Rolling mill equipment point inspection and management method, system and device based on fault knowledge base

Abstract

The application discloses a rolling mill equipment point inspection and management method, system and device based on a fault knowledge base, the method comprising: preprocessing and structuring processing historical fault records, historical fault analysis reports and historical examination mechanism information of collected rolling mill equipment to obtain target field information; based on the target field information, constructing a target fault knowledge base, and based on the target fault knowledge base, constructing a target vector database; performing intention recognition on a received natural language question sentence to obtain an intention recognition type; according to the intention recognition type, searching in the target fault knowledge base and the target vector database to obtain multiple similar historical case information, fault law information and examination requirement information; calling a pre-trained intelligent agent to analyze the multiple similar historical case information, the fault law information and the examination requirement information to output an execution result, thereby solving the technical problem of low intelligent management efficiency of the rolling mill equipment.

Description

Technical Field

[0001] This application relates to the field of intelligent management technology for metallurgical equipment, and in particular to a method, system and device for inspection and management of rolling mill equipment based on a fault knowledge base. Background Technology

[0002] Currently, equipment management in the rolling mill area mainly relies on manual experience for inspection, fault diagnosis, maintenance decision-making, and performance evaluation. This means that the focus of equipment inspection, fault prediction, and maintenance measures are highly dependent on the experience of experienced operators. Because a large number of historical fault records, analysis reports, and performance evaluation mechanisms are scattered and lack a unified, searchable knowledge base, inspection prediction lacks historical data support, fault analysis is inefficient and prone to errors, and it is difficult to achieve closed-loop management from inspection to performance evaluation. This results in the technical problem of low efficiency in the intelligent management of rolling mill equipment. Summary of the Invention

[0003] In view of this, the embodiments of this application provide a method, system and device for inspection and management of rolling mill equipment based on a fault knowledge base, the main purpose of which is to solve the technical problem of low efficiency in intelligent management of rolling mill equipment.

[0004] According to one aspect of this application, a method for inspection and management of rolling mill equipment based on a fault knowledge base is provided. The method includes: collecting historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment; preprocessing and structuring the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain target field information of the rolling mill equipment; constructing a target fault knowledge base based on the target field information, and constructing a target vector database based on the target fault knowledge base; receiving natural language questions from target objects, and performing intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions, wherein the intent recognition type is used to represent the business scenario corresponding to the natural language questions; searching in the target fault knowledge base and the target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information; and calling a pre-trained intelligent agent to analyze the multiple similar historical case information, fault pattern information, and assessment requirement information to output execution results, wherein the execution results are used to represent executable decision information corresponding to the natural language questions.

[0005] Optionally, the original fault records, original fault analysis reports, and original assessment mechanism information of the rolling mill equipment are collected; in response to the fact that the generation time of the original fault record is greater than or equal to the preset generation time, the generation time of the original fault analysis report is greater than or equal to the preset generation time, and the generation time of the original assessment mechanism information is greater than or equal to the preset generation time, the original fault record, original fault analysis report, and original assessment mechanism information are respectively identified as historical fault records, historical fault analysis reports, and historical assessment mechanism information.

[0006] Optionally, a target fault knowledge base is constructed based on the target field information, including: constructing an initial fault knowledge base based on the target field information; obtaining the current fault records, current fault analysis reports, and current assessment mechanism information of the rolling mill equipment, and updating the initial fault knowledge base based on the current fault records, current fault analysis reports, and current assessment mechanism information to obtain the target fault knowledge base.

[0007] Optionally, a target vector database is constructed based on the target fault knowledge base, including: acquiring unstructured data information of the rolling mill equipment and structured data information from the target fault knowledge base; performing vectorization processing on the unstructured data information and the structured data information respectively to obtain vectorized unstructured data information and structured data information; and constructing the target vector database based on the vectorized unstructured data information and structured data information.

[0008] Optionally, a target vector database is constructed based on the vectorized unstructured data and structured data, including: acquiring the business logic management information of the rolling mill equipment; constructing a knowledge association system based on the vectorized unstructured data and structured data, combined with the business logic management information; and constructing the target vector database according to the knowledge association system.

[0009] Optionally, the method further includes: target field information including at least: equipment area, equipment location, fault category, fault time, fault phenomenon, root cause analysis information, handling measures information, rectification result information, assessment clause information, responsible party, and fault handling status; intent identification type including at least: inspection prediction, fault analysis, measure formulation, training and learning, and management reminder; execution result including at least: inspection location of rolling mill equipment, fault causes and similar cases of rolling mill equipment, handling measures of rolling mill equipment, preventive measures of rolling mill equipment, assessment and management information of rolling mill equipment, responsible party, completion time limit of rolling mill equipment inspection task, and verification method of rolling mill equipment inspection task.

[0010] According to another aspect of this application, a rolling mill equipment inspection and management system based on a fault knowledge base is provided. This system includes: a fault knowledge base module for collecting historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment; a data structuring and vectorization module for preprocessing and structuring the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain target field information of the rolling mill equipment; a target fault knowledge base for constructing based on the target field information, and a target vector database for constructing based on the target fault knowledge base; and an intent recognition and question answering module for receiving natural language questions from the target object. The system performs intent recognition on natural language questions to obtain the intent recognition type of the natural language questions. The intent recognition type is used to represent the business scenario corresponding to the natural language questions. The target inspection agent of the rolling mill equipment is combined with multiple similar historical case information, fault pattern information and assessment requirement information to output the execution result. The execution result is used to represent the executable decision information corresponding to the natural language questions. The retrieval enhancement generation module is used to search in the target fault knowledge base and target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information and assessment requirement information.

[0011] According to another aspect of this application, a rolling mill equipment inspection and management device based on a fault knowledge base is provided. The device includes: a data acquisition unit, used to acquire historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, and to preprocess and structure the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain target field information of the rolling mill equipment; a construction unit, used to construct a target fault knowledge base based on the target field information, and to construct a target vector database based on the target fault knowledge base; a first acquisition unit, used to receive natural language questions from a target object, and to perform intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions, wherein the intent recognition type is used to represent the business scenario corresponding to the natural language questions; a second acquisition unit, used to search in the target fault knowledge base and the target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information; and an output unit, used to call a pre-trained intelligent agent to analyze the multiple similar historical case information, fault pattern information, and assessment requirement information to output execution results, wherein the execution results are used to represent executable decision information corresponding to the natural language questions.

