Fault monitoring method and device for power distribution network

Abstract

The invention provides a fault monitoring method and device for a power distribution network. The method comprises the following steps: acquiring time sequence data of various parameters of each node in the monitored power distribution network; calculating fault probability data of the center node at each moment according to the time sequence data of the adjacent nodes of the center node; taking each type of the parameters as a dimension, and obtaining a multi-dimensional parameter vector of each node at each moment; correcting the multidimensional parameter vector of the corresponding node at the corresponding moment through the fault probability data of each node at each moment; calculating the parameter change degree of each node at each moment according to the corrected multi-dimensional parameter vector of each node at each moment; calculating influence data of the central node according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent node of the central node at each moment; and correcting the attention weight between every two nodes according to the influence data of every two nodes.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault monitoring, and in particular to a distribution network fault monitoring method and device. Background Art

[0002] The distribution network has obvious graph characteristics. For example, equipment or components such as transformers and circuit breakers can be regarded as nodes in the graph. Therefore, it is suitable to use graph neural networks with attention mechanisms to diagnose and monitor faults of each of the above nodes. There are also many application examples of related technologies.

[0003] However, since the important feature of the graph neural network with the introduction of the attention mechanism when applied to distribution network fault monitoring is to pay special attention to data changes, the network will still pay too much attention to the data that changes normally. When some equipment has sudden changes in parameters under the normal operation of the distribution network, such as the sudden access of a large load causing changes in the voltage and current of its upstream equipment, the accuracy of the attention weight in the network will deteriorate or even lose its meaning, so the accuracy of fault judgment and monitoring will be seriously affected. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method and device for monitoring faults in a distribution network, which can improve the accuracy of fault judgment in distribution network fault monitoring, thereby ensuring effective monitoring of distribution network faults.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A fault monitoring method for a distribution network comprises the following steps: obtaining time series data of multiple parameters of each node in a monitored distribution network, wherein the nodes in the monitored distribution network are devices or components in the monitored distribution network; regarding each of the nodes as a central node, when a certain node is regarded as a central node, calculating the fault probability data of the central node at each moment according to the time series data of the adjacent nodes of the central node, thereby obtaining the fault probability data of each node at each moment; taking each type of the parameter as a dimension, constructing the parameters of each node at each moment into multidimensional coordinate data, and obtaining a multidimensional parameter vector of each node at each moment; and obtaining a multidimensional parameter vector of each node at each moment by pairing the fault probability data of each node at each moment with each other. The multidimensional parameter vector of the node at the corresponding moment is corrected; the degree of parameter change of each node at each moment is calculated according to the corrected multidimensional parameter vector of each node at each moment; each of the nodes is regarded as a central node, and when a certain node is regarded as a central node, the influence data of the central node is calculated according to the degree of parameter change of the central node at each moment and the degree of parameter change of the adjacent nodes of the central node at each moment, so as to obtain the influence data of each node; the attention weight between each two nodes is corrected according to the influence data of the two nodes; by introducing a graph neural network with an attention mechanism, the fault judgment of the monitored distribution network is realized with the corrected attention weight between each two nodes.

[0007] The initial data of the central node's failure probability at each moment is calculated using the following formula:

[0008]

[0009] Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c represents the cth parameter of the v+1th adjacent node of the Vth central node at the tth time, N1 is the total number of parameter types, and N2 is the total number of adjacent nodes of the Vth central node; the initial data of the failure probability of each node at each moment are normalized to obtain the failure probability data of each node at each moment.

[0010] The modified multidimensional parameter vector of each node at each moment is calculated by the following formula:

[0011] A′ V，t =P′ V，t A V，t

[0012] Among them, A′V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

[0013] The degree of parameter change of each node at each moment is calculated by the following formula:

[0014]

[0015] Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τ The sizes of the two vectors, α and β are the angle influence coefficient and size influence coefficient respectively, and N3 is the total number of moments.

[0016] The influence data of the central node is calculated by the following formula:

[0017]

[0018] Among them, E V Represents the influence data of the Vth central node, M v,t It represents the parameter change degree of the vth adjacent node of the Vth central node at the tth time, N2 is the total number of adjacent nodes of the Vth central node, N3 is the total number at the time, e is a natural constant, and ^ represents exponentiation.

