Safety protection test method and system for breakdown of flexible direct current converter valve lightning arrester

Abstract

The invention discloses a safety protection test method and system for breakdown of a flexible direct current converter valve arrester, and the method comprises the steps: obtaining test data based on a test scheme-test system, and carrying out the data analysis through a data analysis model, and obtaining the test prompt information of the test data; determining reference test information matched with the test prompt information; performing test result analysis processing through a test result analysis model based on the reference test information and the test data to obtain a first test result; the data analysis model is obtained through sample data training, the sample data is data generated by the test result analysis model, data analysis processing performed by the data analysis model at least comprises test result analysis processing performed by the test result analysis model, and test prompt information obtained through the data analysis model can be used as a basis. And the obtained reference test information is used as reference information to assist the test result analysis model in analyzing the test data, so that the test result analysis accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a safety protection test method and system for breakdown of a flexible DC converter valve arrester. Background Art

[0002] The flexible direct current transmission system used in the existing power transmission technology usually has the characteristics of high voltage level, large capacity transmission and flexible control. The flexible direct current converter valve is the core equipment of the flexible direct current transmission system, and the lightning arrester is one of the key protection devices of the flexible direct current converter valve, which is used to prevent lightning overvoltage from damaging the system equipment. However, in the actual operation process, the lightning arrester may have a breakdown failure, causing the system to fail to operate normally.

[0003] In the related art, after completing the simulation test of lightning arrester breakdown, in order to improve the efficiency of test result analysis, a pre-trained artificial intelligence model is often used to analyze the test data. However, the test sample data used in model training is limited, resulting in the possibility that the analyzed test results may not be accurate enough. Summary of the invention

[0004] In order to solve the above technical problems, the embodiments of the present application propose a safety protection test method and system for breakdown of a flexible DC converter valve arrester, which can improve the accuracy of test result analysis.

[0005] In a first aspect, an embodiment of the present application provides a safety protection test method for breakdown of a flexible DC converter valve arrester, comprising: Based on the test data measured by the safety protection test device, data analysis and processing are performed through a data analysis model to obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of the flexible DC converter valve arrester; Determining reference test information that matches the test prompt information; Based on the reference test information and the test data, performing test result analysis and processing through a test result analysis model to obtain a first test result; The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

[0006] Optionally, the reference test information includes first reference test information and second reference test information, and the determining of the reference test information matching the test prompt information includes: Dividing the test prompt information into a plurality of prompt information blocks, and performing feature extraction on each prompt information block to obtain a corresponding first feature and a second feature, wherein the first feature and the second feature respectively represent different test information; Clustering the first features corresponding to the plurality of prompt information blocks to obtain at least one first feature cluster, and clustering the second features corresponding to the plurality of prompt information blocks to obtain at least one second feature cluster; Based on the at least one first feature cluster, first reference test information is determined, and based on the at least one second feature cluster, second reference test information is determined.

[0007] Optionally, the determining first reference test information based on the at least one first feature cluster, and the determining second reference test information based on the at least one second feature cluster, comprises: fusing each first feature in each first feature cluster to obtain a third feature corresponding to each first feature cluster, and fusing each second feature in each second feature cluster to obtain a fourth feature corresponding to each second feature cluster; In a preset first standard test database, information matching each of the third features is obtained as first standard test information of a corresponding first feature cluster, and in a preset second standard test database, information matching each of the fourth features is obtained as second standard test information of a corresponding second feature cluster; The first standard test information of each of the at least one first feature cluster is integrated to obtain first reference test information, and the second standard test information of each of the at least one second feature cluster is integrated to obtain second reference test information.

[0008] Optionally, extracting features from each prompt information block to obtain corresponding first features and second features includes: Each prompt information block is encoded by an encoder to obtain a corresponding first feature and a second feature.

[0009] Optionally, the sample data includes sample prompt information and sample test results, and the training process of the data analysis model includes: Obtain the data analysis model to be trained, as well as sample test data and its corresponding test result labels; Inputting the sample test data into the test result analysis model to obtain the sample prompt information and the sample test result output by the test result analysis model; Based on the test result label, the sample prompt information and the sample test result, the data analysis model to be trained is trained to obtain the data analysis model.

[0010] Optionally, the test result analysis model includes a large model, and the test result analysis and processing is performed by the test result analysis model based on the reference test information and the test data to obtain the first test result, including: Based on the reference test information, determining model prompt information for the large model; The model prompt information and the test data are input into the large model, so that the large model, under the guidance of the model prompt information, performs reasoning according to the test data, thereby completing the test result analysis and processing to obtain the first test result.

[0011] Optionally, the method further comprises: Acquiring test description information associated with the test data input by a user; The step of determining model prompt information for the large model based on the reference test information includes: Based on the test description information, the reference test information is cleaned to obtain the model prompt information.

[0012] Optionally, the method further comprises: Performing correlation detection on the test data and the first test result by using the data analysis model; Based on the detected correlation information, a second test result is determined.

[0013] Optionally, determining the second test result based on the detected correlation information includes: In a case where the correlation information satisfies a preset correlation condition, taking the first test result as the second test result; In the case where the correlation information does not satisfy a preset correlation condition, feeding back the correlation information to the data analysis model, so that the data analysis model updates the test prompt information according to the correlation information; The updated test prompt information is used as new test prompt information to re-execute the step of determining reference test information matching the test prompt information until the re-obtained correlation information meets the preset correlation condition.

