Test result coarse inspection system and method applied to integrated spacer device comprehensive test module

Abstract

The application discloses a test result rough detection system and method applied to an integrated test module of an integrated spacing device, the test result rough detection system runs in the circuit board corresponding to each detection item in the integrated test module of the integrated spacing device, and the architecture of the test result rough detection system is divided into four levels, including a hardware layer, a data layer, an application layer and a master control layer.In the test result rough detection method, the internal logical relationship of multiple indexes and the more refined feature meanings of the multiple indexes are mined by adopting a tree classification algorithm on the index data obtained through multiple detections, so that when new test data is input, the refined logical relationship and the feature meanings can be matched and compared with state indexes in an ideal case, the similarity relationship between the current test result and each ideal case result is obtained, and the final detection result is obtained.The application can greatly improve the automation and intelligent degree of a test instrument and reduce manual work.

Description

Technical Field

[0001] This invention relates to the field of electronic equipment system testing technology in the power industry, and particularly to a preliminary test result inspection system and method for integrated bay equipment comprehensive test modules. Background Technology

[0002] A comprehensive testing module for integrated bay equipment requires testing multiple items, each encompassing various indicators. The combination of these indicators enables the comprehensive testing of a transformer. However, due to the large number of test items, each test requires meticulous inspection and identification, necessitating significant manpower and resulting in high testing costs. This also places a heavy workload on the testing personnel. Summary of the Invention

[0003] The purpose of this invention is to provide a preliminary inspection system and method for test results applied to the comprehensive testing module of integrated bay equipment, thereby improving the accuracy and efficiency of test results during multi-index testing of integrated bay equipment and reducing the manpower costs required for testing. The technical solution adopted by this invention is as follows.

[0004] On one hand, this invention provides a preliminary inspection system for test results applied to a comprehensive testing module for integrated bay equipment. This preliminary inspection system operates on each testing circuit board corresponding to each testing item within the integrated bay equipment comprehensive testing module. The system includes a hardware layer, a data layer, an application layer, and a main control layer. The data layer, application layer, and main control layer are deployed on a main control circuit board. The hardware layer is configured in the test circuit of each testing circuit board and is used to collect and acquire specified index parameters and variables corresponding to the test circuit. The hardware layer outputs the collected index parameters and variables to the data layer. The data layer saves the data input from the hardware layer. The application layer processes the data saved by the data layer to obtain refined classification results. The main control layer manages the hardware layer, data layer, and application layer.

[0005] Optionally, the main control layer has a user interface through which users can register corresponding user information, detect task information, set the data content to be collected, set the internal communication method, and obtain the final measurement results.

[0006] Optionally, the hardware layer includes several data probes, each configured in the test circuit of each test circuit board, used to collect corresponding index parameters and variables. At the hardware level, the data probe is an ADC (Analog-to-Digital Converter), converting and acquiring the hardware data of the test circuit of each circuit board in the integrated interval device comprehensive test module; at the software level, it acts as a data bypass, collecting the main index parameters and variables in the software running the test circuit of each circuit board in the integrated interval device comprehensive test module.

[0007] Optionally, the data layer includes a data redundancy protection module, a persistent database, and a data communication module. The data redundancy protection module is connected to all data probe signals, receives indicator parameters and variable data output from all data probes, and performs multiple acquisitions and cross-validation on the indicator parameters and variable data to obtain accurate indicator parameters and variable data. The data redundancy protection module is communicatively connected to the persistent database, storing the cross-validated indicator parameters and variable data into the persistent database. The data redundancy protection module is also communicatively connected to the application layer, transmitting the cross-validated indicator parameters and variable data to the application layer. The data communication module is used for communication between the data layer and the hardware layer, application layer, and main control layer, as well as communication between the data redundancy protection module and the persistent database. The cross-validation mechanism ensures the absolute accuracy of the main indicator parameters and variables and the digital data collected by the data probes.

[0008] Optionally, the persistent database stores the indicator parameters and variable data after cross-validation by the data redundancy protection module in the form of data snapshots.

[0009] Optionally, the persistent database retrieves and saves the cross-validated index parameters and variable data output by the data redundancy protection module at fixed time intervals.