[0019] The corrected attention weight between every two nodes is calculated by the following formula:

[0020]

[0021] Among them, w′ i,j represents the corrected attention weight between the i-th node and the j-th node, w i,j represents the attention weight between the i-th node and the j-th node before correction, E i Represents the influence data of the i-th node, E j represents the influence data of the jth node, It represents the mean of the difference between the influence data of all nodes.

[0022] A fault monitoring device for a distribution network comprises: an acquisition module, used to acquire time series data of multiple parameters of each node in the monitored distribution network, wherein the nodes in the monitored distribution network are devices or components in the monitored distribution network; a first calculation module, used to regard each of the nodes as a central node, and when a certain node is regarded as a central node, calculate the fault probability data of the central node at each moment according to the time series data of the adjacent nodes of the central node, so as to obtain the fault probability data of each node at each moment; a construction module, used to construct the parameters of each node at each moment into multi-dimensional coordinate data with each type of the parameter as a dimension, and obtain the multi-dimensional parameter vector of each node at each moment; a first correction module, used to correct the fault probability data of each node at each moment according to the fault probability data of each node at each moment. The multidimensional parameter vector at the corresponding moment is corrected; the second calculation module is used to calculate the parameter change degree of each node at each moment according to the corrected multidimensional parameter vector of each node at each moment; the third calculation module is used to regard each of the nodes as a central node, and when a certain node is regarded as a central node, the influence data of the central node is calculated according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent nodes of the central node at each moment, so as to obtain the influence data of each node; the second correction module is used to correct the attention weight between every two nodes according to the influence data of the two nodes; the judgment module is used to realize the fault judgment of the monitored distribution network by introducing a graph neural network with an attention mechanism and using the corrected attention weight between every two nodes.

[0023] The first calculation module calculates the initial data of the failure probability of the central node at each moment by the following formula:

[0024]

[0025] Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c It represents the cth parameter of the v+1th adjacent node of the Vth central node at the tth time, N1 is the total number of parameter types, N2 is the total number of adjacent nodes of the Vth central node, and the first calculation module normalizes the initial data of the failure probability of each node at each time to obtain the failure probability data of each node at each time.

[0026] The first correction module calculates the corrected multidimensional parameter vector of each node at each time by the following formula:

[0027] A′ V,t =P V ' ,t A V,t

[0028] Among them, A′ V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

[0029] The second calculation module calculates the parameter change degree of each node at each moment by the following formula:

[0030]

[0031] Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τ The sizes of the two vectors, α and β are the angle influence coefficient and size influence coefficient respectively, and N3 is the total number of moments.

[0032] Beneficial effects of the present invention:

[0033] The present invention obtains the time series data of multiple parameters of each node in the monitored distribution network to calculate the fault probability data of each node at each moment, and accordingly corrects the multidimensional parameter vector of each node at each moment, and then calculates the parameter change degree of each node at each moment and the influence data of each node in turn, and corrects the attention weight between every two nodes accordingly, and finally uses the corrected attention weight to realize the fault judgment of the distribution network by using the graph neural network with the introduction of attention mechanism. Therefore, by obtaining the actual influence of the node, a more accurate attention weight can be corrected. When the graph neural network with the introduction of attention mechanism is used for distribution network fault monitoring, the accuracy of fault judgment in distribution network fault monitoring can be improved, thereby ensuring the effective monitoring of distribution network faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flow chart of a method for monitoring a fault in a power distribution network according to an embodiment of the present invention;

[0035] Figure 2 The present invention is a block diagram of a fault monitoring device for a distribution network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] The embodiment of the present invention adopts a graph neural network with an attention mechanism to realize fault judgment of the distribution network, thereby realizing fault monitoring of the distribution network. The core of the fault monitoring method and device of the distribution network in the embodiment of the present invention is to correct the attention weight in the above-mentioned graph neural network, so as to improve the accuracy of fault judgment in the fault monitoring of the distribution network.

[0038] like Figure 1 As shown, the fault monitoring method of the distribution network according to the embodiment of the present invention comprises the following steps:

[0039] S1, obtaining time series data of multiple parameters of each node in the monitored distribution network, wherein the node in the monitored distribution network is a device or component in the monitored distribution network.

[0040] First of all, it should be noted that the node in the embodiment of the present invention is a concept in a graph, and the corresponding entity in the distribution network is a device or component such as a transformer or a circuit breaker. If two nodes are connected by an edge, the two nodes are adjacent nodes to each other. The various parameters of the node include current, temperature, operating status, etc., among which the operating status parameter refers to the state of whether the device or component is running, which is the switch state for the switch device. The data corresponding to on and off can be 1 and 0, or 1 and -1, etc.