[0014] In a second aspect, an embodiment of the present application provides a safety protection test system for breakdown of a flexible DC converter valve arrester, comprising: A data analysis module, configured to perform data analysis and processing through a data analysis model based on test data measured by a safety protection test device, and obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of a flexible DC converter valve arrester; A reference test information determination module, used to determine reference test information matching the test prompt information; A first test result acquisition module, configured to perform test result analysis and processing based on the reference test information and the test data through a test result analysis model to obtain a first test result; The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

[0015] In summary, the embodiments of the present application have at least the following beneficial effects: According to an embodiment of the present application, test prompt information corresponding to the test data is obtained by performing data analysis and processing through a data analysis model based on test data measured by a safety protection test device, wherein the safety protection test device is used to simulate the breakdown of a flexible DC converter valve arrester; reference test information matching the test prompt information is determined; based on the reference test information and the test data, test result analysis and processing is performed through a test result analysis model to obtain a first test result; wherein the data analysis model is obtained through sample data training, and the sample data is data generated by the test result analysis model, and the data analysis and processing performed by the data analysis model at least includes the test result analysis and processing performed by the test result analysis model, and the test prompt information obtained through the data analysis model can be used as a basis to obtain reference test information, which is used as reference information to assist the test result analysis model in analyzing the test data, thereby improving the accuracy of the test result analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a safety protection test method for breakdown of a flexible DC converter valve arrester provided in an embodiment of the present application; Figure 2 It is a structural schematic diagram of a safety protection test system for breakdown of a flexible DC converter valve arrester provided in an embodiment of the present application; Figure 3 It is a structural schematic diagram of a safety protection test device provided in an embodiment of the present application; Figure 4 is a circuit diagram of a safety protection test device provided in an embodiment of the present application; Figure 5 It is a schematic diagram of the test points of the safety protection test device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0018] In the description of the present application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, features defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, the meaning of "multiple" is two or more. In the description of the present application, the term "including" and its variations are open inclusions, i.e., "including but not limited to". The term "based on" means "at least partially based on". The term "according to" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments".

[0019] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" 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 it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0020] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0021] First, see Figure 1 , shows a flow chart of a safety protection test method for breakdown of a flexible DC converter valve arrester provided in an embodiment of the present application, the method comprising steps S101-S103, which are as follows: S101, based on the test data measured by the safety protection test device, data analysis and processing are performed through a data analysis model to obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of the flexible DC converter valve arrester.

[0022] In one example, the safety protection test method can be applied to the cloud, which can be deployed with a data analysis model and / or a test result analysis model (if the cloud does not have a corresponding model deployed, the cloud can call a device deployed with the corresponding model to implement the model call). At this time, the cloud can establish a communication connection with the safety protection test device to obtain the measured test data from the safety protection test device. In addition, the safety protection test device can also be configured to send the measured test data to the cloud in response to the measurement of the test data.

[0023] S102: Determine reference test information that matches the test prompt information.

[0024] In one example, step S102 may include: retrieving at least part of the reference test information in a historical test database according to the test prompt information, and / or retrieving at least part of the reference test information in a preset standard test database according to the test prompt information.

[0025] Among them, the standard test database may store standard test information, and the test corresponding to the standard test information may be a test completed by a standard safety protection test device in accordance with a standard test process and / or a standard test environment. In this way, at least part of the reference test information retrieved by this embodiment can be used to prompt the model to compare at least one of the following: the difference between the safety protection test device and the standard safety protection test device, the difference between the test process indicated by the measured test data and the standard test process, and the difference between the test environment indicated by the measured test data and the standard test environment.

[0026] The historical test database may store data of historical tests for simulating arrester breakdown. In this way, by using the data obtained from historical tests as reference test information, this embodiment can fully utilize the data obtained from each historical test in the past, and avoid the waste of computing power caused by retraining and updating the model every time the data obtained from historical tests is obtained.

[0027] S103, based on the reference test information and the test data, perform test result analysis and processing through a test result analysis model to obtain a first test result.

[0028] In one example, step S103 may include: directly using the reference test information as an analysis prompt of the test result analysis model, and performing test result analysis processing on the test data using the test result analysis model.

[0029] The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

[0030] It should be noted that "data analysis processing at least includes test result analysis processing" can be understood as that the data analysis model, after training, can also be used to implement the test result analysis processing performed by the test result analysis model, so that the output of the data analysis model can include corresponding test prompt information.

[0031] In one example, the sample data is suitable for indicating input-output data pairs of a test result analysis model.

[0032] In an optional implementation, the reference test information includes first reference test information and second reference test information, and determining the reference test information matching the test prompt information includes: Dividing the test prompt information into a plurality of prompt information blocks, and performing feature extraction on each prompt information block to obtain a corresponding first feature and a second feature, wherein the first feature and the second feature respectively represent different test information; Clustering the first features corresponding to the plurality of prompt information blocks to obtain at least one first feature cluster, and clustering the second features corresponding to the plurality of prompt information blocks to obtain at least one second feature cluster; Based on the at least one first feature cluster, first reference test information is determined, and based on the at least one second feature cluster, second reference test information is determined.