[0012] According to another aspect of this application, a storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the above-described method for inspection and management of rolling mill equipment based on a fault knowledge base.

[0013] According to another aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor. When the processor executes the program, it implements the above-described method for inspection and management of rolling mill equipment based on a fault knowledge base.

[0014] Using the above technical solution, this application provides a method for inspection and management of rolling mill equipment based on a fault knowledge base. This method first collects historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, and completes preprocessing and structuring to construct a target fault knowledge base and a target vector database. Then, it receives natural language queries from users and identifies the intent recognition type corresponding to their business scenarios. Based on the intent recognition type, it searches the aforementioned target fault knowledge base and target vector database to obtain multiple similar historical case information, fault pattern information, and assessment requirement information. Finally, through a target inspection intelligent agent and combining the aforementioned multiple similar historical case information, fault pattern information, and assessment requirement information, it outputs executable decision information that aligns with actual business operations, thus forming a complete technical chain from data processing to intelligent decision-making. By constructing a reusable knowledge base through structured and vectorized processing of multi-source data, and combining the retrieval and reasoning capabilities of artificial intelligence (AI) agents, the system achieves automatic prediction of key inspection points of rolling mill equipment, rapid analysis of fault causes, automatic generation of handling measures, and automatic matching of assessment requirements. This achieves full-process automation and intelligence from inspection to assessment, solving the technical problem of low efficiency in intelligent management of rolling mill equipment and realizing the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0015] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a method for inspection and management of rolling mill equipment based on a fault knowledge base, provided in an embodiment of this application, is shown. Figure 2 This application provides a schematic flowchart of another method for inspection and management of rolling mill equipment based on a fault knowledge base, as illustrated in an embodiment of this application. Figure 3 This illustration shows a structural diagram of a rolling mill equipment inspection and management system based on a fault knowledge base, provided in an embodiment of this application. Figure 4 This illustration shows a schematic diagram of another rolling mill equipment inspection and management system based on a fault knowledge base provided in an embodiment of this application; Figure 5 This illustration shows a structural schematic diagram of a rolling mill equipment inspection and management device based on a fault knowledge base, provided in an embodiment of this application. Figure 6 A schematic diagram of the device structure of a computer device provided in an embodiment of this application is shown. Detailed Implementation

[0017] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0018] In this embodiment, Figure 1 This illustration shows a flowchart of a rolling mill equipment inspection and management method based on a fault knowledge base, as provided in an embodiment of this application. Figure 1 As shown, the method includes: Step S101: Collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, and preprocess and structure the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain the target field information of the rolling mill equipment.

[0019] In the technical solution provided in step S101 of this application, historical fault records, historical fault analysis reports and historical assessment mechanism information of rolling mill equipment are collected, and the obtained historical fault records, historical fault analysis reports and historical assessment mechanism information are preprocessed and structured to achieve the purpose of obtaining target field information of rolling mill equipment.

[0020] Optionally, historical fault records may include: the time of fault occurrence, the fault symptoms, and records of major incidents, etc. It should be noted that this is merely an example illustrating the possible contents of historical fault records, and does not impose specific limitations on their content.

[0021] Optionally, the historical fault analysis report may include a simple fault analysis and a fault report. Historical assessment mechanism information, which can be simply referred to as assessment mechanism information, is extracted from equipment management assessment system documents. Target field information can be called standardized fields. Target field information includes at least: equipment area, fault phenomenon, and assessment clauses.

[0022] For example, historical fault analysis reports, major accident analysis reports, and equipment management assessment documents of rolling mill equipment in the rolling mill area are collected. During the structured processing, for an original fault record of "main motor overheating", standardized fields are extracted and transformed. That is, through this process, scattered original data are transformed into unified target field information containing dimensions such as equipment area, fault phenomenon, and assessment clauses.

[0023] Step S102: Based on the target field information, construct a target fault knowledge base, and based on the target fault knowledge base, construct a target vector database.

[0024] In the technical solution provided in step S102 of this application, a target fault knowledge base can be constructed based on the target field information obtained in the above steps, and then a target vector database can be constructed based on the target fault knowledge base obtained above. The target vector database can store unstructured data information of the rolling mill equipment.

[0025] For example, structured records of "roll bearing wear" (including target fields such as equipment location, fault symptoms, and root cause analysis) are stored in a target fault knowledge base, forming a standardized relational data table to enable field-based retrieval. Simultaneously, unstructured text descriptions such as the structured "roll bearing wear" records and "roll mill abnormal vibration fault analysis reports" are vectorized into high-dimensional vectors that are understandable by computers and stored in a target vector database. When a user queries, the system can accurately match structured fields and retrieve semantically similar case texts through vector calculations, thus constructing a database system that balances precise querying and semantic reasoning capabilities.

[0026] It should be noted that this is only a preferred implementation method for constructing the target vector database, and the process and method of constructing the target vector database are not specifically limited. As long as the target fault database is used as the basis, the process and method of constructing the target vector database are within the protection scope of this application, and will not be listed here.

[0027] Step S103: Receive a natural language question from the target object, and perform intent recognition on the natural language question to obtain the intent recognition type of the natural language question.

[0028] In the technical solution provided in step S103 of this application, a natural language question statement is received from the target object, and the intent of the natural language question statement is identified through an intent recognition model to obtain the intent recognition type of the natural language question statement. The natural language question statement can be referred to as a user question.

[0029] Optionally, the intent recognition type is used to characterize the business scenario corresponding to the natural language question. The intent recognition model is an algorithmic model based on natural language processing (NLP) technology, used to analyze the natural language question input by the user, extract keywords and semantic features to determine the user's specific business needs.

[0030] Optionally, the target object can be the person asking the question, which can be called a user, such as an inspector, equipment manager, maintenance personnel, etc.

[0031] For example, when receiving a natural language question from an inspector: "What are the key points for this week's inspection of the rolling mill area?", the intent recognition model analyzes the question and identifies that the user's request is to pre-determine the key points of the inspection work. Therefore, the intent recognition type of this natural language question is determined to be "inspection prediction".