[0041] In the embodiment of the present invention, the time series data of various parameters of each node in the monitored power distribution network can be parameters collected at the same sampling interval starting from the same time. For example, the time series data of the current of a certain node can be {I1, I2, I3, ...}, where I x It represents the current value of the node at the xth moment, where x is a positive integer.

[0042] S2, each node is regarded as a central node. When a certain node is regarded as a central node, the failure probability data of the central node at each moment is calculated according to the time series data of the adjacent nodes of the central node, so as to obtain the failure probability data of each node at each moment.

[0043] It should be understood that the central node and the adjacent node in the embodiments of the present invention are both nodes in the monitored power distribution network, and the two are relative concepts that only exist when they appear at the same time. Therefore, in the embodiments of the present invention, the central node, adjacent node, and node in some expressions can be regarded as the same concept according to their actual meaning.

[0044] In one embodiment of the present invention, the initial data of the failure probability of the central node at each moment can be calculated by the following formula:

[0045]

[0046] Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c It represents the cth parameter of the v+1th neighboring node of the Vth central node at the tth time, N1 is the total number of parameter types, and N2 is the total number of neighboring nodes of the Vth central node.

[0047] Then, the initial data of the failure probability of each node at each moment can be normalized to obtain the failure probability data of each node at each moment. The normalization processing here can limit the failure probability data to between 0 and 1, so that it can be used to represent the probability. For example, the initial data of the failure probability can be linearly normalized to obtain the failure probability data between 0 and 1.

[0048] The above P V,t The calculation formula represents the difference between the same parameter of each node and the adjacent nodes at a certain moment. The larger the difference, the greater the possibility of failure of the node. Therefore, the data calculated by the above formula can represent the failure probability of the node, that is, the final result is the failure probability data.

[0049] S3, taking each type of parameter as a dimension, constructs the parameters of each node at each moment into multi-dimensional coordinate data, and obtains the multi-dimensional parameter vector of each node at each moment.

[0050] Taking three parameters as an example, the three-dimensional coordinate data of the Vth node at the tth moment (taking current I, temperature T, and operating state S as examples) are I V,t , TV,t , S V,t , the corresponding three-dimensional parameter vector of the Vth node at the tth time is (I V,t ,T V,t ,S V,t ).

[0051] S4, correcting the multi-dimensional parameter vector of the corresponding node at the corresponding moment through the failure probability data of each node at each moment.

[0052] In one embodiment of the present invention, the modified multi-dimensional parameter vector of each node at each time can be calculated by the following formula:

[0053] A′ V,t =P V ' ,t A V,t

[0054] Among them, A′ V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

[0055] The size (modulus) of the multidimensional parameter vector is obtained by taking the square root of the sum of the squares of the coordinate values ​​of each dimension of the multidimensional parameter vector. Therefore, the above formula only corrects the size of the multidimensional parameter vector and does not change the direction (angle) of the multidimensional parameter vector.

[0056] S5, calculating the parameter change degree of each node at each moment according to the corrected multi-dimensional parameter vector of each node at each moment.

[0057] In one embodiment of the present invention, the parameter change degree of each node at each moment can be calculated by the following formula:

[0058]

[0059] Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τThe sizes of the two vectors, α and β are the angle influence coefficient and the size influence coefficient respectively, and both α and β are positive numbers. When the types of parameters collected by the monitored distribution network are different, the values ​​of α and β are different, which can be set through experiments in each specific embodiment. It is inconvenient to limit them to a specific value here. N3 is the total number of moments.

[0060] The above M V,t The calculation formula represents the cosine similarity data of the multidimensional parameter vector of the Vth node at time t and other different times, and the difference in the size of the multidimensional parameter vector of the Vth node at time t and other different times. The weighted result of these two data, the larger the weighted result, the more obvious the parameter change of the Vth node at time t.

[0061] S6, regard each node as a central node. When a certain node is regarded as a central node, the influence data of the central node is calculated according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent nodes of the central node at each moment, so as to obtain the influence data of each node.