[0033] In one example, the test information represented by the first feature is different from the test information represented by the second feature. For example, the test information represented by the first feature may include a test process, and the test information represented by the second feature may include a test environment, etc.

[0034] In one example, clustering the first features corresponding to each of the plurality of prompt information blocks to obtain at least one first feature cluster, and clustering the second features corresponding to each of the plurality of prompt information blocks to obtain at least one second feature cluster, may include: determining the similarity between each first feature, and clustering the first features corresponding to each of the plurality of prompt information blocks based on the similarity between each first feature to obtain at least one first feature cluster; determining the similarity between each second feature, and clustering the second features corresponding to each of the plurality of prompt information blocks based on the similarity between each second feature to obtain at least one second feature cluster.

[0035] In one example, the features described in any one or more embodiments of the present application can be represented as feature vectors, and the similarity described in any one or more embodiments of the present application can be calculated by at least one of the following: cosine similarity, Euclidean distance, Manhattan distance, Pearson correlation coefficient, etc. It should be understood that this implementation is only an example and does not limit the present application.

[0036] In one example, determining the first reference test information based on the at least one first feature cluster, and determining the second reference test information based on the at least one second feature cluster, may include: inputting the at least one first feature cluster into a large language model to generate corresponding first reference test information; inputting the at least one second feature cluster into a large language model to generate corresponding second reference test information. In this embodiment, the powerful computing power and semantic understanding capabilities of the large language model can be used to generate reasonable reference test information that integrates the information in each feature cluster.

[0037] In an optional embodiment, the determining the first reference test information based on the at least one first feature cluster, and the determining the second reference test information based on the at least one second feature cluster, comprises: fusing each first feature in each first feature cluster to obtain a third feature corresponding to each first feature cluster, and fusing each second feature in each second feature cluster to obtain a fourth feature corresponding to each second feature cluster; In a preset first standard test database, information matching each of the third features is obtained as first standard test information of a corresponding first feature cluster, and in a preset second standard test database, information matching each of the fourth features is obtained as second standard test information of a corresponding second feature cluster; The first standard test information of each of the at least one first feature cluster is integrated to obtain first reference test information, and the second standard test information of each of the at least one second feature cluster is integrated to obtain second reference test information.

[0038] It can be understood that the first standard test database and / or the second standard test database in this embodiment can refer to the description of the above-mentioned standard test database, and the difference may be: taking the example that the test information represented by the first feature includes the test process, and the test information represented by the second feature includes the test environment, the test corresponding to the standard test information stored in the first standard test database can be a test completed by a standard safety protection test device according to the standard test process, and the test corresponding to the standard test information stored in the second standard test database can be a test completed by a standard safety protection test device according to the standard test environment.

[0039] It is not difficult to understand that each third feature in this embodiment integrates the information of each first feature in the corresponding first feature cluster, so that the third feature can characterize the characteristics of the corresponding first feature cluster, so that the first standard test information matched with the third feature obtained corresponds to the characteristics of the corresponding first feature cluster. In addition, the fourth feature is the same and will not be described here.

[0040] In an optional implementation, the extracting features of each prompt information block to obtain the corresponding first feature and second feature includes: Each prompt information block is encoded by an encoder to obtain a corresponding first feature and a second feature.

[0041] In an optional implementation, the sample data includes sample prompt information and sample test results, and the training process of the data analysis model includes: Obtain the data analysis model to be trained, as well as sample test data and its corresponding test result labels; Inputting the sample test data into the test result analysis model to obtain the sample prompt information and the sample test result output by the test result analysis model; Based on the test result label, the sample prompt information and the sample test result, the data analysis model to be trained is trained to obtain the data analysis model.

[0042] In one example, based on the test result label, the sample prompt information and the sample test result, the data analysis model to be trained is trained to obtain the data analysis model, which may include: based on the difference between the test result label and the sample test result, and the association between the sample prompt information and the difference, model training is performed on the data analysis model to be trained.

[0043] It should be understood that the model training can be completed using a general training method, such as the gradient descent method, etc., which is not strictly limited here.

[0044] In an optional embodiment, the test result analysis model includes a large model, and the test result analysis and processing is performed by the test result analysis model based on the reference test information and the test data to obtain the first test result, including: Based on the reference test information, determining model prompt information for the large model; The model prompt information and the test data are input into the large model, so that the large model, under the guidance of the model prompt information, performs reasoning according to the test data, thereby completing the test result analysis and processing to obtain the first test result.

[0045] In an example, the reference test information may also be directly used as the model prompt information for the large model.

[0046] In an optional embodiment, the method further comprises: Acquire test description information associated with the test data input by a user.

[0047] In one example, the test description information can be obtained by the user by inputting according to a pre-set test description template, for example, the user fills in the blanks in the test description template. In this way, the test description information has strict, clear semantics that conform to the test logic. At this time, there is no need to perform semantic understanding / context understanding on the test description information, and there is no need to perform test logic understanding on the test description information, so there is no need to rely on the understanding ability of the large model itself, avoiding the problem of inaccurate information cleaning caused by model understanding deviation.