[0010] Optionally, the application layer includes a tree classification algorithm module, a matching algorithm module, and a detection algorithm module, wherein:

[0011] The tree classification algorithm module is configured to: during preprocessing, acquire the indicator parameters and variable data stored in the persistent database after cross-validation by the data redundancy protection module, calculate and obtain the ideal refined label tree classification model and the historical test result refined label tree classification model; and store the ideal refined label tree classification model and the historical test result refined label tree classification model in the persistent database for use in the testing phase; during the testing phase, the tree classification algorithm module is used to classify the collected indicator parameters and variable data after cross-validation by the data redundancy protection module using the historical test result refined label tree classification model to obtain the initial refined classification result;

[0012] The matching algorithm module is configured to: train the matching algorithm model using the refined label tree classification results under ideal conditions and the refined label tree classification results under historical test results during the preprocessing process; and during the testing phase, the matching algorithm module is used to perform matching operations between the initial refined classification results and the refined label tree classification results under various ideal conditions using the matching algorithm model to obtain a similarity value table between the initial refined classification results and the refined label tree classification results under various ideal conditions.

[0013] The detection algorithm module is configured to: sort the similarity value table, obtain the refined label tree classification result corresponding to the number with the largest value, and use the corresponding classification as the coarse detection result of the current measurement data.

[0014] Optionally, the tree classification algorithm module calculates and obtains the refined label tree classification model under ideal conditions and the refined label tree classification model based on historical test results, including:

[0015] Based on the stored historical test data and results, the XGBOOST algorithm is used to calculate a refined label tree classification model for the historical test results.

[0016] Based on the stored ideal test data and its corresponding ideal test results, the XGBOOST algorithm is used to calculate the refined label tree classification model under ideal conditions.

[0017] The ideal test data can be the calibration data of historical test data, and the ideal test result is the accurate test result corresponding to the historical test data.

[0018] Secondly, the present invention provides a method for preliminary inspection of test results applied to a comprehensive test module of an integrated bay device, comprising:

[0019] According to the preset correspondence between the test items, indicator parameters, and variable types, collect the indicator parameters and variable data corresponding to each test item;

[0020] For each testing item, based on the collected indicator parameters and variable data, a pre-built refined label tree classification model of historical test results is used to perform classification calculations and obtain the corresponding initial refined classification results;

[0021] Based on the initial refined classification results, a pre-trained matching algorithm model is used to perform matching operations between the initial refined classification results and the refined label tree classification results under various ideal conditions, thereby obtaining a similarity numerical table between the initial refined classification results and the refined label tree classification results under various ideal conditions.

[0022] The similarity value table is sorted, and the refined label tree classification result corresponding to the number with the largest value is obtained. The corresponding classification is used as the coarse detection result of the current measurement data.

[0023] The refined label tree classification model for historical test results is obtained by using the XGBOOST algorithm based on historical test data and historical test results corresponding to multiple categories.

[0024] The ideal refined label tree classification model is obtained by using the XGBOOST algorithm based on the ideal test data corresponding to multiple classifications and their corresponding ideal test results.

[0025] The matching algorithm model is obtained by training the model using the classification results of the refined label tree classification model under ideal conditions and the classification results of the refined label tree classification model based on historical test results.

[0026] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for preliminary inspection of test results applied to the integrated test module of an integrated interval device as described in the second aspect.

[0027] Beneficial effects

[0028] A comprehensive testing module for integrated bay equipment requires testing multiple items, each encompassing various indicators. The combination of these indicators completes the comprehensive testing of a transformer. This invention employs a tree classification algorithm to analyze multiple indicators during the testing process, uncovering their internal logical relationships and more refined feature meanings, as well as their impact on result classification. Therefore, when new test data is collected, it can match and compare it with ideal state indicators based on these refined logical relationships and feature meanings, obtaining the similarity between the current test result and the results of various ideal situations, thus obtaining the final test result. This invention significantly improves the automation and intelligence of the testing instrument, reducing manual labor. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the architecture of the preliminary inspection system for test results applied to the integrated interval equipment comprehensive test module of the present invention;

[0030] Figure 2 This is a schematic diagram of the classification tree construction process of the present invention;

[0031] Figure 3 This is a schematic diagram illustrating the implementation process of the method of the present invention. Detailed Implementation

[0032] The following description, in conjunction with the accompanying drawings and specific embodiments, provides further details.