[0062] In one embodiment of the present invention, the influence data of the central node can be calculated by the following formula:

[0063]

[0064] Among them, E V Represents the influence data of the Vth central node, M v,t It represents the parameter change degree of the vth adjacent node of the Vth central node at the tth time, N2 is the total number of adjacent nodes of the Vth central node, N3 is the total number at the time, e is a natural constant, and ^ represents exponentiation.

[0065] The above E V The calculation formula includes the difference between the parameter change degree of the central node and the parameter change degree of its adjacent nodes. The larger the difference, the smaller the influence of the central node on its adjacent nodes. The significance of the power operation with the natural constant as the base and the opposite number as the exponent is that the larger the difference in the above parameter change degree, the greater the influence of the central node, that is, the influence data E V The influence of the central node can be represented, and the data that can represent the influence of the node is limited to between 0 and 1.

[0066] S7, modifying the attention weight between two nodes according to the influence data of each two nodes.

[0067] In one embodiment of the present invention, the modified attention weight between every two nodes may be calculated by the following formula:

[0068]

[0069] Among them, w′ i,j represents the corrected attention weight between the i-th node and the j-th node, w i,j represents the attention weight between the i-th node and the j-th node before correction, E i Represents the influence data of the i-th node, E j represents the influence data of the jth node, It represents the mean of the difference between the influence data of all nodes.

[0070] S8, by introducing the graph neural network with attention mechanism, the fault judgment of the monitored distribution network is realized by using the corrected attention weights between every two nodes.

[0071] It is an existing technology to perform fault diagnosis and monitoring of distribution networks by introducing a graph neural network with an attention mechanism, and the specific process will not be repeated here.

[0072] According to the fault monitoring method of the distribution network of the embodiment of the present invention, the time series data of multiple parameters of each node in the monitored distribution network is obtained to calculate the fault probability data of each node at each moment, and the multi-dimensional parameter vector of each node at each moment is corrected accordingly, and then the parameter change degree of each node at each moment and the influence data of each node are calculated in turn, and the attention weight between every two nodes is corrected accordingly, and finally, the corrected attention weight is used to realize the fault judgment of the distribution network by using the graph neural network with the introduction of attention mechanism. Therefore, by obtaining the actual influence of the node, a more accurate attention weight can be corrected. When the graph neural network with the introduction of attention mechanism is used for distribution network fault monitoring, the accuracy of fault judgment in distribution network fault monitoring can be improved, thereby ensuring the effective monitoring of distribution network faults.

[0073] Corresponding to the distribution network fault monitoring method of the above embodiment, the present invention further proposes a distribution network fault monitoring device.

[0074] like Figure 2As shown, the fault monitoring device of the distribution network of the embodiment of the present invention includes an acquisition module 10, a first calculation module 20, a construction module 30, a first correction module 40, a second calculation module 50, a third calculation module 60, a second correction module 70 and a judgment module 80. Among them, the acquisition module 10 is used to obtain the time series data of multiple parameters of each node in the monitored distribution network, wherein the node in the monitored distribution network is a device or component in the monitored distribution network; the first calculation module 20 is used to regard each node as a central node, and when a certain node is regarded as a central node, according to the time series data of the adjacent nodes of the central node, calculate the fault probability data of the central node at each time, so as to obtain the fault probability data of each node at each time; the construction module 30 is used to take each type of parameter as a dimension, construct the parameters of each node at each time into multi-dimensional coordinate data, and obtain the multi-dimensional parameter vector of each node at each time; the first correction module 40 is used to compare the multi-dimensional parameters of the corresponding node at the corresponding time with the fault probability data of each node at each time. The number vector is corrected; the second calculation module 50 is used to calculate the parameter change degree of each node at each moment according to the corrected multi-dimensional parameter vector of each node at each moment; the third calculation module 60 is used to regard each node as a central node, and when a certain node is regarded as a central node, the influence data of the central node is calculated according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent nodes of the central node at each moment, so as to obtain the influence data of each node; the second correction module 70 is used to correct the attention weight between every two nodes according to the influence data of the two nodes; the judgment module 80 is used to realize the fault judgment of the monitored distribution network by introducing a graph neural network with an attention mechanism and using the corrected attention weight between every two nodes.

[0075] In one embodiment of the present invention, the first calculation module 20 may first calculate the initial data of the failure probability of the central node at each moment by the following formula:

[0076]

[0077] Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c It represents the cth parameter of the v+1th neighboring node of the Vth central node at the tth time, N1 is the total number of parameter types, and N2 is the total number of neighboring nodes of the Vth central node.

