A method and system for fast self-learning of a cable network on-grid topology

Abstract

The present application relates to the technical field of cable automatic testing, in particular to a kind of cable network conduction topology fast self-learning method and system, by the incremental numbering of test node, positive candidate node is selected in order, if its number already exists in connectivity data table, then skip, otherwise as positive access measurement loop, and all nodes greater in number are connected in parallel as negative candidate set, excitation current is applied to measure total loop current;If total current is not greater than threshold, then prune, otherwise if candidate set is only one node, then record conduction, otherwise binary recursive detection, further negative candidate set is pre-divided into multiple subsets, in turn bundled detection, reduce invalid detection, while using the transitivity of conduction relationship, automatically skip covered node, reduce positive traversal times to the number of connected components, the present application can automatically discover unknown cable topology without pre-setting wiring table, effectively reduce the time complexity of detection, and the detection efficiency of large-scale complex cable network is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of automatic cable testing technology, and specifically to a rapid self-learning method and system for cable network conduction topology. Background Technology

[0002] As the "neural network" of modern complex electronic systems, the reliability of the cable network's conductivity directly affects the system's operational safety. With the continuous expansion of the cable network, such as the more than 20,000 test points for launch vehicle cables, the traditional point-to-point successive scanning method can no longer meet the needs of rapid and automated testing. The existing cable continuity testing technologies mainly include the following two types: (1) Point-to-point scanning method based on pre-set wiring tables: For example, the "Design of Portable Cable Continuity Tester" (China Air-to-Air Missile Research Institute, 2019) uses a "learning mode" to pre-store the cable wiring relationships and compares them one by one during testing. However, this method relies on manual input of the wiring table and cannot be used when the cable model under test changes or the connection relationship is unknown, resulting in poor versatility; (2) Full interconnection scanning method: By sequentially selecting any two points for resistance measurement, the time complexity is For example, for The cable network requires testing approximately 520,000 point pairs, but the testing time is too long and cannot meet the needs of rapid on-site testing.

[0003] Therefore, there is an urgent need for a method that can automatically and efficiently complete the learning of the conduction topology of large-scale cable networks. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention aims to provide a rapid self-learning method for cable network continuity topology, solving the problems of reliance on manual wiring tables, low testing efficiency, and numerous invalid measurements in existing cable continuity testing.

[0005] The basic solution provided by this invention is a fast self-learning method for cable network conduction topology, comprising the following steps: Step S1, Initialization: Number all test nodes of the cable network under test in ascending order of numerical values, establish a connectivity data table for storing conduction relationships and initialize it to empty, and preset the conduction current threshold. Step S2, Positive Electrode Traversal and Recursive Bundling Detection: Each test node is selected sequentially as a positive electrode candidate test node in ascending order of its number. For each positive electrode candidate test node, the test node number recorded in the current connectivity data table is obtained. If the current positive electrode candidate test node's number exists in the obtained test node numbers, that positive electrode candidate test node is skipped, and the next positive electrode candidate test node is selected. Otherwise, the current positive electrode candidate test node is connected to the measurement loop as the current positive electrode test node. The current negative electrode candidate set is initialized to all test nodes with numbers greater than the current positive electrode test node, and bundling detection is performed. The bundling detection includes: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] All test nodes in the set are simultaneously connected in parallel to the measurement loop. An excitation current is applied and the total loop current is measured. If the measured total loop current is not greater than the conduction current threshold, the processing of the current negative electrode candidate set ends. If the measured total loop current is greater than the conduction current threshold, the number of test nodes in the current negative electrode candidate set is determined: if it contains only one test node, it is confirmed that the test node is connected to the current positive electrode test node, the connection relationship is recorded and stored in the connectivity data table; if it contains multiple test nodes, the current negative electrode candidate set is divided into two subsets by a binary search method, each subset is used as a new current negative electrode candidate set, and the bundled detection is repeated. Step S3, Output Results: Output the connectivity data table to characterize the conduction topology of the cable network.