[0032] It should be noted that this is only a preferred implementation method for obtaining the intent recognition type of a natural language question statement, and the method and process for obtaining the intent recognition type of a natural language question statement are not specifically limited.

[0033] Step S104: According to the intent recognition type, search the target fault knowledge base and target vector database to obtain information on multiple similar historical cases, fault patterns, and assessment requirements.

[0034] In the technical solution provided in step S104 of this application, the intent recognition type obtained in the above steps can be retrieved in the target fault knowledge base and the target vector database to obtain information on multiple similar historical cases, fault patterns, and assessment requirements. The target fault knowledge base can be simply referred to as the fault knowledge base. The target vector database can be simply referred to as the vector database.

[0035] For example, when the control system of a rolling mill identifies the intent type of a user's question, "How to handle overheating of the main motor of the rolling mill?" as "fault analysis," it first precisely matches the structured fields "equipment location: main motor" and "fault phenomenon: overheating" in the target fault knowledge base to retrieve relevant historical fault records. Simultaneously, it retrieves analysis report fragments describing "abnormal motor cooling system" or "overload" through semantic similarity calculation in the target vector database. Combining the results from both parts, the control system summarizes information on multiple similar historical cases such as "bearing wear" and "cooling fan failure," identifies the fault pattern of "frequent occurrences during the high-temperature summer," and matches assessment requirements related to "deduction of performance points for untimely shutdown inspections," providing data support for subsequent decision-making.

[0036] It should be noted that this is only a preferred implementation method for obtaining information on multiple similar historical cases, fault patterns, and assessment requirements. No specific limitations are made on the method and process of obtaining such information.

[0037] Step S105: Call the pre-trained agent to analyze information on multiple similar historical cases, fault patterns, and assessment requirements to output the execution results.

[0038] In the technical solution provided in step S105 of this application, a pre-trained intelligent agent is invoked to analyze multiple similar historical case information, fault pattern information, and assessment requirement information to achieve the purpose of outputting execution results. The execution results are used to represent executable decision information corresponding to natural language question statements. The pre-trained intelligent agent can be called a pre-trained inspection intelligent agent.

[0039] Optionally, the obtained historical fault records, historical fault analysis reports, and historical assessment mechanism information can be constructed into a training set in the form of "question-answer pairs." For example: Question: "The temperature of the main motor of the rolling mill is abnormally high. What are the possible causes?" Standard answers (extracted from historical reports): "①. Cooling fan filter blockage (occurred 3 times in history); ②. Poor lubrication (occurred 2 times in history); ③. Bearing wear (occurred 1 time in history)." The model is supervisedly trained through a large number of such "question-standard answer" pairs to obtain a pre-trained agent that learns question-answering patterns and logic consistent with the professional scenarios described in this application.

[0040] Furthermore, in practical applications, when a user raises a new question (such as "Key points for this week's inspection"), the control system will retrieve, in real time, historical cases most relevant to the current question from the target fault knowledge base and target vector database (such as faults from the same period or faults in the same location). Generation: When generating the final answer, the AI ​​agent will use these retrieved specific cases with clear historical data support as a basis, integrating them into the answer to ensure that the output is "evidence-based," conforms to the actual situation, and can be displayed in a structured manner. This means it can automatically output according to predetermined formats such as "Possible Causes," "Suggested Measures," "Historical Basis," and "Assessment Requirements," facilitating execution by inspection and maintenance personnel.

[0041] For example, when the intent recognition type is "fault analysis," the control system invokes a pre-trained inspection agent. The inspection agent first reads retrieved historical case information similar to "main motor bearing wear," and combines this with the fault pattern information that "faults often occur during the high-load period in summer" to deduce the main cause of the current fault. Next, based on the clauses regarding "accountability for major equipment accidents" in the assessment requirements, the inspection agent comprehensively analyzes the information and outputs an execution result containing specific handling steps, preventative measures, and compliance recommendations. The output result is: "It is recommended to immediately stop the machine and check the bearing lubrication; the estimated repair time is 4 hours; because it complies with the summer inspection specifications, the assessment points can be waived," thus providing the inspector with accurate and actionable decisions.

[0042] It should be noted that this is only a preferred implementation method for outputting execution results, and the process and method of outputting execution results are not specifically limited. As long as the process and method of calling a pre-trained agent to analyze multiple similar historical case information, fault pattern information and assessment requirement information to output execution results are within the protection scope of this application, and will not be listed here.

[0043] In the technical solutions provided by steps S101 to S105 of this application, historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment are first collected and preprocessed and structured to construct a target fault knowledge base and a target vector database. Then, the user's natural language question is received and the intent recognition type of the corresponding business scenario is identified. The target fault knowledge base and target vector database obtained above are searched according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information. Finally, through the target inspection intelligent agent and combined with the above multiple similar historical case information, fault pattern information, and assessment requirement information, executable decision information that fits the actual business is output to form a full-process technical link from data processing to intelligent decision-making. By constructing a reusable knowledge base through structured and vectorized processing of multi-source data, and combining it with the retrieval and reasoning capabilities of AI agents, the system achieves automatic prediction of key inspection points of rolling mill equipment, rapid analysis of fault causes, automatic generation of handling measures, and automatic matching of assessment requirements. This achieves full-process automation and intelligence from inspection to assessment, solving the technical problem of low efficiency in intelligent management of rolling mill equipment and realizing the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0044] The method described in this embodiment will be further described below.

[0045] As an optional embodiment, the method of collecting historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment includes: collecting original fault records, original fault analysis reports, and original assessment mechanism information of the rolling mill equipment; in response to the fact that the generation time of the original fault record is greater than or equal to a preset generation time, the generation time of the original fault analysis report is greater than or equal to a preset generation time, and the generation time of the original assessment mechanism information is greater than or equal to a preset generation time, the original fault record, the original fault analysis report, and the original assessment mechanism information are respectively identified as historical fault records, historical fault analysis reports, and historical assessment mechanism information.