[0078] Then, the first calculation module 20 may perform normalization processing on the initial data of the failure probability of each node at each moment to obtain the failure probability data of each node at each moment.

[0079] In one embodiment of the present invention, the first correction module 40 can calculate the corrected multi-dimensional parameter vector of each node at each time by the following formula:

[0080] A′ V,t =P V ' ,t A V,t

[0081] Among them, A′ V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

[0082] In one embodiment of the present invention, the second calculation module 50 may calculate the parameter change degree of each node at each moment by the following formula:

[0083]

[0084] Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τ The sizes of the two vectors, α and β are the angle influence coefficient and size influence coefficient respectively, α and β are both positive numbers, and N3 is the total number of moments.

[0085] In one embodiment of the present invention, the third calculation module 60 can calculate the influence data of the central node by the following formula:

[0086]

[0087] Among them, E V Represents the influence data of the Vth central node, M v,tIt represents the parameter change degree of the vth adjacent node of the Vth central node at the tth time, N2 is the total number of adjacent nodes of the Vth central node, N3 is the total number at the time, e is a natural constant, and ^ represents exponentiation.

[0088] In one embodiment of the present invention, the second correction module 70 may calculate the corrected attention weight between every two nodes by the following formula:

[0089]

[0090] Among them, w′ i,j represents the corrected attention weight between the i-th node and the j-th node, w i,j represents the attention weight between the i-th node and the j-th node before correction, E i Represents the influence data of the i-th node, E j represents the influence data of the jth node, It represents the mean of the difference between the influence data of all nodes.

[0091] According to the fault monitoring device of the distribution network of the embodiment of the present invention, the time series data of multiple parameters of each node in the monitored distribution network is obtained to calculate the fault probability data of each node at each moment, and the multi-dimensional parameter vector of each node at each moment is corrected accordingly, and then the parameter change degree of each node at each moment and the influence data of each node are calculated in turn, and the attention weight between every two nodes is corrected accordingly, and finally the corrected attention weight is used to realize the fault judgment of the distribution network by using the graph neural network with the introduction of attention mechanism. Therefore, by obtaining the actual influence of the node, a more accurate attention weight can be corrected. When the graph neural network with the introduction of attention mechanism is used for distribution network fault monitoring, the accuracy of fault judgment in distribution network fault monitoring can be improved, thereby ensuring the effective monitoring of distribution network faults.

[0092] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. "Multiple" means two or more, unless otherwise clearly and specifically defined.

[0093] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0094] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0095] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0096] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0098] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0100] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0101] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A fault monitoring method for a distribution network, characterized in that: The following steps are involved: Acquire time series data of multiple parameters of each node in the monitored distribution network, wherein the node in the monitored distribution network is a device or component in the monitored distribution network; Each of the nodes is regarded as a central node. When a certain node is regarded as a central node, the failure probability data of the central node at each moment is calculated according to the time series data of the adjacent nodes of the central node, so as to obtain the failure probability data of each node at each moment; Taking each type of the parameter as a dimension, constructing the parameter of each node at each moment into multi-dimensional coordinate data, and obtaining a multi-dimensional parameter vector of each node at each moment; The multi-dimensional parameter vector of the corresponding node at the corresponding time is corrected by using the failure probability data of each node at each time; According to the corrected multi-dimensional parameter vector of each node at each moment, the parameter change degree of each node at each moment is calculated; Considering each of the nodes as a central node, when a certain node is considered as a central node, the influence data of the central node is calculated according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent nodes of the central node at each moment, thereby obtaining the influence data of each node; The attention weight between two nodes is modified according to the influence data of each two nodes; By introducing a graph neural network with an attention mechanism, the fault judgment of the monitored distribution network is realized by using the corrected attention weights between every two nodes.

2. The method for monitoring the fault of a power distribution network according to claim 1, characterized in that: The initial data of the central node's failure probability at each moment is calculated using the following formula: Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c represents the cth parameter of the v+1th neighboring node of the Vth central node at the tth time, N1 is the total number of parameter types, and N2 is the total number of neighboring nodes of the Vth central node; The initial data of the failure probability of each node at each moment are normalized to obtain the failure probability data of each node at each moment.