[0048] The step of determining model prompt information for the large model based on the reference test information includes: Based on the test description information, the reference test information is cleaned to obtain the model prompt information.

[0049] In this embodiment, based on the test description information input by the user, the cleaning process can be used to clean out the information in the reference test information that is not sufficiently relevant to the current test data, so that only the information with a high correlation with the current test data is retained in the model prompt information, wherein the correlation can be obtained by calculating the similarity.

[0050] In one example, based on the test description information, the reference test information is cleaned to obtain the model prompt information, which may include: based on the format of the test description template, semantically understanding the test description information to obtain a semantic understanding result corresponding to the test description information, wherein the semantic understanding result is suitable for indicating the test logic corresponding to the test data; in the reference test information, determining information that has insufficient similarity with the semantic understanding result, and removing the information that has insufficient similarity in the reference test information to obtain the model prompt information.

[0051] In an optional embodiment, the method further comprises: Performing correlation detection on the test data and the first test result by using the data analysis model; Based on the detected correlation information, a second test result is determined.

[0052] In an optional implementation, determining the second test result based on the detected correlation information includes: In a case where the correlation information satisfies a preset correlation condition, taking the first test result as the second test result; In the case where the correlation information does not satisfy a preset correlation condition, feeding back the correlation information to the data analysis model, so that the data analysis model updates the test prompt information according to the correlation information; The updated test prompt information is used as new test prompt information to re-execute the step of determining reference test information matching the test prompt information until the re-obtained correlation information meets the preset correlation condition.

[0053] In an example, the preset correlation condition may include that the correlation similarity indicated by the correlation information is lower than a correlation threshold.

[0054] On the second aspect, accordingly, the embodiments of the present application also provide a safety protection test system for the breakdown of a flexible DC converter valve lightning arrester, which can implement all processes of the safety protection test method for the breakdown of a flexible DC converter valve lightning arrester provided in the above embodiments.

[0055] See also Figure 2 , shows a schematic structural diagram of a safety protection test system for a flexible DC converter valve lightning arrester breakdown provided in an embodiment of the present application, the safety protection test system for a flexible DC converter valve lightning arrester breakdown comprises: The data analysis module 201 is used to perform data analysis and processing through a data analysis model based on test data measured by a safety protection test device to obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of a flexible DC converter valve arrester; A reference test information determination module 202, used to determine reference test information matching the test prompt information; A first test result acquisition module 203 is used to perform test result analysis and processing based on the reference test information and the test data through a test result analysis model to obtain a first test result; The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

[0056] In an optional implementation, the reference test information includes first reference test information and second reference test information, and determining the reference test information matching the test prompt information includes: Dividing the test prompt information into a plurality of prompt information blocks, and performing feature extraction on each prompt information block to obtain a corresponding first feature and a second feature, wherein the first feature and the second feature respectively represent different test information; Clustering the first features corresponding to the plurality of prompt information blocks to obtain at least one first feature cluster, and clustering the second features corresponding to the plurality of prompt information blocks to obtain at least one second feature cluster; Based on the at least one first feature cluster, first reference test information is determined, and based on the at least one second feature cluster, second reference test information is determined.

[0057] In an optional embodiment, the determining the first reference test information based on the at least one first feature cluster, and the determining the second reference test information based on the at least one second feature cluster, comprises: fusing each first feature in each first feature cluster to obtain a third feature corresponding to each first feature cluster, and fusing each second feature in each second feature cluster to obtain a fourth feature corresponding to each second feature cluster; In a preset first standard test database, information matching each of the third features is obtained as first standard test information of a corresponding first feature cluster, and in a preset second standard test database, information matching each of the fourth features is obtained as second standard test information of a corresponding second feature cluster; The first standard test information of each of the at least one first feature cluster is integrated to obtain first reference test information, and the second standard test information of each of the at least one second feature cluster is integrated to obtain second reference test information.

[0058] In an optional implementation, the extracting features of each prompt information block to obtain the corresponding first feature and second feature includes: Each prompt information block is encoded by an encoder to obtain a corresponding first feature and a second feature.

[0059] In an optional implementation, the sample data includes sample prompt information and sample test results, and the training process of the data analysis model includes: Obtain the data analysis model to be trained, as well as sample test data and its corresponding test result labels; Inputting the sample test data into the test result analysis model to obtain the sample prompt information and the sample test result output by the test result analysis model; Based on the test result label, the sample prompt information and the sample test result, the data analysis model to be trained is trained to obtain the data analysis model.

[0060] In an optional embodiment, the test result analysis model includes a large model, and the test result analysis and processing is performed by the test result analysis model based on the reference test information and the test data to obtain the first test result, including: Based on the reference test information, determining model prompt information for the large model; The model prompt information and the test data are input into the large model, so that the large model, under the guidance of the model prompt information, performs reasoning according to the test data, thereby completing the test result analysis and processing to obtain the first test result.

[0061] In an optional embodiment, the system further includes: A test description information acquisition module, used to acquire test description information associated with the test data input by a user; The step of determining model prompt information for the large model based on the reference test information includes: Based on the test description information, the reference test information is cleaned to obtain the model prompt information.

[0062] In an optional embodiment, the system further includes: A correlation detection module, used to perform a correlation detection on the test data and the first test result through the data analysis model; The second test result acquisition module is used to determine the second test result based on the detected correlation information.