[0046] In this embodiment, original fault records, original fault analysis reports, and original assessment mechanism information of the rolling mill equipment can be collected. The generation time of the original fault records is then compared with a preset generation time. If the generation time of the original fault records, the generation time of the original fault analysis reports, and the generation time of the original assessment mechanism information are all greater than or equal to the preset generation time, then the original fault records are identified as historical fault records, the original fault analysis reports are identified as historical fault analysis reports, and the original assessment mechanism information is identified as historical assessment mechanism information. The preset generation time can be a manually set time, such as 5 years.

[0047] For example, if the control system is set to a preset generation time of "5 years prior to the current time", then the control system will collect original fault records, original fault analysis reports and original assessment mechanism information related to the rolling mill equipment. When the generation time of the original fault records, original fault analysis reports and original assessment mechanism information all meet the requirement of 5 years or more, they can be identified as historical fault records, historical fault analysis reports and historical assessment mechanism information for building the knowledge base, so as to effectively filter out outdated and invalid data and ensure the timeliness and reference value of the data entering the database.

[0048] Furthermore, it can collect historical fault records of rolling mill equipment over many years (more than 5 years), including fault occurrence time, corresponding specialty, fault phenomenon, simple fault analysis, fault report, major accident record, and corresponding assessment mechanism, so as to carry out unified structured processing and form standardized fields.

[0049] As an optional implementation method, a target fault knowledge base is constructed based on the target field information, including: constructing an initial fault knowledge base based on the target field information; obtaining the current fault records, current fault analysis reports, and current assessment mechanism information of the rolling mill equipment; and updating the initial fault knowledge base based on the current fault records, current fault analysis reports, and current assessment mechanism information to obtain the target fault knowledge base.

[0050] In this embodiment, an initial fault knowledge base is constructed based on the obtained target field information. At the same time, the current fault records, current fault analysis reports, current handling measures, and current assessment mechanism information of the rolling mill equipment are obtained. Based on the current fault records, current fault analysis reports, current handling measures, and current assessment mechanism information, the initial fault knowledge base can be updated to obtain the target fault knowledge base.

[0051] Optionally, the current fault record can be referred to as a newly occurred fault. The current fault analysis report can be referred to as a new fault analysis report. The current assessment mechanism information can be referred to as new assessment requirements. The current handling measures can be referred to as new handling measures.

[0052] For example, new faults, new fault analysis reports, new handling measures, and new assessment requirements are added to the initial fault knowledge base in real time or periodically to continuously optimize the accuracy of prediction and analysis, forming a closed loop for the management of rolling mill equipment.

[0053] As an optional implementation method, a target vector database is constructed based on the target fault knowledge base, including: acquiring unstructured data information of the rolling mill equipment and structured data information from the target fault knowledge base; performing vectorization processing on the unstructured data information and the structured data information respectively to obtain vectorized unstructured data information and structured data information; and constructing the target vector database based on the vectorized unstructured data information and structured data information.

[0054] In this embodiment, unstructured data information of the rolling mill equipment and structured data information from the target fault knowledge base are obtained. Then, the unstructured data information and the structured data information are vectorized respectively to obtain vectorized unstructured data information and structured data information. Based on the above vectorized unstructured data information and structured data information, the target vector database is constructed.

[0055] Optionally, unstructured information is used to represent data that lacks a fixed format and is difficult to logically express using a two-dimensional database table structure. It typically exists in the form of text paragraphs or documents and can be simply referred to as unstructured reports. Structured information is used to represent data that, after preprocessing and structuring, has clearly defined fields and can be stored in database tables; it can be simply referred to as structured data.

[0056] For example, the process first acquires unstructured data, such as a portable document-formatted fault analysis report (PDF fault analysis report) containing a detailed description of "rolling mill vibration anomaly." Simultaneously, it extracts structured data from the target fault knowledge base, such as a record containing standard fields like "Equipment Location: Gearbox," "Fault Phenomenon: Vibration," and "Handling Measures: Tighten Bolts." A text embedding model is then used to vectorize this information. For unstructured data, the model transforms the full-text semantics into a high-dimensional vector; for structured records, the model converts their field content into corresponding feature vectors. Finally, these vectorized data are uniformly stored in the target vector database.

[0057] Furthermore, when a user asks "What should I do about gearbox vibration?", the question can be transformed into a vector, which can then be retrieved from the database and its cosine similarity calculated. At this point, the database can match structured fault records using vectors and also retrieve paragraphs from unstructured reports analyzing the causes of vibration, thus achieving unified semantic retrieval of multi-source heterogeneous data.

[0058] As an optional implementation method, a target vector database is constructed based on vectorized unstructured data information and structured data information, including: obtaining business logic management information of rolling mill equipment; constructing a knowledge association system based on vectorized unstructured data information and structured data information, combined with business logic management information; and constructing the target vector database according to the knowledge association system.

[0059] In this embodiment, the business logic management information of the rolling mill equipment is first obtained. Then, based on the vectorized unstructured data information and structured data information, combined with the business logic management information, a knowledge association system is constructed. Finally, the target vector database can be constructed according to the knowledge association system.

[0060] Optionally, business logic management information is used to characterize preset judgment rules and processing standards, that is, to define the logical conditions for data filtering and flow. Business logic management information includes: physical level logic information of rolling mill equipment, fault lifecycle handling logic information, equipment management element matching logic information, time pattern correlation logic information, and business scenario adaptation logic information.

[0061] Optionally, standardized business tags are bound to all vector data according to several dimensions: physical level logic information of rolling mill equipment, fault lifecycle handling logic information, equipment management element matching logic information, time pattern association logic information, and business scenario adaptation logic information, to establish a basic mapping of "vector data - business logic". At the same time, differentiated similarity weights are set for the tags of different dimensions obtained above (such as equipment part 80%, fault phenomenon 75%, professional category 70%, time pattern 30%, assessment clause 25%, etc.). During retrieval, the control system prioritizes matching business dimensions with high weights to ensure that the results are consistent with the actual needs of on-site operation (such as searching for "roll bearing vibration", prioritizing matching historical cases of the same part and phenomenon, rather than irrelevant data with high algorithm similarity). Finally, based on the above differentiated similarity weights, unstructured information and structured information, a knowledge association system is constructed to ensure that the association retrieval of vector data is consistent with the actual operation of the rolling mill.

[0062] For example, after obtaining structured data and unstructured reports, the structured data and unstructured reports are vectorized. Based on business logic management information, a knowledge association system between the vector database and knowledge is established to obtain the target vector database.