3. The method for monitoring the fault of the distribution network according to claim 2, characterized in that: The modified multidimensional parameter vector of each node at each moment is calculated by the following formula: A′ V,t =P′ V,t A V,t Among them, A′ V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

4. The method for monitoring faults in a power distribution network according to claim 3, characterized in that: The degree of parameter change of each node at each moment is calculated by the following formula: Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τ The sizes of the two vectors, α and β are the angle influence coefficient and size influence coefficient respectively, and N3 is the total number of moments.

5. The method for monitoring faults in a power distribution network according to claim 4, characterized in that: The influence data of the central node is calculated by the following formula: Among them, E V Represents the influence data of the Vth central node, M v,t It represents the parameter change degree of the vth adjacent node of the Vth central node at the tth time, N2 is the total number of adjacent nodes of the Vth central node, N3 is the total number at the time, e is a natural constant, and ^ represents exponentiation.

6. The method for monitoring faults in a power distribution network according to claim 5, characterized in that: The corrected attention weight between every two nodes is calculated by the following formula: Among them, w′ i,j represents the corrected attention weight between the i-th node and the j-th node, w i,j represents the attention weight between the i-th node and the j-th node before correction, E i Represents the influence data of the i-th node, E j represents the influence data of the jth node, It represents the mean of the difference between the influence data of all nodes.

7. A fault monitoring device for a distribution network, characterized in that: include: An acquisition module, used for acquiring time series data of multiple parameters of each node in the monitored distribution network, wherein the node in the monitored distribution network is a device or component in the monitored distribution network; A first calculation module is used to regard each of the nodes as a central node. When a certain node is regarded as a central node, the failure probability data of the central node at each moment is calculated according to the time series data of the adjacent nodes of the central node, so as to obtain the failure probability data of each node at each moment; A construction module, used to construct the parameters of each node at each moment into multi-dimensional coordinate data with each type of the parameter as a dimension, and obtain a multi-dimensional parameter vector of each node at each moment; A first correction module is used to correct the multi-dimensional parameter vector of the corresponding node at the corresponding time by using the failure probability data of each node at each time; The second calculation module is used to calculate the parameter change degree of each node at each moment according to the corrected multi-dimensional parameter vector of each node at each moment; A third calculation module is used to regard each of the nodes as a central node. When a certain node is regarded as a central node, the influence data of the central node is calculated according to the parameter change degree of the central node at each moment and the parameter change degree of the adjacent nodes of the central node at each moment, so as to obtain the influence data of each node; The second correction module is used to correct the attention weight between every two nodes according to the influence data of the two nodes; The judgment module is used to realize the fault judgment of the monitored distribution network by introducing a graph neural network with an attention mechanism and using the corrected attention weights between every two nodes.

8. The fault monitoring device for a power distribution network according to claim 7, characterized in that: The first calculation module calculates the initial data of the failure probability of the central node at each moment by the following formula: Among them, V, t, v, and c represent the central node number, time number, adjacent node number, and parameter type number respectively. V,t represents the initial data of the failure probability of the Vth central node at the tth time, a v,c represents the cth parameter of the vth neighboring node of the Vth central node at time t, a v+1,c represents the cth parameter of the v+1th neighboring node of the Vth central node at the tth time, N1 is the total number of parameter types, N2 is the total number of neighboring nodes of the Vth central node, The first calculation module normalizes the initial data of the failure probability of each node at each moment to obtain the failure probability data of each node at each moment.

9. The fault monitoring device for a power distribution network according to claim 8, characterized in that: The first correction module calculates the corrected multidimensional parameter vector of each node at each time by the following formula: A′ V,t =P V ′ ,t A V,t Among them, A′ V,t represents the modified multidimensional parameter vector of the Vth node at the tth time, A V,t represents the multidimensional parameter vector of the Vth node at time t before correction, P′ V,t Represents the failure probability data of the Vth node at the tth time.

10. The fault monitoring device for a power distribution network according to claim 9, characterized in that: The second calculation module calculates the parameter change degree of each node at each moment by the following formula: Among them, M V,t A′ represents the parameter change degree of the Vth node at the tth time. V,τ represents the modified multidimensional parameter vector of the Vth node at the τth time, CSIM(A′ V,t ,A′ V,τ ) represents A′ V,t , A′ V,τ The cosine similarity of two vectors, |A′ V,t |、|A′ V,τ | respectively represent A′ V,t , A′ V,τ The sizes of the two vectors, α and β are the angle influence coefficient and size influence coefficient respectively, and N3 is the total number of moments.

Images