[0006] The principle of this basic scheme is as follows: By arranging test nodes in ascending order of their numbers and performing parallel measurements, multiple negative pole nodes are simultaneously determined to have a conductive path with the current positive pole node in a single measurement. If no conductive path exists, a large number of test node pairs are skipped. If a conductive path exists and only one node remains in the negative pole candidate set, the conductive relationship is recorded. If a conductive path exists and the negative pole candidate set has more than one test node, the candidate set is continuously divided into smaller subsets using a binary search method, and bundled detection is recursively performed, thereby locating all conductive nodes in logarithmic order. Furthermore, since the conductive relationships in the cable network are transitive, when traversing and selecting the current positive pole test node, this method first retrieves all recorded test node numbers from the connectivity data table. If the current positive pole candidate node number has already been recorded, it means that all possible conductive relationships have been completely covered in the detection of the previous smaller-numbered positive pole nodes, so the node is skipped directly, and the next node is moved on. This reduces the number of positive pole traversals from... The number of actual conductive components is reduced, where is the total number of test nodes in the cable network to be tested. This method reduces the time complexity of traditional point-by-point scanning. Reduce to average This significantly improves the self-learning speed of large-scale cable networks.

[0007] The beneficial effects of this basic scheme are as follows: 1. Blind testing and self-learning capabilities: Existing technologies either require manual input of wiring tables in advance (such as "learning mode") or use full interconnection point-by-point scanning, neither of which can automatically and efficiently complete testing under unknown topologies. This invention directly discovers all continuity relationships during the testing process through bundled detection and binary recursion, without any prior information, thus improving the versatility and emergency response capabilities of the testing system.

[0008] 2. Reduce time complexity: Traditional point-by-point scanning requires testing all node pairs, resulting in a time complexity of O(n log n). The present invention can eliminate some non-conductive intervals with a single bundle detection. Combined with binary recursion, the average number of detections is much less than that of traditional point-by-point scanning. In actual tests, for a cable network with 1024 test nodes and sparse connections, traditional point-by-point scanning requires about 520,000 scans, while the algorithm of the present invention only requires about 10,000 bundle detections, which is about 50 times more efficient. This order of magnitude improvement is a fundamental reduction in algorithm complexity and has practical significance for large-scale cable testing.

[0009] 3. Scalability and versatility: This method is applicable to various cable network tests and is decoupled from specific hardware platforms. Any test equipment with multi-channel parallel measurement capabilities can be implemented through software. The cable network topology obtained by this method can be saved as a configuration file, which facilitates the direct loading of cables of the same type in the future and makes it easier to accurately measure the conduction resistance of subsequent circuits.

[0010] Furthermore, in step S2, for each current positive electrode test node, before performing the bundled detection, the following operations are performed: the current negative electrode candidate set is divided into multiple consecutive negative electrode candidate subsets in numerical order; then, each negative electrode candidate subset is used as a new current negative electrode candidate set in sequence, and the bundled detection is performed.

[0011] By pre-partitioning, the large-scale negative pole candidate set is decomposed into multiple small subsets, which are then bundled and detected separately. Only subsets whose total loop current exceeds the threshold need to undergo further binary recursive detection, while a large number of non-conductive subsets are eliminated at once. This avoids repeated probing of non-conductive intervals in the middle by global binary search, further reducing the number of detections and improving detection efficiency. This is particularly significant for large-scale cables with only sparse continuity in practical applications. For example, when the continuity nodes only exist at both ends of the negative pole candidate set, the basic scheme still needs to perform a deep binary search on the entire interval. However, this scheme significantly reduces the number of detections by directly pruning other subsets and only performing recursive bundle detection on the subsets containing these two nodes.

[0012] Furthermore, the number of negative electrode candidate subsets is determined based on the total number of test nodes in the current negative electrode candidate set and the preset subset baseline capacity.

[0013] The number of negative pole candidate subsets The calculation method is as follows: in, This represents the total number of test nodes in the current negative pole candidate set. The baseline capacity of the subset is... This is the floor function.

