SA-CNN-BiLSTM-based intelligent substation virtual loop automatic generation method

Abstract

The invention discloses an SA-CNN-BiLSTM-based intelligent substation virtual loop automatic generation method, which comprises the following steps: in an SCD file, extracting IED equipment information of historical IED equipment and virtual terminal information contained in the historical IED equipment, and storing the IED equipment information and the virtual terminal information in a historical SCD database; forming a semantic similarity model; and based on the semantic similarity model, finding a final historical IED equipment association pair corresponding to the to-be-designed IED equipment association pair, and instantiating the to-be-designed IED equipment association pair based on the final historical IED equipment association pair. According to the method, the virtual terminator information contained in the historical SCD file is rapidly extracted in a DOM mode and stored in the database in a classified mode, and the query and matching speed is remarkably increased. The semantic similarity model can effectively extract semantic features and endow the semantic features with higher weights, the semantic recognition success rate is high, and the method has a good effect in intelligent substation virtual connection automatic design.

Description

Technical Field

[0001] The invention belongs to the field of electric power technology, and in particular relates to a method for automatically generating a virtual circuit of an intelligent substation based on SA-CNN-BiLSTM. Technical Background

[0002] Smart grid has a profound impact on multiple core links such as power generation, transmission, substation, distribution, information communication, and system scheduling. As an indispensable part of smart grid, smart substation has ushered in a wide and in-depth construction wave in recent years. However, with the continuous expansion of the construction scale of smart substations, the problems faced in its design and implementation stages are becoming increasingly prominent and need to be solved urgently. The substation configuration description (SCD) file is the core configuration file, which describes in detail the instance configuration and communication parameter information of all IEDs and the communication configuration between IEDs. Its configuration efficiency and accuracy will be directly related to the debugging process of the entire smart substation. However, the existing virtual circuit configuration is mostly carried out manually. At the same time, due to the lack of relevant knowledge such as IEC61850 and XML, the virtual circuit configuration is inefficient and has a high error rate. For this reason, it is urgent to carry out research on the automatic design method of virtual circuits in smart substations to improve the efficiency and accuracy of virtual circuit configuration.

[0003] At present, researchers at home and abroad have conducted research on the automatic design, visualization and other fields of virtual circuits. The ongoing automatic generation of virtual connections is roughly divided into two steps: building a historical database and matching virtual terminals. For the construction of the historical database, since it is manually constructed, there is a lot of subjectivity when constructing the keyword dictionary. Under non-standard naming conditions, it is difficult to accurately match, so it is not universal. For virtual terminal matching, the target connection is found in the database mainly through device coding and string-fuzzy matching. However, since the semantic features of the text cannot be accurately extracted, the matching accuracy of similar text information and different semantic features is not high. In view of the above problems, the present invention proposes a data-driven intelligent substation virtual circuit intelligent design method. The virtual terminal information of the historical SCD file is quickly extracted in DOM mode to build a historical SCD database; a semantic similarity model with SA-CNN-BiLSTM as the core is established; based on the model, a virtual circuit automatic generation and virtual circuit autonomous verification scheme is proposed, which significantly improves the efficiency and accuracy of virtual terminal configuration of smart substations. Summary of the invention

[0004] The purpose of the present invention is to provide a method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM in order to solve the problems of complicated configuration of virtual connections and low degree of automation in smart substations.

[0005] The above-mentioned purpose of the present invention can be achieved by the following technical means:

[0006] A method for automatically generating a virtual circuit in a smart substation based on SA-CNN-BiLSTM comprises the following steps:

[0007] Step 1: In the SCD file, extract the IED device information of the historical IED device and the virtual terminal information it contains and store them in the historical SCD database;

[0008] Step 2: Construct a semantic similarity model;

[0009] Step 3: Based on the semantic similarity model, the final historical IED device association pair corresponding to the IED device association pair to be designed is found, and the IED device association pair to be designed is instantiated based on the final historical IED device association pair.

[0010] As mentioned above, the IED device information of the historical IED device and the virtual terminal information contained therein are classified and stored in the historical SCD database according to voltage level, device type and transmission mode.

[0011] As mentioned above, the semantic similarity model includes the BERT model, the CNN convolutional model, the SA-BiLSTM module, the vector concatenation layer, the fully connected layer, and the output layer.

