Method and system for diagnosing device defects based on artificial intelligence knowledge graph

Abstract

This application relates to the field of equipment defect diagnosis technology, and discloses a method and system for equipment defect diagnosis based on artificial intelligence knowledge graphs. The method includes: presetting a monitoring cycle and an update cycle; acquiring first data, second data, and third data; training a defect diagnosis knowledge graph; performing unilateral defect diagnosis and collaborative defect diagnosis; and updating the defect diagnosis knowledge graph. The system corresponds to this method. This application achieves unilateral defect diagnosis of primary equipment by acquiring the first data, unilateral defect diagnosis of secondary equipment by acquiring the second data, and collaborative defect diagnosis of primary and secondary equipment by acquiring the third data. It also provides a data foundation for full-cycle data mining and improvement of artificial intelligence knowledge graph technology. By combining the monitoring cycle and the update cycle, it achieves full-cycle data mining; by combining collaborative defect diagnosis, it achieves improvement of artificial intelligence knowledge graph technology; and finally, it realizes equipment defect diagnosis based on artificial intelligence knowledge graphs.

Description

Technical Field

[0001] This application relates to the field of equipment defect diagnosis technology, specifically a method and system for equipment defect diagnosis based on artificial intelligence knowledge graphs. Background Technology

[0002] As the hardware foundation for maintaining the stable operation of a network, monitoring equipment is essential. Existing equipment monitoring systems can collect data throughout the entire equipment lifecycle and issue warnings when defects occur. However, it's clear that this data contains crucial information such as the events leading up to the defect, the specific defect, and effective solutions. Properly utilizing this information could not only diagnose equipment defects but also provide early warnings. In practice, however, difficulties in utilizing this data result in its underutilization. Therefore, mining data from the entire equipment lifecycle is a key technological direction for optimizing equipment defect diagnosis.

[0003] Artificial intelligence knowledge graph technology is an emerging and comprehensive technology capable of rapidly retrieving matching information from massive amounts of data. Its existing technical framework, with adaptive modifications, can be applied to equipment defect diagnosis. Therefore, improving artificial intelligence knowledge graph technology to better suit defect diagnosis is a reasonable technical direction for optimizing equipment defect diagnosis.

[0004] In a device network, primary and secondary devices form an effective integrated whole. However, current equipment supervision often focuses on the supervision of individual devices, neglecting the interconnectedness of them as a whole. Therefore, exploring the connections between primary and secondary devices and achieving collaborative diagnosis is a reasonable technical direction for optimizing equipment defect diagnosis.

[0005] Chinese patent number CN202010937208.1 discloses a method, system, device and storage medium for diagnosing defects in secondary power equipment. However, it relies solely on a pre-set defect database for obtaining diagnostic evidence and only monitors secondary equipment, failing to fully utilize data and comprehensively monitor equipment.

[0006] In summary, there is an urgent need for a technical solution that can achieve collaborative defect diagnosis of primary and secondary equipment based on full-cycle data mining and artificial intelligence knowledge graph technology. Summary of the Invention

[0007] The purpose of this application is to provide a device defect diagnosis method and system based on artificial intelligence knowledge graphs to solve the technical problems mentioned in the background.

[0008] To achieve the above objectives, this application discloses the following technical solutions:

[0009] In a first aspect, this application discloses a device defect diagnosis method based on artificial intelligence knowledge graphs, the method comprising:

[0010] S1: Preset monitoring period and update period, wherein the update period is a positive integer multiple of the monitoring period;

[0011] S2: Obtain first data from the primary equipment, second data from the secondary equipment, and third data associated with the primary and secondary equipment for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary equipment; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary equipment; the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between the primary and secondary equipment.