[0063] As an optional embodiment, the method further includes: target field information including at least: equipment area, equipment location, fault category, fault time, fault phenomenon, root cause analysis information, handling measures information, rectification result information, assessment clause information, responsible party, and fault handling status; intent identification type including at least: inspection prediction, fault analysis, measure formulation, training and learning, and management reminder; execution result including at least: inspection location of rolling mill equipment, fault causes and similar cases of rolling mill equipment, handling measures of rolling mill equipment, preventive measures of rolling mill equipment, assessment and management information of rolling mill equipment, responsible party, completion time limit of rolling mill equipment inspection task, and verification method of rolling mill equipment inspection task.

[0064] In this embodiment, the target field information includes at least: equipment area, equipment location, fault category, fault time, fault phenomenon, root cause analysis information, handling measures information, rectification result information, assessment clause information, responsible party, and fault handling status. Here, the fault category can be simply referred to as the category. The responsible party can be referred to as the responsible person. The root cause analysis information can be simply referred to as the root cause analysis. The handling measures information can be simply referred to as the handling measures. The rectification result information can be referred to as the rectification result.

[0065] Optionally, the equipment area is used to characterize the production operation area to which the fault or inspection of the rolling mill equipment belongs, such as the rolling mill area, roughing mill area, finishing mill area, heating furnace area, etc. The equipment part is used to characterize the specific location of the rolling mill equipment under the equipment area where the fault or inspection needs to be performed, such as the roll bearing, main motor, hydraulic pipeline, reducer, cooling system, etc.

[0066] Optionally, the fault category is used to characterize the technical specialty to which the fault belongs, such as maintenance, mechanical, or hydraulic specialties. Fault time is used to characterize the specific time and periodic characteristics of the fault occurrence, such as day shifts or high summer temperatures. Fault phenomenon is used to characterize the abnormal external characteristics exhibited by the rolling mill equipment, such as abnormal vibration, high temperature, abnormal noise, leakage, insufficient pressure, or shutdown. Root cause analysis information is used to characterize the most fundamental and original cause of the fault, such as bearing wear, insufficient lubrication, design defects, or inadequate maintenance. Handling measures information is used to characterize the specific methods taken to address the fault, such as checking bearing clearance, tightening bolts, replacing seals, cleaning filters, and improving lubrication.

[0067] Optionally, the rectification result information is used to characterize the final effect and verification conclusion after fault handling, such as repaired, operating normally, or requiring continued observation. The assessment clause information is used to characterize the management assessment criteria corresponding to the current fault or inspection, such as assessment for missed inspections, unplanned downtime, or overdue rectification. The responsible party is used to characterize the personnel or team responsible for the equipment, the current fault, or the inspection task, such as the rolling mill inspector, hydraulic maintenance team, equipment manager, or shift leader. The fault handling status is used to characterize the current progress of the fault from occurrence to completion, such as reported, in progress, awaiting review, closed loop, or overdue.

[0068] Optionally, the intent recognition types may include at least: inspection prediction, fault analysis, action formulation, training and learning, and management reminders. For example, when receiving natural language questions from users such as inspectors, equipment managers, and maintenance personnel, such as: "What are the key points for this week's inspection of the rolling mill area?"; "Possible causes of this fault and corresponding solutions?"; "What typical faults do new employees need to master?", intent recognition can categorize these into intent recognition types such as inspection prediction, fault analysis, action formulation, training and learning, and management reminders.

[0069] Optionally, the execution results may include at least: the inspection locations of the rolling mill equipment, the causes of rolling mill equipment failures and similar cases, the handling measures for the rolling mill equipment, the preventive measures for the rolling mill equipment, the assessment and management information of the rolling mill equipment, the responsible parties, the completion deadlines for the rolling mill equipment inspection tasks, and the verification methods for the rolling mill equipment inspection tasks. Among these, the inspection locations may be referred to as key inspection areas. The causes of failures may be referred to as possible causes of failures. Similar cases may be referred to as historical similar cases. The handling measures may be referred to as standard handling measures. The assessment and management information may be referred to as assessment clauses and management requirements. The completion deadlines for the inspection tasks may be simply referred to as completion deadlines. The verification methods for the inspection tasks may be simply referred to as verification methods.

[0070] For example, based on user questions, multiple similar historical cases, concurrent fault patterns, corresponding professional points, and corresponding assessment requirements are retrieved from the target vector database and target fault knowledge base to form supporting evidence. Then, the AI ​​agent is used to combine the retrieved historical data to output the key inspection areas and their basis for this week or period; possible causes of faults, similar historical cases; standard handling measures, preventive measures; corresponding assessment clauses and management requirements; responsible persons, completion deadlines, and verification methods.

[0071] In this embodiment, historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment are first collected and preprocessed and structured to construct a target fault knowledge base and a target vector database. Then, natural language queries from users are received, and the intent recognition type corresponding to the business scenario is identified. Searches are performed in the aforementioned target fault knowledge base and target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information. Finally, through a target inspection intelligent agent, combined with the aforementioned multiple similar historical case information, fault pattern information, and assessment requirement information, executable decision information that fits the actual business is output, forming a complete technical link from data processing to intelligent decision-making. By considering the structured and vectorized processing of multi-source data to construct a reusable knowledge base, and combining it with the retrieval and reasoning capabilities of the AI ​​intelligent agent, automatic prediction of key inspection points of the rolling mill equipment, rapid analysis of fault causes, automatic generation of handling measures, and automatic matching of assessment requirements are achieved. This achieves the goal of full-process automation and intelligence from inspection to assessment, solving the technical problem of low efficiency in intelligent management of rolling mill equipment and realizing the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0072] Furthermore, as a refinement and extension of the specific implementation methods of the above embodiments, in order to fully explain the specific implementation process of this embodiment, Figure 2 This application provides a flowchart illustrating another method for inspection and management of rolling mill equipment based on a fault knowledge base, as illustrated in this embodiment. Figure 2 As shown, the method includes: Step S201: Knowledge base construction and data structuring.