[0012] The BERT model converts input information into a semantic information vector in vector form;

[0013] The CNN convolution model includes a CNN convolution layer and a pooling layer. The semantic information vector output by the BERT model is input into the CNN convolution layer, and the pooling layer is used to sample the output features of the CNN convolution layer.

[0014] The output feature vector of the CNN convolutional model is input into the SA-BiLSTM module;

[0015] After the two pieces of information to be compared are processed by the BERT model, the CNN convolutional model, and the SA-BiLSTM module in sequence, the SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are obtained. The SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are input into the vector concatenation layer, and the text feature vectors are obtained as output.

[0016] The text feature vector is input into the fully connected layer and compressed through the output of the fully connected layer;

[0017] The output layer uses the SoftMax function to calculate the output vector of the fully connected layer to obtain the similarity of the two pieces of information to be compared.

[0018] As described above, finding the final historical IED device association pair corresponding to the IED device pair to be designed includes the following steps:

[0019] Step 3.1: A pair of IED devices to be designed connected by a virtual circuit is formed into an IED device association pair to be designed, and IED device information and virtual terminal information of two IED devices to be designed in the IED device association pair to be designed are parsed;

[0020] Step 3.2: define the two IED devices to be designed in the IED device association pair, namely the IED device to be designed at the local end and the IED device to be designed at the opposite end.

[0021] The semantic similarity model of SA-CNN-BiLSTM is used to select IED devices in the historical SCD database whose IED device information similarity is greater than the similarity threshold as pre-matching local historical IED devices, and multiple pre-matching opposite-end historical IED devices connected to the virtual circuits that match the local historical IED devices are found in the historical SCD database;

[0022] Using the semantic similarity model of SA-CNN-BiLSTM, the peer historical IED devices whose IED device information similarity is greater than the similarity threshold are selected as pre-matched peer historical IED devices.

[0023] Each pair of pre-matched local historical IED devices and pre-matched opposite historical IED devices is respectively regarded as a pair of historical IED device association pairs;

[0024] Step 3.3: For one pair of historical IED device association pairs: virtual terminal association is performed between the virtual terminal information of the IED device to be designed at the local end and the virtual terminal information of the pre-matched historical IED device at the local end; virtual terminal association is performed between the virtual terminal information of the pre-matched historical IED device at the opposite end connected to the virtual circuit of the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the IED device to be designed at the opposite end; based on the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the pre-matched historical IED device at the opposite end that are successfully associated with the virtual terminal, the number of virtual circuit associations of the association pair of the IED device to be designed is obtained,

[0025] Traverse each pair of historical IED device association pairs, obtain the historical IED device association pair with the largest virtual circuit association number as the final historical IED device association pair, and instantiate the IED device association pair to be designed based on the final historical IED device association pair.

[0026] The basis for associating the virtual terminals as described above is that the similarity of the virtual terminal information is the highest and is greater than a set similarity threshold.

[0027] A method for automatically generating a virtual circuit of an intelligent substation based on SA-CNN-BiLSTM comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the generation method when executing the computer program.

[0028] A computer-readable storage medium stores a computer program, which implements the steps of the above-mentioned generation method when executed by a processor.

[0029] A computer program product comprises a computer program, which implements the steps of the above-mentioned generation method when executed by a processor.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] (1) This method quickly extracts the virtual terminal information contained in the historical SCD file through the DOM method and stores it in the database in a classified manner, which significantly improves the query and matching speed.

[0032] (2) The semantic similarity model constructed by this method can effectively extract semantic features and assign higher weights to them. The success rate of semantic recognition is high and it has a good effect in the automatic design of virtual connections in smart substations.

[0033] (3) By manually constructing a standard template library, the automatically generated virtual circuits are verified. On the one hand, the accuracy of automatic generation of virtual circuits is improved, and on the other hand, the verification workload of designers is significantly reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the process of the present invention;

[0035] Figure 2 This is a schematic diagram of the SCD file of a 220kV smart substation;

[0036] Figure 3 The IED equipment information and internal virtual connection relationship of 220kV busbar measurement and control set A;

[0037] Figure 4 It is the SV receiving virtual terminal table of 220kV busbar measurement and control set A;

[0038] Figure 5 This is a comparison chart of the effects of three semantic similarity models;

[0039] Figure 6 It is the virtual terminal information after autonomous verification;

[0040] Figure 7 Send a virtual connection for GOOSE of MM2255B220kV busbar PT merging unit B set PCS221BB". DETAILED DESCRIPTION

[0041] In order to facilitate those skilled in the art to understand and implement the present invention, the present invention is further described in detail below in conjunction with implementation examples. The implementation examples described herein are only used to illustrate and explain the present invention, and are not intended to limit the present invention.