[0012] S3: Train the defect diagnosis knowledge graph; the defect diagnosis knowledge graph includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning; the defect diagnosis knowledge graph stores an expert defect database and a data defect database; the expert defect database is obtained from the professional knowledge of domain experts, and the data defect database is obtained from the results of data analysis;

[0013] S4: Based on the first data and using the unilateral map, perform unilateral defect diagnosis of the primary equipment and output the first defect diagnosis result;

[0014] S5: Based on the second data and using the unilateral map, perform unilateral defect diagnosis of the secondary equipment and output the second defect diagnosis result;

[0015] S6: Based on the third data and using the collaborative graph, perform collaborative defect diagnosis of primary and secondary equipment, and output the third defect diagnosis result;

[0016] S7: Obtain defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph.

[0017] Based on the above, unilateral defect diagnosis of primary equipment is achieved by acquiring first data, unilateral defect diagnosis of secondary equipment is achieved by acquiring second data, and collaborative defect diagnosis of primary and secondary equipment is achieved by acquiring third data. Furthermore, acquiring the first, second, and third data provides a data foundation for full-cycle data mining and artificial intelligence knowledge graph technology improvement. Full-cycle data mining is achieved by combining monitoring and update cycles, and artificial intelligence knowledge graph technology improvement is achieved by combining collaborative defect diagnosis. This enables collaborative defect diagnosis of primary and secondary equipment based on full-cycle data mining and artificial intelligence knowledge graph technology improvement without affecting the unilateral defect diagnosis of primary or secondary equipment.

[0018] Preferably, the acquisition and processing of the first data and the second data specifically include:

[0019] The first and second raw data are acquired using the acquisition devices set on the device and the corresponding communication interface;

[0020] Preprocess the first and second raw data to obtain the first training data and the second training data.

[0021] The first training data and the second training data are feature extracted using a pre-set expert defect database, and the data is stored in the form of a data chain associated with operational status defect resolution measures, thus obtaining the corresponding first data and the second data.

[0022] Preferably, the acquisition and processing of the third data specifically includes:

[0023] Obtain the first data and the second data, and perform time-series-based matching on the first data and the second data;

[0024] A preset association time threshold is set. When the absolute value of the difference between the occurrence time of the primary equipment defect in the first data and the occurrence time of the secondary equipment defect in the second data is less than or equal to the association time threshold, the expert defect database is traversed. When no matching defect is found, the first data and the second data are associated and defined as candidate third data.

[0025] A preset association frequency threshold is set. When the number of times the candidate third data appears in one update cycle is greater than or equal to the association frequency threshold, the candidate third data is stored in the form of an association running status association defect association solution association data chain to obtain the third data and store it in the data defect database.

[0026] Preferably, the training process of the one-sided map specifically includes the following steps:

[0027] A1: Store the first data and the second data into a preset one-sided graph architecture;

[0028] A2: Train the single-sided graph architecture after execution step A1, so that it can parse the input first data or second data to obtain the corresponding first running state or second running state;

[0029] A3: Using the corresponding first or second operating state obtained in step A2, perform one-sided device identification, one-sided relationship matching, and one-sided knowledge representation; the one-sided device identification is used to identify the corresponding primary or secondary device based on the operating state; the one-sided relationship matching is used to match the corresponding first defect and corresponding first solution and second defect and corresponding second solution based on the operating state using the expert defect library and the data defect library; the one-sided knowledge representation is used to represent the operating state, the corresponding primary or secondary device, and the corresponding first defect and corresponding first solution and second defect and corresponding second solution.

[0030] A4: Output the one-sided graph architecture after executing step A3 and define it as the one-sided graph.

[0031] Preferably, the training process of the collaborative graph specifically includes the following steps:

[0032] B1: Store the third data into a preset collaborative graph architecture;

[0033] B2: Train the collaborative graph architecture after execution step B1, so that it can parse the input third data to obtain the corresponding associated running state;

[0034] B3: Using the associated operating status obtained in step B2, perform collaborative device identification, collaborative relationship matching, and collaborative knowledge representation; the collaborative device identification is used to identify the corresponding primary and secondary device associations based on the associated operating status; the collaborative relationship matching is used to match the corresponding third defect and the corresponding third solution based on the associated operating status using the expert defect library and the data defect library; the collaborative knowledge representation is used to represent the associated operating status, the corresponding primary and secondary device associations, and the corresponding third defect and the corresponding third solution.