[0014] Based on baseline capacity Control the maximum size of each negative candidate subset and the number of negative candidate subsets. Compared with the total number of test nodes in the current negative pole candidate set Positive correlation means that the negative candidate subset is adaptively adjusted. The baseline capacity can be preset according to the actual hardware detection capability and the expected detection granularity (such as a typical value of 16~64). This ensures that the binary recursive localization within each negative candidate subset will not fall into too deep a recursion depth due to the set being too large, and also avoids the frequent number of bundled detections caused by too many negative candidate subsets.

[0015] Furthermore, the connectivity data table uses a linked list structure to store a list of test nodes connected to each test node, with the test nodes in each list arranged in ascending order by number.

[0016] In this scheme, each node maintains a variable-length list that only stores the actual connected nodes. Compared with a fixed-size adjacency matrix, this saves storage space and eliminates the need to pre-allocate large blocks of contiguous memory, making it particularly suitable for online testing and embedded systems with a large number of test nodes. At the same time, the linked list structure combined with ascending order allows for efficient sequential access, facilitating manual review and program parsing. Furthermore, the ascending order list is beneficial for operations such as building adjacency lists, generating reports, and drawing cable topology diagrams.

[0017] The present invention also discloses a fast self-learning system for cable network conduction topology, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. Attached Figure Description

[0018] Figure 1 This is a flowchart of Embodiment 1 of a fast self-learning method for cable network conduction topology according to the present invention.

[0019] Figure 2 This is a schematic diagram of the test node binding detection process numbered 1 in Embodiment 1 of the present invention.

[0020] Figure 3This is a flowchart of Embodiment 2 of a fast self-learning method for cable network conduction topology according to the present invention. Detailed Implementation

[0021] The following detailed description illustrates the specific implementation method: Example 1 is basically as shown in the appendix. Figure 1 As shown: Taking a small-scale cable network with 16 test nodes as an example, the cable network under test has 16 test nodes, numbered sequentially from 1 to 16. Assume the continuity relationships are: (1,5), (1,8), (2,6), (3,7), (5,8). The excitation current of the measurement circuit is provided by a constant voltage applied to the measurement circuit. The conduction current threshold is 10mA. If there is no continuity relationship in the current detection, the entire measurement circuit cannot form a complete closure, and the total measurement circuit current is approximately 0, which is less than the conduction current threshold. The specific implementation process is as follows: Step S1, Initialization: Number nodes 1 to 16, and create an empty connectivity data table to store the conduction relationship. The connectivity data table uses a linked list structure to store the list of test nodes connected to each test node. The test nodes in each list are arranged in ascending order by number. The preset current threshold is 10mA.