[0042] Embodiment 1:

[0043] A method for automatically generating a virtual circuit in a smart substation based on SA-CNN-BiLSTM specifically comprises the following steps:

[0044] Step 1: Use DOM to convert the SCD file into a tree structure, and only access IED nodes and key nodes containing their description information to reduce reading time. Extract the IED device information of the historical IED device and the virtual terminal information it contains, and store the extracted IED device information of the historical IED device and the virtual terminal information it contains in the historical SCD database according to voltage level, device type and transmission mode. The IED device information includes device name, device description, device type, device manufacturer, and version number.

[0045] Step 2: Integrate the advantages of self-attention mechanism, convolutional neural network, and bidirectional long short-term memory neural network to form a semantic similarity model with SA-CNN-BiLSTM as the core. The semantic similarity model includes:

[0046] (1) BERT model:

[0047] The BERT model is used to convert the input information into a semantic information vector in vector form. By jointly adjusting the bidirectional Transformer in all layers for training, the semantic information of the input text is fully extracted and the semantic information vector is output.

[0048] Among them, T rm is the Transformer encoder; E 1 、E 2 , L, E n is the input vector of the BERT model; T 1 , T 2 , L, T n It is the semantic information vector output by the BERT model.

[0049] (2) CNN convolutional model:

[0050] The CNN convolutional model includes CNN convolutional layers and pooling layers.

[0051] The semantic information vector output by the BERT model is input into the CNN convolutional layer to extract the features of the semantic information vector. The calculation process is shown in formula (1).

[0052]

[0053] Among them, C is the output feature of the CNN convolutional layer; is the convolution operation; W is the weight vector of the convolution kernel; b represents the offset; f is the activation function; X 1 is the input semantic information vector.

[0054] The output features of the CNN convolutional layer are sampled using the pooling layer to retain its important information features. The output vector of the pooling layer is used as the output feature vector of the CNN convolutional model. The height and width of the output feature vector are shown in the following formula:

[0055]

[0056] Among them, K H and K w Represents the height and width of the pooling kernel; I H and I W Represents the height and width of the output features of the CNN convolutional layer; S represents the stride of the pooling kernel; O H and O W Represent the height and width of the output feature vector respectively.

[0057] (3) SA-BiLSTM module:

[0058] The SA-BiLSTM module includes two reverse LSTM units, namely the forward LSTM network and the backward LSTM network. The output feature vector of the CNN convolutional model is input into the forward LSTM network and the backward LSTM network for calculation, and the self-attention mechanism is combined to give higher weight to the key information, and finally the output feature vector of the SA-BiLSTM module is obtained.

[0059] (4) Vector concatenation layer:

[0060] After the two pieces of information to be compared (the two IED device information to be compared, or the two virtual terminal information to be compared) are processed by the BERT model, the CNN convolutional model, and the SA-BiLSTM module respectively, the SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are obtained. The SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are input into the vector concatenation layer, and the text feature vectors are output.

[0061] (5) Fully connected layer:

[0062] The text feature vector is input into the fully connected layer, and the output of the fully connected layer is compressed to convert the high-dimensional feature vector into a low-dimensional feature vector. The calculation process is as follows:

[0063] The text feature vector Y output by the vector concatenation layer is combined with the weight matrix W of the fully connected layer d Calculate according to formula (4).

[0064] M = tanh(W d ·Y+b d ) (4)

[0065] where b d is the bias parameter of the fully connected layer; M is the output vector of the fully connected layer.

[0066] (6) Output layer:

[0067] The SoftMax function is used to calculate the output vector of the fully connected layer to obtain the similarity of the two pieces of information to be compared. The calculation process is shown in formula (5).

[0068] probability=softmax(W x ·M+b) (5)

[0069] Among them, W x is the weight matrix of the output layer; M is the output vector of the fully connected layer; b is the bias function of the output layer; probability is the similarity of the two pieces of information to be compared.