[0073] In this embodiment, historical fault records of rolling mill equipment over many years (more than 5 years), including fault occurrence time, corresponding specialty, fault phenomenon, simple fault analysis, fault report, major accident records, and corresponding assessment mechanisms, can be collected. This data will then be processed in a unified structure to form standardized fields: equipment area, equipment location, specialty category, fault time, fault phenomenon, root cause analysis, handling measures, rectification results, assessment clauses, responsible person, and fault handling status. The structured data and unstructured reports will be vectorized to establish a vector database and knowledge association system.

[0074] Step S202: User question and intent recognition.

[0075] In this embodiment, when natural language questions are received from users such as inspectors, equipment managers, and maintenance personnel, such as: "What are the key points for this week's inspection of the rolling mill area?"; "Possible causes of this fault and corresponding solutions?"; "What typical faults do new employees need to master?", intent recognition can categorize these into intent recognition types such as inspection prediction, fault analysis, solution formulation, training and learning, and management reminders.

[0076] Step S203: Retrieval and reasoning based on historical data.

[0077] In this embodiment, based on the user's question, multiple similar historical cases, concurrent fault patterns, corresponding professional points, and corresponding assessment requirements are retrieved from the target vector database and the target fault knowledge base to form supporting evidence.

[0078] Step S204: Generate an executable result with historical data.

[0079] In this embodiment, the AI ​​agent is used in conjunction with retrieved historical data to output the key inspection areas and their basis for this week or period; possible causes of failures and similar historical cases; standard handling measures and preventive measures; corresponding assessment clauses and management requirements; responsible persons, completion deadlines, and verification methods.

[0080] Step S205: Closed-loop management and knowledge base update.

[0081] In this embodiment, newly occurring faults, new fault analysis reports, new handling measures, and new assessment requirements are added to the initial fault knowledge base in real time or periodically to continuously optimize the accuracy of prediction and analysis, forming a closed loop for the management of rolling mill equipment.

[0082] For example, when a user asks: "Based on historical data, what should I focus on during this week's inspection of the rolling mill area?" the intent recognition type is: Predictive inspection focus; retrieve fault data from the same period and shift over the past 3-5 years; output this week's key inspection points for the rolling mill area: Roll bearings: focus on checking temperature, vibration, and lubrication (based on the highest frequency of faults in this area historically); Main motor and cooling system: check fans, filters, and temperature (based on the correlation between this and many historical high-temperature faults); Hydraulic system pipelines and joints: check for leaks (based on historical leaks leading to machine shutdowns); Gearbox sound, oil level, and oil temperature (based on the high frequency of abnormal noise faults historically). Responsible person: Rolling mill area inspector; Execution method: record each item at each location and compare with historical data.

[0083] Furthermore, when a user asks: "Abnormal vibration is occurring in the rolling mill, what are the possible causes and solutions?", the intent is identified as: fault analysis and solution development; historical similar vibration fault cases, corresponding professional analysis, handling solutions, and assessment requirements are retrieved; possible causes are output: bearing wear or excessive clearance (the highest percentage of similar historical faults); loose anchor bolts; misalignment of the coupling. Recommended measures: immediately check vibration values ​​and temperature, comparing them to historical thresholds; check bolt tightness; check bearings and lubrication; if necessary, stop the machine and open the cover for inspection; simultaneously, match the corresponding equipment management assessment requirements to form a standardized handling procedure.

[0084] For another example, when a user asks: What key points should a new person master when taking over rolling mill inspection? The solution should include: the locations and symptoms of high-frequency faults over the past three years; key points for diagnosing typical faults; inspection standards and key inspection areas; common problem handling procedures; and relevant assessment mechanisms and management requirements. This will enable rapid onboarding and reduce the risk of personnel replacement.

[0085] In this embodiment, a reusable knowledge base is constructed by structuring and vectorizing multi-source data. Combined with the retrieval and reasoning capabilities of AI agents, the automatic prediction of key inspection points of rolling mill equipment, rapid analysis of fault causes, automatic generation of handling measures, and automatic matching of assessment requirements are realized. This achieves the goal of full-process automation and intelligence from inspection to assessment, solves the technical problem of low efficiency in intelligent management of rolling mill equipment, and realizes the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0086] Furthermore, as Figure 1 The specific implementation of the method, Figure 3 This illustration shows a structural diagram of a rolling mill equipment inspection and management system based on a fault knowledge base, as provided in an embodiment of this application. Figure 3As shown, the rolling mill equipment inspection and management system 300 based on the fault knowledge base includes: a fault knowledge base module 301, a data structuring and vectorization module 302, an intent recognition and question answering module 303, and a retrieval enhancement generation module 304.

[0087] The fault knowledge base module 301 is used to collect historical fault records, historical fault analysis reports and historical assessment mechanism information of rolling mill equipment.

[0088] The data structuring and vectorization module 302 is used to preprocess and structure historical fault records, historical fault analysis reports and historical assessment mechanism information to obtain target field information of rolling mill equipment; based on the target field information, a target fault knowledge base is constructed, and based on the target fault knowledge base, a target vector database is constructed.

[0089] The intent recognition and question answering module 303 is used to receive natural language questions from the target object and perform intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions. The intent recognition type is used to represent the business scenario corresponding to the natural language questions. The target inspection agent of the rolling mill equipment is combined with multiple similar historical case information, fault pattern information and assessment requirement information to output the execution result. The execution result is used to represent the executable decision information corresponding to the natural language questions.

[0090] The retrieval enhancement generation module 304 is used to search the target fault knowledge base and target vector database according to the intent recognition type to obtain information on multiple similar historical cases, fault patterns, and assessment requirements.

[0091] In this embodiment, the fault knowledge base module is also used to store and manage years of structured fault data, fault reports, analysis conclusions, and assessment mechanisms; the data structuring and vectorization module is also used to clean, annotate, and vectorize the original documents; the intent recognition and question answering module is also used to understand user questions and match them with scenarios such as inspection, faults, measures, assessments, and training; and the retrieval enhancement generation (RAG) module is also used to retrieve relevant cases from the historical database and generate evidence-based answers.

[0092] Furthermore, in Figure 3 On this basis, Figure 4 This paper illustrates a schematic diagram of another rolling mill equipment inspection and management system based on a fault knowledge base, as provided in an embodiment of this application. Figure 4 As shown, the system also includes: an early warning and closed-loop management module 305, an interface module 306, an AI agent 307, and a user interface 308.