[0033] Example 1

[0034] This embodiment describes a preliminary test result inspection system applied to the integrated testing module of an integrated bay device. (See also...) Figure 1 The preliminary inspection system for test results operates on the test circuit boards corresponding to each test item within the integrated interval equipment comprehensive test module. The hardware of this preliminary inspection system includes a main control circuit board. The system architecture is divided into four layers: hardware layer, data layer, application layer, and main control layer. The data layer, application layer, and main control layer are deployed on the main control circuit board.

[0035] The hardware layer consists of data probes within the system, primarily ADCs (Analog-to-Digital Converters) that convert hardware data from the circuitry into digital data and acquire it. At the software layer, it mainly functions as a data bypass, collecting key performance parameters and variables from the software algorithms running within the tester. The data collected by the probes is output to the data redundancy protection module.

[0036] The data layer mainly comprises three modules: a data redundancy protection module, a persistent database, and a data communication module. Its primary function is to ensure the normal and stable operation of data acquisition, while simultaneously storing historical data from past tests to support the scope of the current test.

[0037] Data Redundancy Protection Module: During data acquisition, noise exists in the hardware circuitry, and undiscovered bugs may exist in the software. Therefore, a data redundancy protection module is proposed. Multiple acquisitions are performed, and cross-validation is applied to ensure the absolute accuracy of the acquired data. The cross-validated data from the data redundancy protection module is: 1) stored in a persistent database; 2) output to the algorithm module.

[0038] Persistent Database: The persistent database stores historical data from all previously run tests. This data is saved as snapshots; at fixed time intervals, the cross-validated data output by the data redundancy protection module is retrieved and saved.

[0039] Data communication module: Ensures stable and normal data communication between various modules in the system.

[0040] The application layer mainly consists of a tree classification algorithm module, a matching algorithm module, and a detection algorithm module.

[0041] During preprocessing, the tree classification algorithm module retrieves historical data stored in a persistent database and calculates two types of tree classification models: a refined label tree classification model under ideal conditions and a refined label tree classification model based on historical test results. Both tree classification models are stored in the persistent database for use in the testing phase. During the formal calculation process, the tree classification algorithm module uses the refined label tree classification model based on historical test results to classify the collected data, obtaining the final refined classification result. .

[0042] During preprocessing, the matching algorithm module trains its model using both ideal, refined label tree classification results and historical test results. In the actual computation, the matching algorithm model applies these refined classification results. The results are matched with the refined label tree classification results under various ideal conditions to obtain a similarity table between the above results and the classification results under various ideal conditions.

[0043] The detection algorithm module sorts the above similarity value table and obtains the result corresponding to the largest number as the coarse detection result of this measurement data.

[0044] The main control layer consists primarily of the main control module. This module is mainly responsible for the user interface. Users register relevant user information, detect task information, set the data content to be collected, configure internal communication methods, etc., through this module, and obtain the final measurement results.

[0045] A specific embodiment of the test result preliminary inspection system of the present invention applied to the integrated test module of the integrated interval equipment is as follows:

[0046] See Figure 2First, a preprocessing process is performed. During preprocessing, the tree classification algorithm module retrieves historical data stored in a persistent database and calculates two types of tree classification models: a refined label tree classification model under ideal conditions and a refined label tree classification model based on historical test results. Both tree classification models are then stored in the persistent database for use in the testing phase. During preprocessing, the matching algorithm module trains its matching algorithm model using the refined label tree classification results from both the ideal and historical test results.

[0047] See Figure 3 Then, the formal measurement process begins. The user first registers relevant user information through the main control module, checks task information, sets the data content to be collected, configures internal communication methods, etc., and starts the test task. Data probes at the hardware layer acquire hardware and software measurement process index data, which is then passed to the data redundancy protection module to obtain the final collected index data. This data is stored in a persistent database as a training sample set for future tests, and also enters the application layer. During the formal calculation process, the tree classification algorithm module uses a refined label tree classification model based on historical test results to classify the collected data, obtaining the final refined classification result. The result is fed into the matching algorithm module, where it is matched against the refined label tree classification results for each ideal scenario to obtain a similarity table between the above result and the classification results for each ideal scenario. This table is then fed into the detection algorithm module, sorted, and the result corresponding to the largest value is obtained as the coarse detection result of this measurement data. Finally, the result is passed to the main control module and submitted to the user. Simultaneously, the detection result is also stored in a persistent database as training labels for subsequent tests, enriching the training sample set.