[0070] Step 3: Use the SA-CNN-BiLSTM-based semantic similarity model to automatically generate virtual loops. The steps are as follows:

[0071] Step 3.1: A pair of IED devices to be designed that are connected by a virtual loop (without specific virtual terminal information, but clearly associated with virtual terminals) are combined into an IED device association pair to be designed, and the ICD file corresponding to the IED device to be designed in the IED device association pair to be designed is parsed to extract the IED device information to be designed and the virtual terminal information. The ICD file is a device performance description file provided by the device manufacturer, which includes the technical data model and services provided by the IED, but does not include the actual IED name and communication parameters, so it needs to be further instantiated, that is, to realize the virtual connection between the IED devices to be designed associated with the virtual terminals. By parsing the ICD file, the IED device information to be designed and the virtual terminal information it contains can be obtained. The IED device information includes the device name, device description, device type, device manufacturer, and version number.

[0072] Step 3.2: define the two IED devices to be designed of the associated pair of the IED devices to be designed, namely the IED device to be designed at the local end and the IED device to be designed at the opposite end. Use the semantic similarity model of SA-CNN-BiLSTM to select the IED devices in the historical SCD database whose IED device information similarity is greater than the similarity threshold as the pre-matched historical IED devices at the local end. Find multiple pre-matched historical IED devices at the opposite end connected to the virtual circuit matching the historical IED device at the local end in the historical SCD database. The IED device information includes device name, device description, device type, device manufacturer, and version number. The similarity threshold can be selected as 0.7.

[0073] One local IED device to be designed may correspond to multiple pre-matched local historical IED devices that meet the similarity threshold; one pre-matched local historical IED device may be connected to multiple opposite historical IED devices through a virtual circuit;

[0074] Furthermore, the semantic similarity model of SA-CNN-BiLSTM is used to select the historical IED devices on the opposite end whose IED device information similarity is greater than the similarity threshold as the pre-matching historical IED devices on the opposite end relative to the IED device to be designed.

[0075] Through the above steps, multiple pairs of pre-matched local historical IED devices and pre-matched opposite historical IED devices corresponding to the local IED device to be designed can be obtained, and each pair of pre-matched local historical IED devices and pre-matched opposite historical IED devices is respectively used as a pair of historical IED device association pairs.

[0076] As shown in formula (6), the IED device A to be designed at this end m With the IED device A to be designed at the other end n The association pairs formed are the association pairs of IED devices to be designed. The semantic similarity model based on SA-CNN-BiLSTM is used to find the historical IED device association pairs Q with IED device information similarity greater than 0.7 in the historical SCD database of the same type. 1 Q 2 LQ n If no matching pairs that meet the conditions can be found, the query conditions are not limited and only the device name is used for filtering.

[0077] (A m , A n )→{Q 1 Q 2 LQ n}→

[0078]

[0079] Where n is the number of historical IED device association pairs screened; a is the virtual circuit contained in the i-th historical IED device association pair, 1≤t≤n; i is the number of virtual circuits contained in the i-th historical IED device association pair; They are respectively from the first virtual circuit of the i-th historical IED device association pair to the first sending virtual terminal; They are respectively the receiving virtual terminals of the i-th historical IED device associated pair.

[0080] Step 3.3: For one pair of historical IED device association pairs: virtual terminal association is performed between the virtual terminal information of the IED device to be designed at the local end and the virtual terminal information of the pre-matched historical IED device at the local end; virtual terminal association is performed between the virtual terminal information of the pre-matched historical IED device at the opposite end connected to the virtual circuit of the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the IED device to be designed at the opposite end; the basis for the virtual terminal association is: the similarity of the virtual terminal information is the highest and is greater than the set similarity threshold; based on the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the pre-matched historical IED device at the opposite end that has been successfully associated, the number of successfully associated virtual circuit associations is obtained. Then the number of successfully associated virtual circuit associations of the IED device association pair to be designed is obtained.

[0081] Traverse each pair of historical IED device association pairs, obtain the historical IED device association pair with the largest virtual circuit association number as the final historical IED device association pair, and instantiate the IED device association pair to be designed based on the final historical IED device association pair.

