the impact of extreme climate on the power grid, the environmental adaptability coefficient of the sub-region when transient events occur is calculated: ； Where Re is the environmental adaptability coefficient of the sub-region, Pb(h,t) is the probability of equipment failure at the h-th node in the sub-region in the t-th time monitoring interval, and f(h,t) is the equipment spatial distribution density function of the h-th node in the sub-region in the t-th time monitoring interval, generated by kernel density estimation, t∈(1,2,…,W), W is the total number of time monitoring intervals; The probability of equipment failure is calculated using the following formula: ； Where Pb is the probability of equipment failure, and β0, β1, and β2 are calculation parameters.

10. The photovoltaic grid-connected power generation online monitoring and status assessment system under transient interference according to claim 1, characterized in that: Based on the transient characteristics of the node and the power quality, operating efficiency, and environmental adaptability of the sub-area when a transient event occurs, an integrated multi-source data fusion algorithm and deep learning evaluation model are constructed to perform node status assessment, including the following steps: Obtain transient characteristic data from the node's historical output data when a pause event occurs and for a preset time period before the pause event occurs, and calculate the power quality, operating efficiency, and environmental adaptability of the corresponding transient event; Based on whether a node fails in a preset time period before and after a pause event, labels for potential fault data or normal data are added based on the corresponding transient characteristic data, as well as the power quality, operating efficiency, and environmental adaptability when the transient event occurs. Build an integrated multi-source data fusion algorithm and deep learning evaluation model, dynamically calculate the weights of power quality, operating efficiency, and environmental adaptability when transient events occur based on the entropy weight method, and perform weighted fusion; The transient feature data and the weighted fusion data are input into the LSTM-Attention model for training. The model outputs node data as potential fault data or normal data. The power quality, operating efficiency, and environmental adaptability data of the corresponding sub-area of ​​the most recently monitored node at the time of a transient event are weighted and fused, and then input into the trained LSTM-Attention model together with the transient feature data to identify the data as potential fault data or normal data.

Images