[0035] B4: Output the collaborative graph architecture after executing step B3 and define it as the collaborative graph.

[0036] Preferably, the output process of the first defect diagnosis result and the second defect diagnosis result specifically includes the following steps:

[0037] C1: Obtain the real-time first data and real-time second data at the end of one of the monitoring cycles, and execute steps C2 and C3 in parallel;

[0038] C2: Input the real-time first data into the one-sided graph and run it. When the first defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the diagnosis result of the first defect.

[0039] C3: Input the real-time second data into the one-sided graph and run it. When the second defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the diagnosis result of the second defect.

[0040] Preferably, the output process of the third defect diagnosis result specifically includes the following steps:

[0041] D1: Obtain real-time third data at the end of one of the monitoring cycles, and execute step D2 in parallel;

[0042] D2: Input the real-time third data into the collaborative graph and run it. When the third defect is matched, perform the corresponding collaborative knowledge representation, output the result of the collaborative knowledge representation and define it as the diagnosis result of the third defect.

[0043] Preferably, the process of outputting the third defect diagnosis result further includes the following steps:

[0044] D3: Analyze the matched third defect, and execute step D4 if the first defect exists or the second defect does not occur;

[0045] D4: Based on the third defect diagnosis result, obtain the first or second device corresponding to the first or second defect that has not occurred, and output the corresponding warning generated by the first or second device.

[0046] Preferably, the acquisition of defect diagnosis records and early warning records, and the updating of the defect diagnosis knowledge graph, specifically include:

[0047] Based on the monitoring cycle, the diagnostic results of the defect diagnosis records, including the unilateral defect diagnosis and the collaborative defect diagnosis, are obtained. The accuracy of the diagnostic results is judged, and the expert defect database and the data defect database are updated using the judgment results.

[0048] Based on the update cycle, the data defect database is updated when new third data is added.

[0049] Secondly, this application discloses a device defect diagnosis system based on artificial intelligence knowledge graphs. This system is applicable to the device defect diagnosis method based on artificial intelligence knowledge graphs as described in any of the preceding claims. The system includes:

[0050] A time module, configured with a preset monitoring period and an update period; the update period is a positive integer multiple of the monitoring period;

[0051] The data module is configured to: acquire first data from a primary device, second data from a secondary device, and third data associated with the primary and secondary devices for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary device; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary device; and the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between the primary and secondary devices.

[0052] The training module is configured to: train the defect diagnosis knowledge graph; the defect diagnosis knowledge graph includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning; the defect diagnosis knowledge graph stores an expert defect database and a data defect database; the expert defect database is obtained from the professional knowledge of domain experts, and the data defect database stores the results of data analysis;

[0053] The application module is configured to: perform unilateral defect diagnosis of primary equipment based on the first data and using the unilateral map, and output a first defect diagnosis result; perform unilateral defect diagnosis of secondary equipment based on the second data and using the unilateral map, and output a second defect diagnosis result; and perform collaborative defect diagnosis of primary and secondary equipment based on the third data and using the collaborative map, and output a third defect diagnosis result.

[0054] The update module is configured to: acquire defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph.

[0055] Beneficial effects: The equipment defect diagnosis method and system based on artificial intelligence knowledge graph of this application realizes unilateral defect diagnosis of primary equipment by acquiring first data, unilateral defect diagnosis of secondary equipment by acquiring second data, and collaborative defect diagnosis of primary and secondary equipment by acquiring third data. Moreover, the acquisition of first, second and third data provides a data foundation for full-cycle data mining and artificial intelligence knowledge graph technology improvement. Full-cycle data mining is realized by combining monitoring cycle and update cycle, and artificial intelligence knowledge graph technology improvement is realized by combining collaborative defect diagnosis. It realizes collaborative defect diagnosis of primary and secondary equipment based on full-cycle data mining and artificial intelligence knowledge graph technology improvement without affecting the unilateral defect diagnosis of primary or secondary equipment. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating the device defect diagnosis method based on artificial intelligence knowledge graph provided in this application embodiment;