[0082] The standard template library is used to realize the automatic verification of virtual circuits. In order to reduce the workload of designers to verify the automatically generated virtual circuits, the standard template library is constructed according to the voltage level and equipment type. In addition, when constructing the standard template library, only the key information of the IED equipment is stored, and irrelevant information is manually removed to improve the verification accuracy. The steps for autonomous verification of virtual circuits are as follows:

[0083] Search for a suitable template in the standard template library based on the voltage level, device type, and other information of the IED device to be designed.

[0084] The virtual terminal information of the automatically instantiated IED device to be designed is input into the semantic similarity model for comparison with the virtual terminal information contained in the standard template to determine whether the connection is correct.

[0085] According to the judgment results, the automatically generated virtual connections are divided into three categories: normal, missing, and wrong, and the designer is prompted to make corrections.

[0086] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0087] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0088] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0089] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0090] Embodiment 2:

[0091] Take the SCD file of a well-designed 220kV smart substation as an example for verification. Some codes are as follows: Figure 2 Parse the SCD file to obtain the 220kV busbar measurement and control A set of IED equipment information and virtual terminal information as shown below. Figure 3 The SV receiving virtual terminal table of the 220kV busbar measurement and control set A is as follows: Figure 4 shown.

[0092] Steps 1-3 in Example 1 were carried out.

[0093] The semantic similarity model is used to match IED devices and virtual terminals in the historical SCD database. Taking the similarity matching of some IED names as an example, the similarity judgment results are shown in Table 1.

[0094] Table 1 Matching status of some IED devices

[0095] Historical IED Name Target IED Name Similarity Matching 220KV Dashi Line A Set 220KV Zhenghe Line A Set 0.97 yes 220KV Dashi Line A Set 220KV busbar measurement and control set A 0.65 no 220KV Dashi Line A Set 35KV busbar measurement and control set A 0.21 no 220KV Dashi Line A Set 220KV busbar measurement and control set A 0.47 no

[0096] For some IED devices that are similar but should not be matched, such as "220kV Dashi Line A Set" and "220kV Busbar Measurement and Control A Set", when extracting semantic features, "line" and "busbar" are extracted as separate semantic features and given a larger weight for similarity calculation. The final result fails to exceed the threshold and cannot be matched successfully, which is consistent with the actual situation; for devices that should be matched, such as "220kV Dashi Line A Set" and "220kV Zhenghe Line A Set", more attention is paid to "220kV", "line" and "A Set" when extracting semantic features. Therefore, the similarity calculated in the end is larger, the match is successful, which is consistent with the actual situation; when the historical IED name is significantly different from the target IED name, the similarity is also lower, such as "220kV Dashi Line A Set" and "35kV Busbar Measurement and Control A Set", which cannot be matched successfully, which is consistent with the actual situation. It can be seen that the semantic similarity model constructed in this paper can effectively extract semantic features and assign higher weights to them. The success rate of semantic recognition is high, and it has a good effect in the automatic design of virtual connections in smart substations.

[0097] In order to further illustrate the advantages of the semantic similarity model constructed by the present invention, a semantic similarity model based on a recurrent neural network (RNN) and a long short-term memory neural network (LSTM) is used to automatically design virtual connections. The results are shown in FIG. Figure 5 shown.

[0098] It can be seen that the application of recurrent neural network (RNN) in semantic recognition is affected by factors such as gradient explosion and gradient decay, and the virtual terminal cannot accurately extract information features, and the success rate of virtual connection is very low; while the long short-term memory neural network (LSTM) has made improvements on this basis and overcome this problem to a certain extent, but it still fails to fully extract semantic features; however, in the invention, we use a bidirectional long short-term memory neural network (BiLSTM) to extract semantic features well, and introduce a v attention mechanism to give important information a higher weight for comparison. The matching success rate is independent of the device type and has high efficiency.

[0099] In order to improve the readability of virtual circuits and facilitate designers to add, delete, and modify virtual connections, this method uses visualization and other technologies to build a relevant platform to display the verified virtual connections, such as Figure 6 As shown. GOOSE communication is indicated by a blue arrow, and SV communication is indicated by a yellow arrow. The arrow shows the number of virtual circuits between the two devices. The red "×" indicates that there is an error in the current virtual circuit. Click to further view the GOOSE sent virtual connection list of "MM2255B220kV busbar PT merging unit B set PCS221BB", such as Figure 7 shown.