[0048] This algorithm combines the XGBOOST and LR algorithms to classify the test results of the transformer field rapid integrated tester and perform regression matching with each type of result to obtain the similarity value between the current test result and each type of result. Finally, by sorting the similarity values, the result that is closest to the current test result is obtained as the coarse inspection result of the current test result, so as to achieve the purpose of automated preliminary detection.

[0049] This algorithm learns from past test data using the XGBOOST algorithm to refine the classification of that data, resulting in a more detailed test classification tree label model α. Simultaneously, it learns from ideal test results using the XGBOOST algorithm to mine an ideal classification tree label model β, which is even more detailed than the ideal result data itself. After each test, the collected data is fed into the aforementioned test classification tree label model α to calculate the classification result for that test. This classification result is then compared with the ideal classification tree label β using the Logistic Regression (LR) algorithm to obtain the similarity score of the test result relative to various other results. .

[0050] Should The data structure is as follows:

[0051]

[0052] Regarding the above The data is sorted, and the one with the highest score is the result type of this test data.

[0053] Example 2

[0054] Based on the same inventive concept as the preliminary classification of test data in Example 1, this embodiment introduces a preliminary test result method applied to the integrated test module of an integrated bay device, including:

[0055] According to the preset correspondence between the test items, indicator parameters, and variable types, collect the indicator parameters and variable data corresponding to each test item;

[0056] For each testing item, based on the collected indicator parameters and variable data, a pre-built refined label tree classification model of historical test results is used to perform classification calculations and obtain the corresponding initial refined classification results;

[0057] Based on the initial refined classification results, a pre-trained matching algorithm model is used to perform matching operations between the initial refined classification results and the refined label tree classification results under various ideal conditions, thereby obtaining a similarity numerical table between the initial refined classification results and the refined label tree classification results under various ideal conditions.

[0058] The similarity value table is sorted, and the refined label tree classification result corresponding to the number with the largest value is obtained. The corresponding classification is used as the coarse detection result of the current measurement data.

[0059] refer to Figure 3As shown, the implementation of the method in this embodiment requires first analyzing historical indicator data before actual testing to uncover the internal logical relationships of multiple indicators in each testing project, as well as the more detailed feature meanings of multiple indicators and their impact on result classification. Thus, when new test data is collected, it is possible to match and compare it with the state indicators under ideal conditions based on the detailed logical relationships and feature meanings, obtain the similarity relationship between the current test results and the results of each ideal condition, and obtain the final test results.

[0060] The analysis of historical indicator data includes: for each detection item, based on the corresponding multi-class historical test data and historical test results, the XGBOOST algorithm is used to calculate the refined label tree classification model of the historical test results of the corresponding detection item;

[0061] For each testing item, based on the ideal test data for the corresponding multi-class classification and its corresponding ideal test results, the XGBOOST algorithm is used to calculate the refined label tree classification model under ideal conditions. The ideal test data can be calibration data of historical test data, and the ideal test results are the actual accurate test results corresponding to the known historical test data.

[0062] The matching algorithm model is trained using the classification results of the refined label tree classification model under ideal conditions and the classification results of the refined label tree classification model based on historical test results.

[0063] Example 3

[0064] This embodiment describes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for preliminary inspection of test results applied to the integrated test module of an integrated interval device as described in Embodiment 2.