[0093] The early warning and closed-loop management module 305 is used to remind users of overdue inspections, incomplete rectification, and high-fault areas.

[0094] Interface module 306 is used to interface with existing equipment management systems and inspection systems.

[0095] AI agent 307 is used to output specific execution results (such as fault handling solutions, inspection suggestions, etc.).

[0096] User interface 308: A visual window or channel used to represent the interaction between the system and the user.

[0097] In this embodiment, a fault knowledge base module collects historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment; a data structuring and vectorization module preprocesses and structures the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain target field information of the rolling mill equipment; based on the target field information, a target fault knowledge base is constructed, and based on the target fault knowledge base, a target vector database is constructed; an intent recognition and question-answering module receives natural language questions from the target object, performs intent recognition on the natural language questions, and obtains the intent recognition type of the natural language questions, wherein the intent recognition... The different types are used to represent the business scenarios corresponding to natural language questions; the target inspection agent of the rolling mill equipment is combined with multiple similar historical case information, fault pattern information and assessment requirement information to output the execution result, which is used to represent the executable decision information corresponding to the natural language question; through the retrieval enhancement generation module, according to the intent recognition type, the target fault knowledge base and target vector database are searched to obtain multiple similar historical case information, fault pattern information and assessment requirement information, thereby solving the technical problem of low efficiency in intelligent management of rolling mill equipment and realizing the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0098] Furthermore, as Figure 1 The specific implementation of the method, Figure 5 This illustration shows a structural diagram of a rolling mill equipment inspection and management device based on a fault knowledge base, as provided in an embodiment of this application. Figure 5 As shown, the rolling mill equipment inspection and management device 500 based on the fault knowledge base includes: a data acquisition unit 501, a data construction unit 502, a first acquisition unit 503, a second acquisition unit 504, and an output unit 505.

[0099] The acquisition unit 501 is used to acquire historical fault records, historical fault analysis reports and historical assessment mechanism information of the rolling mill equipment, and to preprocess and structure the historical fault records, historical fault analysis reports and historical assessment mechanism information to obtain the target field information of the rolling mill equipment.

[0100] Construction unit 502 is used to construct a target fault knowledge base based on target field information, and to construct a target vector database based on the target fault knowledge base.

[0101] The first acquisition unit 503 is used to receive natural language questions from the target object, and to perform intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions. The intent recognition type is used to characterize the business scenario corresponding to the natural language questions.

[0102] The second acquisition unit 504 is used to search the target fault knowledge base and target vector database according to the intent recognition type to obtain information on multiple similar historical cases, fault patterns, and assessment requirements.

[0103] The output unit 505 is used to call a pre-trained agent to analyze multiple similar historical case information, fault pattern information and assessment requirement information, so as to output the execution result. The execution result is used to represent the executable decision information corresponding to the natural language question statement.

[0104] Optionally, the device is also used to: collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, including: collecting original fault records, original fault analysis reports, and original assessment mechanism information of the rolling mill equipment; in response to the generation time of the original fault record being greater than or equal to a preset generation time, the generation time of the original fault analysis report being greater than or equal to a preset generation time, and the generation time of the original assessment mechanism information being greater than or equal to a preset generation time, the original fault record, the original fault analysis report, and the original assessment mechanism information are respectively identified as historical fault records, historical fault analysis reports, and historical assessment mechanism information.

[0105] Optionally, the device is also used to: construct a target fault knowledge base based on target field information, including: constructing an initial fault knowledge base based on target field information; obtaining current fault records, current fault analysis reports, and current assessment mechanism information of the rolling mill equipment, and updating the initial fault knowledge base based on the current fault records, current fault analysis reports, and current assessment mechanism information to obtain the target fault knowledge base.

[0106] Optionally, the device is also used to: construct a target vector database based on a target fault knowledge base, including: acquiring unstructured data information of the rolling mill equipment and structured data information in the target fault knowledge base; performing vectorization processing on the unstructured data information and the structured data information respectively to obtain vectorized unstructured data information and structured data information; and constructing a target vector database based on the vectorized unstructured data information and structured data information.

[0107] Optionally, the device is also used to: construct a target vector database based on vectorized unstructured data information and structured data information, including: acquiring business logic management information of rolling mill equipment; constructing a knowledge association system based on vectorized unstructured data information and structured data information, combined with business logic management information; and constructing the target vector database according to the knowledge association system.

[0108] Optionally, the target field information shall include at least: equipment area, equipment location, fault category, fault time, fault phenomenon, root cause analysis information, handling measures information, rectification result information, assessment clause information, responsible party, and fault handling status; the intent identification type shall include at least: inspection prediction, fault analysis, measure formulation, training and learning, and management reminder; the execution result shall include at least: inspection location of rolling mill equipment, fault causes and similar cases of rolling mill equipment, handling measures of rolling mill equipment, preventive measures of rolling mill equipment, assessment and management information of rolling mill equipment, responsible party, completion time limit of rolling mill equipment inspection task, and verification method of rolling mill equipment inspection task.

[0109] In this embodiment, a data acquisition unit collects historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment. This data is then preprocessed and structured to obtain target field information for the rolling mill equipment. A construction unit builds a target fault knowledge base based on this target field information, and a target vector database is constructed based on the target fault knowledge base. A first acquisition unit receives natural language questions from the target object and performs intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions. The intent recognition type is used to represent the business scenario corresponding to the natural language questions. A second acquisition unit searches the target fault knowledge base and the target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information. An output unit calls a pre-trained intelligent agent to analyze the multiple similar historical case information, fault pattern information, and assessment requirement information to output execution results. These execution results represent executable decision information corresponding to the natural language questions, thereby solving the technical problem of low efficiency in the intelligent management of rolling mill equipment and achieving the technical effect of improving the efficiency of intelligent management of rolling mill equipment.

[0110] It should be noted that other corresponding descriptions of the functional units involved in the rolling mill equipment inspection and management device based on a fault knowledge base provided in this application embodiment can be found by referring to... Figure 1 The corresponding descriptions in the method will not be repeated here.