[0063] In an optional implementation, determining the second test result based on the detected correlation information includes: In a case where the correlation information satisfies a preset correlation condition, taking the first test result as the second test result; In the case where the correlation information does not satisfy a preset correlation condition, feeding back the correlation information to the data analysis model, so that the data analysis model updates the test prompt information according to the correlation information; The updated test prompt information is used as new test prompt information to re-execute the step of determining reference test information matching the test prompt information until the re-obtained correlation information meets the preset correlation condition.

[0064] In addition, among some existing technologies, flexible direct current transmission technology is an important means of large-scale new energy transmission. In recent years, the research, design and manufacturing technology of ultra-high voltage flexible direct current converter valves have been continuously developed and have reached the ±800kV 8GW level. The ±800kV 8GW ultra-high voltage flexible direct current transmission technology has a capacity comparable to conventional direct current. In terms of operation control, it has the characteristics of 4-quadrant flexible control and does not rely on the voltage support of the AC grid, while conventional direct current can only achieve 2-quadrant control and relies on the support of the AC grid.

[0065] The flexible DC converter valve is the core equipment of the flexible DC transmission system. The converter valve is composed of multiple sub-modules. Taking a project using ±800kV 8GW as an example, the project requires at least thousands of sub-modules.

[0066] The number of submodules and the selection of submodule capacitance values ​​will greatly affect the project cost. Configuring valve arresters can optimize the converter valve design, reduce the number of submodules, and reduce the project cost, but the arrester breakdown conditions and protection issues need to be considered.

[0067] The valve arrester can limit the overvoltage level of the converter valve during the energy absorption process to protect the converter valve. If the valve arrester flashes and breaks down during the energy absorption process, it is equivalent to short-circuiting the high and low voltage ends of the converter valve group, which will cause the large capacitor in the converter valve submodule to discharge uncontrolled. Under the most severe conditions, it will cause the converter valve to explode, causing equipment loss and will cause power outages in the transmission system.

[0068] In order to solve the problem that the arrester breakdown may cause damage to the flexible DC converter valve and further expand the fault range, it is usually necessary to design corresponding protection measures for the arrester breakdown problem to ensure the safety of the submodules that need to be protected after the arrester breakdown. However, since the test needs to be carried out directly with the combined transmission system and the arrester, there will be at least one of the following difficulties: 1. The voltage level of the transmission system is relatively high, and its voltage level is usually hundreds of kilovolts or even higher than that of the arrester. It is extremely difficult and dangerous to test the two large high-voltage equipment together. At this time, plan 1 can be adopted: disassemble the flexible direct current converter valve group and use one or more converter valve sub-modules as the basic test unit to simulate the transient breakdown test of the arrester.

[0069] 2. When conducting joint tests, it is necessary to consider the characteristics and mutual influence of the lightning arrester and the transmission system at the same time to ensure that the test conditions can truly simulate the actual operation conditions. This poses a high challenge to the design and technical requirements of the test plan. At this time, plan 2 can be adopted: through multiple rounds of electromagnetic-electromechanical simulation, comprehensively consider the boundary conditions of the test, conduct simulation tests under strict test conditions and verify the effectiveness of protective measures.

[0070] 3. The complex structure of the power transmission system and the arrester makes it difficult to install and disassemble the equipment under test, which not only increases the time and cost of the test, but may also cause damage to the arrester and the power transmission system. In this case, the above-mentioned solution 1 can be adopted.

[0071] 4. The requirements for the test site and test equipment are high. It is necessary to select a test hall with test conditions, and professional test equipment with high voltage tolerance is also required. At this time, the above solution 1 can be adopted.

[0072] However, the above schemes 1 and 2 have many difficulties in the safety protection verification experiment when simulating the accidental breakdown of the lightning arrester. Therefore, it is particularly important to study a safety protection test device for simulating the breakdown of the lightning arrester of the ultra-high voltage flexible DC converter valve. The safety protection test device is designed to evaluate whether the protection measures set after the lightning arrester breaks down in the ultra-high voltage flexible DC transmission system through simulation tests under laboratory conditions can effectively protect the flexible DC converter valve from being affected.

[0073] In view of this, see Figure 3 and Figure 4 The present application provides an example of a safety protection test device for simulating the breakdown of a lightning arrester of a UHV flexible DC converter valve. The device can be used to simulate the working condition of an accidental breakdown of the lightning arrester during the voltage rise of the valve group when the flexible DC valve is charged, so as to verify whether the protection measures set to cope with such working conditions can effectively protect the converter valve from damage. The safety protection test device includes: Test module 301; A pre-charging loop module 302, two ends of which are electrically connected to a first end and a second end of the tested module respectively; The arrester protection simulation module 303 comprises a first switch and a protection circuit connected in series, wherein two ends of the first switch and the protection circuit connected in series are electrically connected to a first end and a second end of the tested module respectively; A discharge loop module 304, two ends of which are electrically connected to the first end and the second end of the tested module respectively; and The controller 305 is respectively connected to the tested module, the pre-charging circuit module, the arrester protection simulation module, and the discharge circuit module, and is configured as follows: In response to a pre-charging instruction, the pre-charging circuit module is controlled to output current to the tested module, and when it is determined that the voltage of the tested module reaches a first preset voltage, the pre-charging circuit module is controlled to stop outputting current; When it is determined that the voltage of the tested module reaches the first preset voltage, in response to a discharge instruction, the discharge circuit module is controlled to output current to the tested module, and when it is determined that the voltage of the tested module reaches the second preset voltage, the first switch is controlled to be turned on so that at least part of the current output by the discharge circuit module flows into the protection circuit, wherein the second preset voltage is greater than the first preset voltage.