[0065] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0066] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0067] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0068] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0069] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A preliminary inspection system for test results applied to a comprehensive test module of an integrated bay device, wherein the preliminary inspection system operates on each test circuit board corresponding to each test item in the comprehensive test module of the integrated bay device, characterized in that: The preliminary test result detection system includes a hardware layer, a data layer, an application layer, and a main control layer. The data layer, application layer, and main control layer are deployed on a main control circuit board. The hardware layer is equipped with several data probes, each of which is configured in the test circuit of each test circuit board to collect and acquire specified index parameters and variables corresponding to the test circuit. The hardware layer outputs the collected index parameters and variables to the data layer. The data layer stores the data input from the hardware layer. The data layer includes a data redundancy protection module, a persistent database, and a data communication module. The data redundancy protection module is connected to all data probe signals, receives indicator parameters and variable data output from all data probes, and performs multiple samplings and cross-validation on the indicator parameters and variable data to obtain accurate indicator parameters and variable data. The data redundancy protection module is communicatively connected to the persistent database, storing the cross-validated indicator parameters and variable data into the persistent database. The data redundancy protection module is also communicatively connected to the application layer, transmitting the cross-validated indicator parameters and variable data to the application layer. The data communication module is used for communication between the data layer and the hardware layer, application layer, and main control layer, as well as communication between the data redundancy protection module and the persistent database. The application layer processes the data stored in the data layer to obtain refined classification results. The application layer includes a tree classification algorithm module, a matching algorithm module, and a detection algorithm module. Specifically, the tree classification algorithm module is configured to: during preprocessing, acquire indicator parameters and variable data from the persistent database after cross-validation by the data redundancy protection module; calculate and obtain an ideal refined label tree classification model and a historical test result refined label tree classification model; and store these models in the persistent database for use during the testing phase. During the testing phase, the tree classification algorithm module is used to classify the indicator parameters and variable data after cross-validation by the data redundancy protection module using the historical test result refined label tree classification model to obtain... The initial refined classification result is obtained. The matching algorithm module is configured to: train the matching algorithm model using the refined label tree classification result under ideal conditions and the refined label tree classification result under historical test results during the preprocessing process; during the testing phase, the matching algorithm module is used to perform matching operations between the initial refined classification result and the refined label tree classification result under each ideal condition using the matching algorithm model to obtain a similarity value table between the initial refined classification result and the refined label tree classification result under each ideal condition; the detection algorithm module is configured to: sort the similarity value table, obtain the refined label tree classification result corresponding to the largest number, and use the corresponding classification as the coarse detection result of the current measurement data; The main control layer manages the hardware layer, data layer, and application layer.

2. The preliminary inspection system for test results applied to the integrated testing module of an integrated bay device according to claim 1, characterized in that: The main control layer has a user interface through which users can register relevant user information, detect task information, set the data content to be collected, set the internal communication method, and obtain the final measurement results.

3. The preliminary inspection system for test results applied to the integrated testing module of an integrated bay device according to claim 1, characterized in that: The persistent database stores the indicator parameters and variable data, which have been cross-validated by the data redundancy protection module, in the form of data snapshots.

4. The preliminary inspection system for test results applied to the integrated testing module of an integrated bay device according to claim 1, characterized in that: The persistent database retrieves and saves the cross-validated index parameters and variable data output by the data redundancy protection module at fixed time intervals.

5. The preliminary inspection system for test results applied to the integrated testing module of an integrated bay device according to claim 1, characterized in that: The tree classification algorithm module calculates and obtains both the ideal refined label tree classification model and the historical test result refined label tree classification model, including: Based on the stored historical test data and results, the XGBOOST algorithm is used to calculate a refined label tree classification model for the historical test results. Based on the stored ideal test data and its corresponding ideal test results, the XGBOOST algorithm is used to calculate the refined label tree classification model under ideal conditions.

6. A method for preliminary inspection of test results based on the preliminary inspection system for comprehensive test results of integrated interval equipment as described in any one of claims 1-5, characterized in that, include: According to the preset correspondence between the test items, indicator parameters, and variable types, collect the indicator parameters and variable data corresponding to each test item; For each testing item, based on the collected indicator parameters and variable data, a pre-built refined label tree classification model of historical test results is used to perform classification calculations and obtain the corresponding initial refined classification results; Based on the initial refined classification results, a pre-trained matching algorithm model is used to perform matching operations between the initial refined classification results and the refined label tree classification results under various ideal conditions, thereby obtaining a similarity numerical table between the initial refined classification results and the refined label tree classification results under various ideal conditions. The similarity value table is sorted, and the refined label tree classification result corresponding to the number with the largest value is obtained. The corresponding classification is used as the coarse detection result of the current measurement data. The refined label tree classification model for historical test results is obtained by using the XGBOOST algorithm based on historical test data and historical test results corresponding to multiple categories. The ideal refined label tree classification model is obtained by using the XGBOOST algorithm based on the ideal test data corresponding to multiple classifications and their corresponding ideal test results. The matching algorithm model is obtained by training the model using the classification results of the refined label tree classification model under ideal conditions and the classification results of the refined label tree classification model based on historical test results.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the coarse detection method for test results as described in claim 6.

Images