[0111] This application also provides a computer device, which may specifically be a personal computer, a server, a network device, etc. Figure 6 This application provides a schematic diagram of the device structure of a computer device according to an embodiment of the present application. Figure 6 As shown, the computer device includes a bus, a processor, memory, and a communication interface, and may also include an input / output interface and a display device. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores location information. The network interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0112] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0113] In one embodiment, a computer-readable storage medium is provided, which may be non-volatile or volatile, having stored thereon a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0114] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0115] It should be noted that the user personal information involved in the embodiments of this application is all authorized (with the knowledge and consent) by the relevant parties or fully authorized by all parties, and the executing entity can obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with the relevant laws and regulations of the relevant countries and regions, and do not violate public order and good morals. It should be noted that if any software tools or components other than those of this company appear in the embodiments of this application, they are merely illustrative examples and do not represent actual use.

[0116] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0117] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0118] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for inspection and management of rolling mill equipment based on a fault knowledge base, characterized in that, The method includes: Collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, and preprocess and structure the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain the target field information of the rolling mill equipment; Based on the target field information, a target fault knowledge base is constructed, and based on the target fault knowledge base, a target vector database is constructed. Receive a natural language question from a target object, and perform intent recognition on the natural language question to obtain the intent recognition type of the natural language question, wherein the intent recognition type is used to characterize the business scenario corresponding to the natural language question; According to the intent recognition type, a search is performed in the target fault knowledge base and the target vector database to obtain information on multiple similar historical cases, fault patterns, and assessment requirements. A pre-trained agent is invoked to analyze multiple similar historical case information, fault pattern information, and assessment requirement information to output execution results, wherein the execution results are used to characterize the executable decision information corresponding to the natural language question statement.

2. The method according to claim 1, characterized in that, Collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of rolling mill equipment, including: Collect the original fault records, original fault analysis reports, and original assessment mechanism information of the rolling mill equipment; In response to the fact that the generation time of the original fault record is greater than or equal to the preset generation time, the generation time of the original fault analysis report is greater than or equal to the preset generation time, and the generation time of the original assessment mechanism information is greater than or equal to the preset generation time, the original fault record, the original fault analysis report, and the original assessment mechanism information are respectively determined as the historical fault record, the historical fault analysis report, and the historical assessment mechanism information.

3. The method according to claim 1, characterized in that, Based on the target field information, a target fault knowledge base is constructed, including: Based on the target field information, an initial fault knowledge base is constructed; The current fault records, current fault analysis reports, and current assessment mechanism information of the rolling mill equipment are obtained, and the initial fault knowledge base is updated based on the current fault records, current fault analysis reports, and current assessment mechanism information to obtain the target fault knowledge base.

4. The method according to claim 1, characterized in that, Based on the aforementioned target fault knowledge base, a target vector database is constructed, including: Obtain unstructured data information of the rolling mill equipment, as well as structured data information from the target fault knowledge base; The unstructured data information and the structured data information are respectively vectorized to obtain the vectorized unstructured data information and the structured data information; The target vector database is constructed based on the vectorized unstructured data information and the structured data information.

5. The method according to claim 4, characterized in that, Based on the vectorized unstructured data information and the structured data information, the target vector database is constructed, including: Obtain the business logic management information of the rolling mill equipment; Based on the vectorized unstructured data information and the structured data information, combined with the business logic management information, a knowledge association system is constructed. The target vector database is constructed according to the aforementioned knowledge association system.

6. The method according to claim 1, characterized in that, The method further includes: The target field information includes at least: equipment area, equipment location, fault category, fault time, fault phenomenon, root cause analysis information, handling measures information, rectification result information, assessment clause information, responsible party, and fault handling status. The intent recognition types include at least: inspection prediction, fault analysis, measure formulation, training and learning, and management reminders; The execution results include at least: the inspection points of the rolling mill equipment, the causes of failures and similar cases of the rolling mill equipment, the handling measures of the rolling mill equipment, the preventive measures of the rolling mill equipment, the assessment and management information of the rolling mill equipment, the responsible parties, the completion time limit of the inspection tasks of the rolling mill equipment, and the verification method of the inspection tasks of the rolling mill equipment.

7. A rolling mill equipment inspection and management system based on a fault knowledge base, characterized in that, The system includes: The fault knowledge base module is used to collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of rolling mill equipment. The data structuring and vectorization module is used to preprocess and structure the historical fault records, the historical fault analysis reports, and the historical assessment mechanism information to obtain the target field information of the rolling mill equipment; based on the target field information, a target fault knowledge base is constructed, and based on the target fault knowledge base, a target vector database is constructed. The intent recognition and question answering module is used to receive natural language questions from a target object, and to perform intent recognition on the natural language questions to obtain the intent recognition type of the natural language questions. The intent recognition type is used to represent the business scenario corresponding to the natural language questions. The module also calls a pre-trained agent to analyze multiple similar historical case information, fault pattern information, and assessment requirement information to output execution results. The execution results are used to represent the executable decision information corresponding to the natural language questions. The retrieval enhancement generation module is used to search the target fault knowledge base and the target vector database according to the intent recognition type to obtain multiple similar historical case information, fault pattern information, and assessment requirement information.

8. A rolling mill equipment inspection and management device based on a fault knowledge base, characterized in that, The device includes: The data acquisition unit is used to collect historical fault records, historical fault analysis reports, and historical assessment mechanism information of the rolling mill equipment, and to preprocess and structure the historical fault records, historical fault analysis reports, and historical assessment mechanism information to obtain the target field information of the rolling mill equipment. A construction unit is used to construct a target fault knowledge base based on the target field information, and to construct a target vector database based on the target fault knowledge base; The first acquisition unit is used to receive a natural language question statement from a target object, and to perform intent recognition on the natural language question statement to obtain the intent recognition type of the natural language question statement, wherein the intent recognition type is used to characterize the business scenario corresponding to the natural language question statement. The second acquisition unit is used to search the target fault knowledge base and the target vector database according to the intent recognition type to obtain information on multiple similar historical cases, fault patterns, and assessment requirements. The output unit is used to call a pre-trained agent to analyze multiple similar historical case information, fault pattern information and assessment requirement information to output execution results, wherein the execution results are used to characterize executable decision information corresponding to the natural language question statement.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

Images