[0058] Figure 2 This is a structural block diagram of an artificial intelligence knowledge graph-based device defect diagnosis system provided in an embodiment of this application. Detailed Implementation

[0059] The technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0060] In this document, the term "comprising" is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] This embodiment discloses, in its first aspect, as follows: Figure 1 The method for diagnosing equipment defects based on artificial intelligence knowledge graphs is shown below. The method includes:

[0062] S1: Preset monitoring period and update period, with the update period being a positive integer multiple of the monitoring period.

[0063] S2: Obtain first data of primary equipment, second data of secondary equipment, and third data associated with primary and secondary equipment for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary equipment; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary equipment; the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between primary and secondary equipment; in a simple example, primary equipment may be, but is not limited to, equipment directly used for production, transportation, and distribution, such as generators, transformers, circuit breakers, and busbars; secondary equipment may be, but is not limited to, equipment used for monitoring, controlling, regulating, and protecting primary equipment, such as relay protection devices, measuring instruments, and control circuits.

[0064] S3: Train the defect diagnosis knowledge graph; the defect diagnosis knowledge graph includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning; the defect diagnosis knowledge graph stores an expert defect library and a data defect library; the expert defect library is obtained from the professional knowledge of domain experts, and the data defect library is stored from the results of data analysis; it should be noted that the expert defect library in this embodiment also needs to be verified by data analysis.

[0065] S4: Based on the first data and using the one-sided spectrum, perform a one-sided defect diagnosis of the equipment and output the first defect diagnosis result.

[0066] S5: Based on the second data and using the one-sided spectrum, perform one-sided defect diagnosis of the secondary equipment and output the second defect diagnosis result.

[0067] S6: Based on the third data and using the collaborative graph, perform collaborative defect diagnosis of primary and secondary equipment, and output the third defect diagnosis result.

[0068] S7: Obtain defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph.

[0069] By acquiring first data, unilateral defect diagnosis of primary equipment is achieved; by acquiring second data, unilateral defect diagnosis of secondary equipment is achieved; and by acquiring third data, collaborative defect diagnosis of primary and secondary equipment is achieved. Furthermore, acquiring the first, second, and third data provides a data foundation for full-cycle data mining and improvements to artificial intelligence knowledge graph technology. Full-cycle data mining is achieved by combining monitoring and update cycles, and improvements to artificial intelligence knowledge graph technology are achieved by combining collaborative defect diagnosis. This enables collaborative defect diagnosis of primary and secondary equipment based on full-cycle data mining and improvements to artificial intelligence knowledge graph technology without affecting unilateral defect diagnosis of primary or secondary equipment.

[0070] Specifically, the acquisition and processing of the first and second data include:

[0071] The first and second raw data are acquired using acquisition devices set on the equipment and the corresponding communication interface. In a simple example, the first raw data includes electrical parameters, environmental parameters, vibration data, partial discharge data and insulation resistance, etc., and the second raw data includes the action signals and fault records of the protection device, the status information of the control circuit and the operation data of the automation system.

[0072] Preprocess the first and second raw data to obtain the first and second training data.

[0073] The first and second training data are feature extracted using a pre-defined expert defect database, and the data is stored in the form of a data chain relating operational status defect resolution measures, thus obtaining the corresponding first and second data.

[0074] The acquisition of first and second data provides a data foundation for unilateral defect diagnosis, and the storage of data in the form of a data chain linking operational status defect solutions provides technical support for constructing a unilateral map, which ensures the normal conduct of routine unilateral equipment defect diagnosis.

[0075] Specifically, the acquisition and processing of third-party data includes:

[0076] Obtain the first data and the second data, and perform time-series-based matching on the first data and the second data;

[0077] A preset association time threshold is set. When the absolute value of the difference between the occurrence time of the primary equipment defect in the first data and the occurrence time of the secondary equipment defect in the second data is less than or equal to the association time threshold, the expert defect database is traversed. When no matching defect is found, the first data and the second data are associated and defined as candidate third data. The comparison with the association time threshold provides an initial judgment basis for data association.