[0100] Virtual circuits that are detected to be erroneous are marked in red to facilitate designers to modify them.

[0101] It should be noted that the specific embodiments described in the present invention are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the attached claims.

Claims

1. A method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM, characterized in that: The following steps are involved: Step 1: In the SCD file, extract the IED device information of the historical IED device and the virtual terminal information it contains and store them in the historical SCD database; Step 2: Construct a semantic similarity model; Step 3: Based on the semantic similarity model, the final historical IED device association pair corresponding to the IED device association pair to be designed is found, and the IED device association pair to be designed is instantiated based on the final historical IED device association pair.

2. According to claim 1, a method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM is characterized in that: The IED device information of the historical IED device and the virtual terminal information contained therein are classified and stored in the historical SCD database according to voltage level, device type and transmission mode.

3. According to claim 1, a method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM is characterized in that: The semantic similarity model includes a BERT model, a CNN convolutional model, a SA-BiLSTM module, a vector concatenation layer, a fully connected layer, and an output layer in sequence. The BERT model converts input information into a semantic information vector in vector form; The CNN convolution model includes a CNN convolution layer and a pooling layer. The semantic information vector output by the BERT model is input into the CNN convolution layer, and the pooling layer is used to sample the output features of the CNN convolution layer. The output feature vector of the CNN convolutional model is input into the SA-BiLSTM module; After the two pieces of information to be compared are processed by the BERT model, the CNN convolutional model, and the SA-BiLSTM module in sequence, the SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are obtained. The SA-BiLSTM module output feature vectors corresponding to the two pieces of information to be compared are input into the vector concatenation layer, and the text feature vectors are obtained as output. The text feature vector is input into the fully connected layer and compressed through the output of the fully connected layer; The output layer uses the SoftMax function to calculate the output vector of the fully connected layer to obtain the similarity of the two pieces of information to be compared.

4. According to claim 3, a method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM is characterized in that: The method of finding the final historical IED device association pair corresponding to the IED device pair to be designed comprises the following steps: Step 3.1: A pair of IED devices to be designed connected by a virtual circuit is formed into an IED device association pair to be designed, and IED device information and virtual terminal information of two IED devices to be designed in the IED device association pair to be designed are parsed; Step 3.2: define the two IED devices to be designed in the IED device association pair, namely the IED device to be designed at the local end and the IED device to be designed at the opposite end. The semantic similarity model of SA-CNN-BiLSTM is used to select IED devices in the historical SCD database whose IED device information similarity is greater than the similarity threshold as pre-matching local historical IED devices, and multiple pre-matching opposite-end historical IED devices connected to the virtual circuits that match the local historical IED devices are found in the historical SCD database; Using the semantic similarity model of SA-CNN-BiLSTM, the peer historical IED devices whose IED device information similarity is greater than the similarity threshold are selected as pre-matched peer historical IED devices. Each pair of pre-matched local historical IED devices and pre-matched peer historical IED devices is respectively regarded as a pair of historical IED device association pairs; Step 3.3: For one pair of historical IED device association pairs: virtual terminal association is performed between the virtual terminal information of the IED device to be designed at the local end and the virtual terminal information of the pre-matched historical IED device at the local end; virtual terminal association is performed between the virtual terminal information of the pre-matched historical IED device at the opposite end connected to the virtual circuit of the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the IED device to be designed at the opposite end; based on the virtual terminal information of the pre-matched historical IED device at the local end and the virtual terminal information of the pre-matched historical IED device at the opposite end that are successfully associated with the virtual terminal, the number of virtual circuit associations of the association pair of the IED device to be designed is obtained, Traverse each pair of historical IED device association pairs, obtain the historical IED device association pair with the largest virtual circuit association number as the final historical IED device association pair, and instantiate the IED device association pair to be designed based on the final historical IED device association pair.

5. According to claim 4, a method for automatically generating virtual circuits in smart substations based on SA-CNN-BiLSTM is characterized in that: The basis for associating the virtual terminals is that the similarity of the virtual terminal information is the highest and is greater than a set similarity threshold.

6. A method for automatically generating a virtual circuit in a smart substation based on SA-CNN-BiLSTM, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the generation method according to any one of claims 1 to 5 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the generation method according to any one of claims 1 to 5 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the generation method according to any one of claims 1 to 5 are implemented.

Images