[0022] Step S2, Positive Electrode Traversal and Recursive Bundling Detection: Each test node is selected sequentially as a positive electrode candidate test node in ascending order of its number. For each positive electrode candidate test node, the test node number recorded in the current connectivity data table is obtained. If the current positive electrode candidate test node's number exists in the obtained test node numbers, that positive electrode candidate test node is skipped, and the next positive electrode candidate test node is selected. Otherwise, the current positive electrode candidate test node is connected to the measurement loop as the current positive electrode test node. The current negative electrode candidate set is initialized to all test nodes with numbers greater than the current positive electrode test node, and bundling detection is performed. The bundling detection includes: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] All test nodes in the set are simultaneously connected in parallel to the measurement loop. An excitation current is applied and the total loop current is measured. If the measured total loop current is not greater than the conduction current threshold, the processing of the current negative electrode candidate set ends. If the measured total loop current is greater than the conduction current threshold, the number of test nodes in the current negative electrode candidate set is determined: if it contains only one test node, it is confirmed that the test node is connected to the current positive electrode test node, the connection relationship is recorded and stored in the connectivity data table; if it contains multiple test nodes, the current negative electrode candidate set is divided into two subsets by a binary search method, each subset is used as a new current negative electrode candidate set, and the bundled detection is repeated. In this embodiment 1, initially, the positive electrode candidate test node is numbered 1, as shown in the attached figure. Figure 2As shown, the connectivity data table is empty at this time, and there is no test node numbered 1. Therefore, the positive candidate test node numbered 1 is connected to the measurement loop as the current positive test node. The current negative candidate set is initialized to [2,3,...,16] and bundled detection is performed, that is, all test nodes in the current negative candidate set are connected in parallel as the current negative test nodes. The measured total loop current is greater than the conduction current threshold, which indicates that there is a conduction relationship. At this time, it is determined that the current negative candidate set contains multiple test nodes, so it is divided into two subsets by a binary search method. When dividing, the two subsets are divided into two equal subsets according to the number order. The number of test nodes in the two subsets does not differ by more than 1. In this embodiment, if a set containing an odd number of test nodes is divided into two subsets, the first subset has 1 more test nodes, so it is divided into a first subset [2,...,9] and a second subset [10,...,16]. The first subset [2,...,9] and the second subset [10,...,16] are respectively used as the new current negative candidate sets, and bundled detection is performed: For the first subset [2,...,9], after connecting the test nodes in parallel to the measurement loop, the total measured loop current is greater than the conduction current threshold, indicating a conduction relationship. Further bisection is performed to divide it into [2,...,5] and [6,...,9]. When [2,...,5] is connected, the total measured loop current is greater than the conduction current threshold, indicating a conduction relationship. Further bisection is performed to divide it into [2,3] and [4,5]. When [2,3] is connected, the total measured loop current is not greater than the conduction current threshold, indicating no conduction. The process ends here. The current negative electrode candidate set is processed as follows: when [4,5] is connected, the measured total loop current is greater than the conduction current threshold, indicating a conduction relationship. It is then divided into [4] and [5]. When [4] is connected, the measured total loop current is not greater than the conduction current threshold, indicating no conduction. The processing of the current negative electrode candidate set ends. When [5] is connected, the measured total loop current is greater than the conduction current threshold, indicating a conduction relationship. The current negative electrode candidate set (i.e., [5]) contains only one test node. The conduction relationship (1,5) is recorded and stored in the connectivity data table. For the subsequent subset [10,...,16], the same process is followed to confirm that the test node numbered 8 is connected to the test node numbered 1, and the connection relationship (1,8) is recorded in the connectivity data table. Finally, for the test node numbered 1, the linked list information in its connectivity data table is (1,5,8). After the above process, the processing of the current positive test node numbered 1 ends. The test node numbered 2 is then selected as the positive candidate test node, and the above process is performed, recording the continuity relationship (2,6) and storing it in the connectivity data table. For the positive candidate test node numbered 3, the above process is performed, recording the continuity relationship (3,7) and storing it in the connectivity data table. For the positive candidate test node numbered 4, after it is selected as the current positive test node, the current negative candidate set is initialized to [5,...,16]. At this point, after parallel connection to the measurement circuit, the total measurement circuit is not greater than the conduction current threshold, i.e., there is no continuity. Therefore, it is directly pruned, and the processing of the current negative candidate set ends. Processing: For the positive candidate test node numbered 5, obtain the test node number in the current connectivity data table. Since test node number 5 already exists, skip this positive candidate test node and continue to select the next test node as the positive candidate test node. Similarly, for test nodes numbered 6, 7, and 8, since their test node numbers exist in the connectivity data table, skip their selection. For the subsequent positive candidate test nodes numbered 9 to 15, since there is no conductivity during the first bundled detection, prune them directly. When selecting test node numbered 16 as the positive candidate test node, since this number exists in the connectivity data table, skip selecting this test node as the current positive test node.

[0023] Step S3, Output Results: Output the connectivity data table. The linked list information is (1,5,8), (2,6), (3,7), which is used to characterize the conduction topology of the cable network. The conduction relationships can be read from it as (1,5), (1,8), (2,6), (3,7), (5,8).

[0024] In this embodiment, the detection is performed using this method, requiring fewer than 120 detections compared to the traditional point-by-point scanning.