[0074] In this way, this embodiment can reduce the difficulty and time consumed in the safety protection test of lightning arrester breakdown, improve the safety of the test, and can control the safety protection test device through the controller to perform relevant tests, so that more accurate test data (such as circuit information in each module) can be measured more efficiently and conveniently during the test process to facilitate subsequent data processing.

[0075] The following is a further description of the safety protection test device.

[0076] In one example, the tested module 301 can be used to simulate equipment that needs protection.

[0077] In one example, the pre-charging circuit module 302 is suitable for providing a pre-charging current that makes the voltage of the tested module reach a first preset voltage; the discharging circuit module 304 is suitable for outputting a current to the tested module to increase the voltage of the tested module after the voltage of the tested module reaches the first preset voltage; when the voltage of the tested module reaches a second preset voltage, the first switch is turned on, so that at least part of the current output by the discharging circuit module flows into the protection circuit.

[0078] In one example, when the first switch is turned on, it can be used to simulate a lightning arrester breakdown.

[0079] In one example, the protection circuit can be used to simulate a protection device whose protection performance needs to be evaluated, and the protection performance is suitable for indicating the safety protection capability that the protection device can provide to the tested module 301 when the lightning arrester breaks down.

[0080] It is worth noting that the tested module 301 used in each embodiment of the present application may include a flexible DC converter valve in a flexible DC power transmission system, or various modules inside the flexible DC converter valve, wherein the flexible DC converter valve is used in conjunction with a lightning arrester.

[0081] In an optional embodiment, in combination Figure 3 and Figure 4 , the pre-charging circuit module comprises: A first resistor R3 and a second switch QS01 connected in series; and, The first direct current supply unit has one end electrically connected to the second end of the tested module via the first resistor R3 and the second switch QS01 connected in series, and the other end electrically connected to the first end of the tested module.

[0082] In one example, the first DC power supply unit includes a first AC power source AC1 and a first three-phase rectifier bridge circuit 401. The first three-phase rectifier bridge circuit 401 is suitable for converting AC power provided by the first AC power source AC1 into DC power output by the first DC power supply unit.

[0083] In an example, the second switch QS01 may be an isolating switch.

[0084] In an optional embodiment, in combination Figure 3 and Figure 4 , the discharge circuit module comprises: a second direct current supply unit; and The first capacitor C1 has a first end electrically connected to one end of the second DC power supply unit, a second end electrically connected to the other end of the second DC power supply unit, and a first end electrically connected to a first end of the tested module, and a second end electrically connected to a second end of the tested module after being connected in series with a thyristor, a reactor, and a switch.

[0085] In one example, the second DC power supply unit includes a second AC power source AC2 and a second three-phase rectifier bridge circuit 402. The second three-phase rectifier bridge circuit 402 is suitable for converting AC power provided by the second AC power source AC2 into DC power output by the second DC power supply unit.

[0086] In one example, the first capacitor C1 has a value of 14 mF.

[0087] In an optional embodiment, in combination Figure 3 and Figure 4 , the discharge circuit module also includes: A second resistor R2 and a third switch K13 are connected in series, wherein a first end of the first capacitor C1 is electrically connected to one end of the second DC power supply unit via the second resistor R2 and the third switch K13 connected in series.

[0088] In an example, the third switch K13 may be an isolating switch.

[0089] In this embodiment, the second resistor R2 can play a role of current limiting.

[0090] In an optional embodiment, in combination Figure 3 and Figure 4 , the discharge circuit module also includes: A reactance L1, a first thyristor SCR1 and a fourth switch K01 are connected in series, wherein the first end of the first capacitor C1 is electrically connected to the second end of the tested module via the reactance L1, the first thyristor SCR1 and the fourth switch K01.

[0091] In one example, the reactance L1 has a value of 2.5 mH.

[0092] In an example, the fourth switch K01 may be an isolating switch.

[0093] In an optional embodiment, in combination Figure 3 and Figure 4 , the discharge circuit module also includes: A third resistor R4, two ends of which are electrically connected to the first end and the second end of the first capacitor C1 respectively; and / or, At least one first diode D3, the anode of each of the first diodes D3 is electrically connected to the first end of the first capacitor C1, and the cathode of each of the first diodes D3 is electrically connected to the second end of the first capacitor C1.

[0094] In this embodiment, the third resistor R4 mainly plays the role of a voltage-equalizing resistor.

[0095] In an optional embodiment, in combination Figure 3 and Figure 4 , the protection circuit includes at least one inductor L2, and the first switch includes a second thyristor SCR2.

[0096] In one example, the value of the inductor L2 is 125 μH.