[0078] A preset association frequency threshold is set. When the number of times a candidate third data appears in an update cycle is greater than or equal to the association frequency threshold, the candidate third data is stored in the form of an association running status association defect association solution association data chain, and the third data is obtained and stored in the data defect database. The comparison with the association frequency threshold provides the final judgment basis for data association.

[0079] Acquiring third-party data provides a data foundation for collaborative defect diagnosis, and storing it in the form of a data chain that links operational status, defects, and solutions provides technical support for constructing a collaborative graph. This improves upon existing artificial intelligence knowledge graph technology, providing the necessary technical means to conduct more comprehensive defect diagnosis of the effective combination of primary and secondary equipment.

[0080] Specifically, the training process for a one-sided atlas includes the following steps:

[0081] A1: Store the first and second data into the preset one-sided graph architecture.

[0082] A2: Train the one-sided graph architecture after step A1, so that it can parse the first or second input data to obtain the corresponding first or second running state.

[0083] A3: Using the corresponding first or second operating state obtained in step A2, perform one-sided device identification, one-sided relationship matching, and one-sided knowledge representation; one-sided device identification is used to identify the corresponding primary or secondary device based on the operating state; one-sided relationship matching is used to match the corresponding first defect and corresponding first solution and second defect and corresponding second solution based on the operating state using the expert defect library and the data defect library; one-sided knowledge representation is used to represent the operating state, the corresponding primary or secondary device, and the corresponding first defect and corresponding first solution and second defect and corresponding second solution.

[0084] A4: Output the one-sided graph architecture after executing step A3 and define it as a one-sided graph.

[0085] Training unilateral maps provides technical support for unilateral defect diagnosis.

[0086] Specifically, the training process of the collaborative graph includes the following steps:

[0087] B1: Store the third data into the preset collaborative graph architecture.

[0088] B2: Train the collaborative graph architecture after step B1, so that it can parse the input third data to obtain the corresponding associated running state.

[0089] B3: Using the corresponding associated operating status obtained in step B2, perform collaborative device identification, collaborative relationship matching, and collaborative knowledge representation; collaborative device identification is used to identify the corresponding primary and secondary device associations based on the associated operating status; collaborative relationship matching is used to match the corresponding third defect and the corresponding third solution based on the associated operating status using the expert defect library and the data defect library; collaborative knowledge representation is used to represent the associated operating status, the corresponding primary and secondary device associations, and the corresponding third defect and the corresponding third solution.

[0090] B4: Output the collaborative graph architecture after executing step B3 and define it as a collaborative graph.

[0091] Training a collaborative graph provides technical support for diagnosing collaborative defects.

[0092] Specifically, the output process of the first defect diagnosis result and the second defect diagnosis result includes the following steps:

[0093] C1: Obtain the first real-time data and the second real-time data at the end of a monitoring cycle, and execute steps C2 and C3 in parallel; it should be noted that the parallel execution in this embodiment aims to improve the timeliness of unilateral defect monitoring, making the defect diagnosis results more timely.

[0094] C2: Input the real-time first data into the one-sided graph and run it. When the first defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the first defect diagnosis result.

[0095] C3: Input the real-time second data into the one-sided graph and run it. When a second defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the second defect diagnosis result.

[0096] By outputting one or more of the first and second defect results, equipment managers can obtain information on unilateral defects, the corresponding equipment, and the corresponding solutions in a timely manner, thereby improving the level of intelligence in equipment defect diagnosis.

[0097] Specifically, the output process of the third defect diagnosis results includes the following steps:

[0098] D1: Obtain real-time third-party data at the end of a monitoring cycle and execute step D2 in parallel.

[0099] D2: Input real-time third data into the collaborative graph and run it. When a third defect is matched, perform the corresponding collaborative knowledge representation, output the result of the collaborative knowledge representation and define it as the third defect diagnosis result.