[0025] Example 2 is basically as shown in the appendix. Figure 3 As shown: Taking a large-scale cable network with 1024 test nodes as an example, it is assumed that the continuity relationship is extremely sparse: only the continuity relationships (1,2), (1,1024), and (2,1024) exist: Step S1: Initialization: Number nodes 1 to 1024 and create an empty connectivity data table to store the conduction relationships. The connectivity data table can use a linked list structure to store the list of test nodes connected to each test node. The test nodes in each list are arranged in ascending order by number. The preset current threshold is 10mA.

[0026] Step S2, Positive Electrode Traversal and Bundled Recursive Detection: Each test node is selected sequentially as a positive electrode candidate test node in ascending order of its number. For each positive electrode candidate test node, the test node number recorded in the current connectivity data table is obtained. If the current positive electrode candidate test node's number exists in the obtained test node numbers, that positive electrode candidate test node is skipped, and the next positive electrode candidate test node is selected. Otherwise, the current positive electrode candidate test node is connected to the measurement loop as the current positive electrode test node. The current negative electrode candidate set is initialized with all test nodes whose numbers are greater than the current positive electrode test node. The current negative electrode candidate set is divided into multiple consecutive negative electrode candidate subsets according to their numbering order. Then, each negative electrode candidate subset is sequentially used as a new current negative electrode candidate set. Then, the bundled detection is performed. The bundled detection includes: simultaneously connecting all test nodes in the current negative electrode candidate set to the measurement loop in parallel, applying an excitation current and measuring the total loop current. If the measured total loop current is not greater than the conduction current threshold, the processing of the current negative electrode candidate set ends. If the measured total loop current is greater than the conduction current threshold, the number of test nodes in the current negative electrode candidate set is determined: if only one test node is included, it is confirmed that the test node is connected to the current positive electrode test node, the connection relationship is recorded and stored in the connectivity data table; if multiple test nodes are included, the current negative electrode candidate set is divided into two subsets by a binary search method, each subset is used as a new current negative electrode candidate set, and the bundled detection is repeated. In this embodiment 2, the subset baseline capacity The default value is 256. For the positive candidate test node numbered 1, it is connected to the measurement loop as the current positive test node. At this time, the current negative candidate set is [2,3,...,1024], which is the total number of test nodes in the current negative candidate set. It is 1024, according to the formula The current negative pole candidate set is divided into 4 consecutive negative pole candidate subsets. In this embodiment 2, if it is not possible to divide them completely equally, the last negative pole candidate subset will have 1 fewer test node, that is, it will be divided into [2,...,257], [258,...,513], [514,...,769], [770,...,1024]. These four negative pole candidate subsets are used as the new current negative pole candidate set, and the bundled detection is performed. For the negative electrode candidate subset [2,...,257], after parallel connection, the measured total loop current is greater than the conduction current threshold, indicating a conduction relationship. A binary recursive positioning is performed within this negative electrode candidate subset to confirm that test node number 1 and test node number 2 are connected, and the conduction relationship (1,2) is recorded in the connectivity data table. For the negative electrode candidate subsets [258,...,513] and [514,...,769], the measured total loop current is not greater than the conduction current threshold, indicating no conduction relationship, and they are directly pruned. For the negative electrode candidate subset [770,...,1024], after parallel connection, the measured total loop current... If the current exceeds the conduction current threshold, a conduction relationship exists. A binary recursive search is performed within this negative electrode candidate subset to confirm that test node 1 and test node 1024 are connected. The conduction relationship (1, 1024) is recorded in the connectivity data table. That is, for test node 1, its linked list information in the connectivity data table is (1, 2, 1024). Therefore, test node 2 is skipped as the current positive electrode test node. When test nodes 3 to 1023 are subsequently selected as positive electrode test nodes, no conduction is detected after binding. For test node 1024, it exists in the connectivity data table and is skipped.

[0027] Step S3, Output Results: Output the connectivity data table, the linked list information is (1,2,1024), which is used to characterize the conduction topology of the cable network. The conduction relationship can be read from it as (1,2), (1,1024), (2,1024).