[0097] In an optional embodiment, in combination Figure 3 and Figure 4 , the tested module comprises: A second diode D1 and a third diode D2; a first IGBT (Insulate-Gate Bipolar Transistor) T1, whose collector is electrically connected to the cathode of the second diode D1, and whose emitter is electrically connected to the anode of the second diode D1; A second IGBT T2, whose collector is electrically connected to the cathode of the third diode D2, the emitter of the first IGBTT1, and the anode of the second diode D1, and whose emitter is electrically connected to the anode of the third diode D2, wherein the first end of the tested module is electrically connected between the emitter of the first IGBTT1 and the collector of the second IGBT T2; A fourth resistor R1 and a second capacitor C2 are connected in parallel, and one end of the fourth resistor R1 and the second capacitor C2 after being connected in parallel is electrically connected to the collector of the first IGBT T1 and the cathode of the second diode D1, respectively; the other end of the fourth resistor R1 and the second capacitor C2 after being connected in parallel is electrically connected to the anode of the third diode D2 and the emitter of the second IGBT T2, respectively; and one end of the fourth resistor R1 and the second capacitor C2 after being connected in parallel is electrically connected to the anode of the third diode D2 and the emitter of the second IGBT T2, and is used as the first end of the test module.

[0098] The beneficial effects of the above safety protection test device include at least one of the following: (1) This embodiment creatively uses a capacitor bank to charge the submodule under test, so as to simulate the situation in which a large current is injected into the submodule to increase its voltage under actual working conditions; (2) This embodiment uses the opening and closing of power electronic devices to simulate the process of energy transfer from the submodule to the arrester before and after the arrester breaks down; (3) The circuit of this embodiment is simple and easy to operate, and only very few test resources are needed to complete whether the safety protection measures of the converter valve are effective when the lightning arrester breaks down.

[0099] In one example, see Figure 4 , it can be considered that the receiving end fault has been locked, but due to communication delay, the sending end energy is still sending power to the receiving end. At this time, the DC at the sending end continues to increase, which may cause the arrester to break down. At this time, if the arrester breaks down, the diode will reversely recover, and the reverse recovery of the diode under high voltage and high current is easy to be damaged. In order to avoid this problem, a 25mH reactor is connected in series with the arrester (no direction in the figure, 25mH is used in actual engineering, and it is equivalent to 125uH, i.e. L2, for experimental verification below) to limit the di / dt value of the diode (i.e. D1) when the arrester breaks down transiently to prevent damage to the diode. In order to verify the feasibility of this safety protection scheme, an experiment is designed for verification.

[0100] Carry out the verification test as follows: (1) To prevent the submodule from being damaged due to accidental breakdown of the lightning arrester, a 25mH reactor is connected in series with the lightning arrester for protection. This solution is equivalent to a 125µH reactor connected in series with the submodule. (2) In accordance with Figure 4 Build a test circuit; (3) In accordance with Figure 5 Schematic diagram of the submodule test circuit layout. The layout of each test point I (indicating reactance current), V (used to simulate the voltage across the thyristor when the arrester breaks down), ID, VD, V dc ; (4) After the test circuit is built, check the internal hardware circuit and software system of the test platform and set up a high-voltage protection warning area; (5) Send a closing command to QS01 in the pre-charging circuit through the control system (i.e., the controller) to pre-charge the module under test. After the pre-charging voltage reaches a first preset voltage (e.g., 400V), charging is stopped. At this time, all boards of the module under test have been powered on. (6) The control system sends a closing command to K13 in the discharge circuit module to pre-charge C1 in the discharge circuit module. The charging stops when the pre-charge voltage reaches 6 kV. At this time, the energy stored in C1 is sufficient to charge the test module to more than 4 kV. (7) After the above charging is completed, disconnect K13 and QS01, and close K01; (8) After the above operation is completed, the first capacitor C1 enters the charging phase for the tested module, and the control system sends a turn-on command to SCR1. At this time, C1 will charge C2 of the tested module through L1 and D1, which simulates the process of the DC voltage of the UHV flexible DC converter valve rising continuously; (9) When the voltage of the tested module rises to the preset verification voltage (i.e., the second preset voltage) of 3900V, the thyristor (i.e., SCR2) that controls the breakdown of the lightning arrester is turned on, and the charging current for charging the tested module will be at least partially transferred to the thyristor that simulates the breakdown of the lightning arrester. At this time, it is equivalent to the actual breakdown of the lightning arrester, and the charging current is transferred from the tested module to the working condition of the lightning arrester; (10) After the above test steps, wait for the device to discharge, analyze the test waveform, check the status of the submodule, and check whether the electrical stress on the diode is within the safe range during the process of transferring the charging current from the submodule to the reactor when the arrester breaks down. This can help determine whether the protective measures configured with the reactor can effectively deal with the accidental breakdown of the arrester.

[0101] In summary, the embodiments of the present application have at least the following beneficial effects: According to an embodiment of the present application, test prompt information corresponding to the test data is obtained by performing data analysis and processing through a data analysis model based on test data measured by a safety protection test device, wherein the safety protection test device is used to simulate the breakdown of a flexible DC converter valve arrester; reference test information matching the test prompt information is determined; based on the reference test information and the test data, test result analysis and processing is performed through a test result analysis model to obtain a first test result; wherein the data analysis model is obtained through sample data training, and the sample data is data generated by the test result analysis model, and the data analysis and processing performed by the data analysis model at least includes the test result analysis and processing performed by the test result analysis model, and the test prompt information obtained through the data analysis model can be used as a basis to obtain reference test information, which is used as reference information to assist the test result analysis model in analyzing the test data, thereby improving the accuracy of the test result analysis.