[0100] By outputting third-party defect diagnosis results, equipment managers can obtain relevant defect information, corresponding equipment, and corresponding solutions in a timely manner, thereby improving the level of intelligence in equipment defect diagnosis.

[0101] Specifically, the output process of the third defect diagnosis results also includes the following steps:

[0102] D3: Parse the matched third defect. If the first defect exists or the second defect has not occurred, proceed to step D4.

[0103] D4: Based on the third defect diagnosis result, obtain the first or second device corresponding to the first or second defect that has not occurred, and output the corresponding early warning generated by the first or second device; it should be noted that the judgment based on the associated time threshold in this embodiment provides a buffer time between the occurrence of defects in the primary or secondary device, which increases the possibility of issuing an early warning and realizes the early prevention of equipment defects.

[0104] By issuing corresponding early warnings, equipment managers can take appropriate measures to address the corresponding equipment in advance, even if the associated defects have not yet occurred, thereby improving the level of equipment protection and enhancing the intelligence level of equipment defect diagnosis.

[0105] Specifically, this involves acquiring defect diagnosis records and early warning records, and updating the defect diagnosis knowledge graph, including:

[0106] Based on the monitoring cycle, the diagnostic results of defect diagnosis records, including unilateral defect diagnosis and collaborative defect diagnosis, are obtained. The accuracy of the diagnostic results is judged, and the judgment results are used to update the expert defect database and the data defect database.

[0107] Based on the update cycle, the data defect database is updated when new third data is added.

[0108] By periodically updating the expert defect database and the data defect database, full-cycle data mining was achieved, which improved data utilization and optimized data integration. It also provided data support for improving the application of artificial intelligence knowledge graph technology in equipment defect diagnosis.

[0109] This embodiment discloses, in a second aspect, as follows: Figure 2 The system illustrates an equipment defect diagnosis system based on an artificial intelligence knowledge graph. This system is applicable to the equipment defect diagnosis method based on an artificial intelligence knowledge graph as described in any of the preceding claims. The system includes:

[0110] The time module is configured with a preset monitoring period and an update period; the update period is a positive integer multiple of the monitoring period.

[0111] The data module is configured to: acquire first data from primary equipment, second data from secondary equipment, and third data associated with primary and secondary equipment for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary equipment; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary equipment; and the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between primary and secondary equipment.

[0112] The training module is configured as follows: It trains a defect diagnosis knowledge graph, which includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning. The defect diagnosis knowledge graph stores an expert defect database and a data defect database. The expert defect database is derived from the professional knowledge of domain experts, while the data defect database stores data derived from data analysis results.

[0113] The application module is configured to: perform unilateral defect diagnosis of primary equipment based on the first data and using a unilateral graph, and output the first defect diagnosis result; perform unilateral defect diagnosis of secondary equipment based on the second data and using a unilateral graph, and output the second defect diagnosis result; and perform collaborative defect diagnosis of primary and secondary equipment based on the third data and using a collaborative graph, and output the third defect diagnosis result.

[0114] The update module is configured to: obtain defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph.

[0115] It should be noted that the device defect diagnosis system based on artificial intelligence knowledge graph in this embodiment corresponds to the aforementioned device defect diagnosis method based on artificial intelligence knowledge graph. Therefore, the contents of the device defect diagnosis system based on artificial intelligence knowledge graph in this embodiment that are not specifically described, including but not limited to specific functional definitions, working principles and technical effects, are the same as those of the aforementioned device defect diagnosis method based on artificial intelligence knowledge graph, and will not be repeated here.