[0028] Example 2 uses pre-partitioning to decompose a large set of negative electrode candidates into multiple small subsets, which are then bundled and detected separately. Only subsets with continuity relationships need to be further recursively detected by binary search, while a large number of non-conductive subsets are eliminated at once. This avoids repeated probing of non-conductive intervals by global binary search, and has better adaptability in actual large-scale sparse cable networks.

[0029] After completing the self-learning process described above, the obtained connectivity data table can be saved as a configuration file (e.g., JSON format). When performing batch testing on the same type of cable, this configuration file can be directly loaded to perform precise four-wire resistance measurements (applying current, acquiring voltage, and calculating resistance values) on each conducting node pair, quickly obtaining milliohm-level precision on-resistance values ​​and generating a test report.

[0030] The present invention also provides a rapid self-learning system for cable network conduction topology, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above method. The test equipment used to implement Embodiment 2 can adopt a multi-board parallel architecture (8 measurement boards, each board managing one or more negative pole candidate subsets), and the main controller coordinates through a backplane bus.

[0031] The above are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A fast self-learning method for a cable network commissioning topology, characterized in that, Includes the following steps: Step S1, Initialization: Number all test nodes of the cable network under test in ascending order of numerical values, establish a connectivity data table for storing conduction relationships and initialize it to empty, and preset the conduction current threshold. Step S2, Positive Electrode Traversal and Recursive Bundling Detection: Each test node is selected sequentially as a positive electrode candidate test node in ascending order of its number. For each positive electrode candidate test node, the test node number recorded in the current connectivity data table is obtained. If the current positive electrode candidate test node's number exists in the obtained test node numbers, that positive electrode candidate test node is skipped, and the next positive electrode candidate test node is selected. Otherwise, the current positive electrode candidate test node is connected to the measurement loop as the current positive electrode test node. The current negative electrode candidate set is initialized to all test nodes with numbers greater than the current positive electrode test node, and bundling detection is performed. The bundling detection includes: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] All test nodes in the set are simultaneously connected in parallel to the measurement loop. An excitation current is applied and the total loop current is measured. If the measured total loop current is not greater than the conduction current threshold, the processing of the current negative electrode candidate set ends. If the measured total loop current is greater than the conduction current threshold, the number of test nodes in the current negative electrode candidate set is determined: if it contains only one test node, it is confirmed that the test node is connected to the current positive electrode test node, the connection relationship is recorded and stored in the connectivity data table; if it contains multiple test nodes, the current negative electrode candidate set is divided into two subsets by a binary search method, each subset is used as a new current negative electrode candidate set, and the bundled detection is repeated. Step S3, Output Results: Output the connectivity data table to characterize the conduction topology of the cable network.

2. A method of fast self-learning of a cable network commissioning topology according to claim 1, characterized in that, In step S2, for each current positive electrode test node, before performing the bundled detection, the following operations are performed: the current negative electrode candidate set is divided into multiple consecutive negative electrode candidate subsets in numerical order; then, each negative electrode candidate subset is used as a new current negative electrode candidate set in sequence, and the bundled detection is performed.

3. A method of fast self-learning of a cable network commissioning topology according to claim 2, characterized in that, The number of negative electrode candidate subsets is determined based on the total number of test nodes in the current negative electrode candidate set and the preset subset baseline capacity.

4. A method of fast self-learning of a cable network commissioning topology according to claim 3, characterized in that, The number of the negative electrode candidate subsets The calculation method is as follows: Wherein, is the total number of test nodes of the current negative electrode candidate set, is the reference capacity of the subset, is the upward rounding function.

5. A method of fast self-learning of a cable network commissioning topology according to claim 4, characterized in that, The connectivity data table uses a linked list structure to store a list of test nodes connected to each test node, with the test nodes in each list arranged in ascending order by their numbers.

6. A fast self-learning system for a cable network commissioning topology, characterized by: The method includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method according to any one of claims 1 to 5.

Images