[0102] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary hardware platform, and of course, it can also be implemented entirely by hardware. Based on such an understanding, all or part of the contribution of the technical solution of the present application to the background technology can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application or some parts of the embodiments.

[0103] The above is a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications are also considered to be within the scope of protection of the present application.

Claims

1. A safety protection test method for breakdown of a flexible DC converter valve arrester, characterized in that: include: Based on the test data measured by the safety protection test device, data analysis and processing are performed through a data analysis model to obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of the flexible DC converter valve arrester; Determining reference test information that matches the test prompt information; Based on the reference test information and the test data, performing test result analysis and processing through a test result analysis model to obtain a first test result; The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

2. The safety protection test method according to claim 1, characterized in that: The reference test information includes first reference test information and second reference test information, and the determining of the reference test information matching the test prompt information includes: Dividing the test prompt information into a plurality of prompt information blocks, and performing feature extraction on each prompt information block to obtain a corresponding first feature and a second feature, wherein the first feature and the second feature respectively represent different test information; Clustering the first features corresponding to the plurality of prompt information blocks to obtain at least one first feature cluster, and clustering the second features corresponding to the plurality of prompt information blocks to obtain at least one second feature cluster; Based on the at least one first feature cluster, first reference test information is determined, and based on the at least one second feature cluster, second reference test information is determined.

3. The safety protection test method according to claim 2, characterized in that: The determining of first reference test information based on the at least one first feature cluster, and the determining of second reference test information based on the at least one second feature cluster, include: fusing each first feature in each first feature cluster to obtain a third feature corresponding to each first feature cluster, and fusing each second feature in each second feature cluster to obtain a fourth feature corresponding to each second feature cluster; In a preset first standard test database, information matching each of the third features is obtained as first standard test information of a corresponding first feature cluster, and in a preset second standard test database, information matching each of the fourth features is obtained as second standard test information of a corresponding second feature cluster; The first standard test information of each of the at least one first feature cluster is integrated to obtain first reference test information, and the second standard test information of each of the at least one second feature cluster is integrated to obtain second reference test information.

4. The safety protection test method according to claim 2, characterized in that: The extracting features of each prompt information block to obtain the corresponding first feature and second feature includes: Each prompt information block is encoded by an encoder to obtain a corresponding first feature and a second feature.

5. The safety protection test method according to claim 1, characterized in that: The sample data includes sample prompt information and sample test results. The training process of the data analysis model includes: Obtain the data analysis model to be trained, as well as sample test data and its corresponding test result labels; Inputting the sample test data into the test result analysis model to obtain the sample prompt information and the sample test result output by the test result analysis model; Based on the test result label, the sample prompt information and the sample test result, the data analysis model to be trained is trained to obtain the data analysis model.

6. The safety protection test method according to claim 1, characterized in that: The test result analysis model includes a large model, and based on the reference test information and the test data, the test result analysis model is used to perform test result analysis processing to obtain a first test result, including: Based on the reference test information, determining model prompt information for the large model; The model prompt information and the test data are input into the large model, so that the large model, under the guidance of the model prompt information, performs reasoning according to the test data, thereby completing the test result analysis and processing to obtain the first test result.

7. The safety protection test method according to claim 6, characterized in that: The method further comprises: Acquiring test description information associated with the test data input by a user; The step of determining model prompt information for the large model based on the reference test information includes: Based on the test description information, the reference test information is cleaned to obtain the model prompt information.

8. The safety protection test method according to any one of claims 1 to 7, characterized in that: The method further comprises: Performing correlation detection on the test data and the first test result by using the data analysis model; Based on the detected correlation information, a second test result is determined.

9. The safety protection test method according to claim 8, characterized in that: The determining of the second test result based on the detected correlation information includes: In a case where the correlation information satisfies a preset correlation condition, taking the first test result as the second test result; In the case where the correlation information does not satisfy a preset correlation condition, feeding back the correlation information to the data analysis model, so that the data analysis model updates the test prompt information according to the correlation information; The updated test prompt information is used as new test prompt information to re-execute the step of determining reference test information matching the test prompt information until the re-obtained correlation information meets the preset correlation condition.

10. A safety protection test system for breakdown of flexible DC converter valve arrester, characterized in that: include: A data analysis module, configured to perform data analysis and processing through a data analysis model based on test data measured by a safety protection test device, and obtain test prompt information corresponding to the test data, wherein the safety protection test device is used to simulate the breakdown of a flexible DC converter valve arrester; A reference test information determination module, used to determine reference test information matching the test prompt information; A first test result acquisition module, configured to perform test result analysis and processing based on the reference test information and the test data through a test result analysis model to obtain a first test result; The data analysis model is obtained through training with sample data, the sample data is data generated by the test result analysis model, and the data analysis processing performed by the data analysis model at least includes the test result analysis processing performed by the test result analysis model.

Images