[0116] In summary, the equipment defect diagnosis method and system based on artificial intelligence knowledge graphs in this embodiment achieve unilateral defect diagnosis of primary equipment by acquiring first data, unilateral defect diagnosis of secondary equipment by acquiring second data, and collaborative defect diagnosis of primary and secondary equipment by acquiring third data. Furthermore, acquiring the first, second, and third data provides a data foundation for full-cycle data mining and improvement of artificial intelligence knowledge graph technology. Full-cycle data mining is achieved by combining monitoring and update cycles, and artificial intelligence knowledge graph technology is improved by combining collaborative defect diagnosis. This achieves collaborative defect diagnosis of primary and secondary equipment based on full-cycle data mining and improved artificial intelligence knowledge graph technology without affecting unilateral defect diagnosis of primary or secondary equipment.

[0117] In the embodiments provided in this application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to implement the functions described herein, or combinations thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the associated hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of a computer program from one place to another. Storage media may be any available medium accessible to a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code having the form of instructions or data structures and accessible to a computer.

[0118] Finally, it should be noted that the above description is only a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for diagnosing equipment defects based on artificial intelligence knowledge graphs, characterized in that, The method includes: S1: Preset monitoring period and update period, wherein the update period is a positive integer multiple of the monitoring period; S2: Obtain first data from the primary equipment, second data from the secondary equipment, and third data associated with the primary and secondary equipment for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary equipment; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary equipment; the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between the primary and secondary equipment. S3: Train the defect diagnosis knowledge graph; the defect diagnosis knowledge graph includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning; the defect diagnosis knowledge graph stores an expert defect database and a data defect database; the expert defect database is obtained from the professional knowledge of domain experts, and the data defect database is obtained from the results of data analysis; S4: Based on the first data and using the unilateral map, perform unilateral defect diagnosis of the primary equipment and output the first defect diagnosis result; S5: Based on the second data and using the unilateral map, perform unilateral defect diagnosis of the secondary equipment and output the second defect diagnosis result; S6: Based on the third data and using the collaborative graph, perform collaborative defect diagnosis of primary and secondary equipment, and output the third defect diagnosis result; S7: Obtain defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph; The acquisition and processing of the first data and the second data specifically include: The first and second raw data are acquired using the acquisition devices set on the device and the corresponding communication interface; Preprocess the first and second raw data to obtain the first training data and the second training data. The first training data and the second training data are feature extracted using a pre-set expert defect database, and the data is stored in the form of a data chain associated with operational status defect resolution measures, so as to obtain the corresponding first data and the second data. The acquisition and processing of the third data specifically includes: Obtain the first data and the second data, and perform time-series-based matching on the first data and the second data; A preset association time threshold is set. When the absolute value of the difference between the occurrence time of the primary equipment defect in the first data and the occurrence time of the secondary equipment defect in the second data is less than or equal to the association time threshold, the expert defect database is traversed. When no matching defect is found, the first data and the second data are associated and defined as candidate third data. A preset association frequency threshold is set. When the number of times the candidate third data appears in one update cycle is greater than or equal to the association frequency threshold, the candidate third data is stored in the form of an association running status association defect association solution association data chain to obtain the third data and store it in the data defect database.

2. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 1, characterized in that, The training process of the one-sided map specifically includes the following steps: A1: Store the first data and the second data into a preset one-sided graph architecture; A2: Train the single-sided graph architecture after execution step A1, so that it can parse the input first data or second data to obtain the corresponding first running state or second running state; A3: Using the corresponding first or second operating state obtained in step A2, perform one-sided device identification, one-sided relationship matching, and one-sided knowledge representation; the one-sided device identification is used to identify the corresponding primary or secondary device based on the operating state; the one-sided relationship matching is used to match the corresponding first defect and corresponding first solution and second defect and corresponding second solution based on the operating state using the expert defect library and the data defect library; the one-sided knowledge representation is used to represent the operating state, the corresponding primary or secondary device, and the corresponding first defect and corresponding first solution and second defect and corresponding second solution. A4: Output the one-sided graph architecture after executing step A3 and define it as the one-sided graph.

3. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 1, characterized in that, The training process of the collaborative graph specifically includes the following steps: B1: Store the third data into a preset collaborative graph architecture; B2: Train the collaborative graph architecture after execution step B1, so that it can parse the input third data to obtain the corresponding associated running state; B3: Using the associated operating status obtained in step B2, perform collaborative device identification, collaborative relationship matching, and collaborative knowledge representation; the collaborative device identification is used to identify the corresponding primary and secondary device associations based on the associated operating status; the collaborative relationship matching is used to match the corresponding third defect and the corresponding third solution based on the associated operating status using the expert defect library and the data defect library; the collaborative knowledge representation is used to represent the associated operating status, the corresponding primary and secondary device associations, and the corresponding third defect and the corresponding third solution. B4: Output the collaborative graph architecture after executing step B3 and define it as the collaborative graph.

4. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 2, characterized in that, The output process of the first defect diagnosis result and the second defect diagnosis result specifically includes the following steps: C1: Obtain the real-time first data and real-time second data at the end of one of the monitoring cycles, and execute steps C2 and C3 in parallel; C2: Input the real-time first data into the one-sided graph and run it. When the first defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the diagnosis result of the first defect. C3: Input the real-time second data into the one-sided graph and run it. When the second defect is matched, perform the corresponding one-sided knowledge representation, output the result of the one-sided knowledge representation and define it as the diagnosis result of the second defect.

5. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 3, characterized in that, The process of outputting the third defect diagnosis result specifically includes the following steps: D1: Obtain real-time third data at the end of one of the monitoring cycles, and execute step D2 in parallel; D2: Input the real-time third data into the collaborative graph and run it. When the third defect is matched, perform the corresponding collaborative knowledge representation, output the result of the collaborative knowledge representation and define it as the diagnosis result of the third defect.

6. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 5, characterized in that, The process of outputting the third defect diagnosis result also includes the following steps: D3: Analyze the matched third defect, and execute step D4 if the first defect exists or the second defect does not occur; D4: Based on the third defect diagnosis result, obtain the first or second device corresponding to the first or second defect that has not occurred, and output the corresponding warning generated by the first or second device.

7. The equipment defect diagnosis method based on artificial intelligence knowledge graph according to claim 1, characterized in that, The acquisition of defect diagnosis records and early warning records, and the updating of the defect diagnosis knowledge graph, specifically include: Based on the monitoring cycle, the diagnostic results of the defect diagnosis records, including the unilateral defect diagnosis and the collaborative defect diagnosis, are obtained. The accuracy of the diagnostic results is judged, and the expert defect database and the data defect database are updated using the judgment results. Based on the update cycle, the data defect database is updated when new third data is added.

8. A device defect diagnosis system based on artificial intelligence knowledge graph, wherein the system is applicable to the device defect diagnosis method based on artificial intelligence knowledge graph as described in any one of claims 1-7, characterized in that, The system includes: A time module, configured with a preset monitoring period and an update period; the update period is a positive integer multiple of the monitoring period; The data module is configured to: acquire first data from a primary device, second data from a secondary device, and third data associated with the primary and secondary devices for training a preset defect diagnosis knowledge graph; the first data is used to characterize the first operating state, first defect, and corresponding first solution of the primary device; the second data is used to characterize the second operating state, second defect, and corresponding second solution of the secondary device; and the third data is used to characterize the associated operating state, associated defect, and corresponding associated solution between the primary and secondary devices. The training module is configured to: train the defect diagnosis knowledge graph; the defect diagnosis knowledge graph includes a one-sided graph and a collaborative graph, and is used for defect diagnosis and early warning; the defect diagnosis knowledge graph stores an expert defect database and a data defect database; the expert defect database is obtained from the professional knowledge of domain experts, and the data defect database stores the results of data analysis; The application module is configured to: perform unilateral defect diagnosis of primary equipment based on the first data and using the unilateral map, and output a first defect diagnosis result; perform unilateral defect diagnosis of secondary equipment based on the second data and using the unilateral map, and output a second defect diagnosis result; and perform collaborative defect diagnosis of primary and secondary equipment based on the third data and using the collaborative map, and output a third defect diagnosis result. The update module is configured to: acquire defect diagnosis records and early warning records, and update the defect diagnosis knowledge graph.

Images