A method, system, and media for diagnosis-treatment co-generation in predictive maintenance of industrial equipment

Abstract

The application discloses a diagnosis-disposal joint generation method and system in predictive maintenance of industrial equipment. The method comprises the following steps: based on the fault position, fault nature and severity in the diagnosis conclusion, querying a preset diagnosis-disposal mapping library to obtain a basic disposal suggestion template matched with the fault position, fault nature and severity; differentially filling the basic disposal suggestion template according to the diagnosis confidence level to generate disposal suggestion content matched with the confidence level; and structurally associating and outputting the diagnosis conclusion and the disposal suggestion, so that each inspection action in the disposal suggestion refers to the corresponding evidence identifier in the diagnosis conclusion. Through the diagnosis-disposal joint generation mechanism, the application solves the technical problems that the diagnosis conclusion and the disposal suggestion are disjointed, the disposal suggestion is not specific and executable enough, and the suggestion is too general when the confidence level is low in the predictive maintenance scene of the industrial equipment.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for industrial equipment, specifically to a method, system, and medium for generating diagnosis-disposal linkage in predictive maintenance scenarios for industrial equipment, particularly applicable to predictive maintenance systems for industrial equipment based on an intelligent agent architecture, wherein the intelligent agent has the ability to generate diagnostic conclusions and disposal suggestions in a linkage manner. Background Technology

[0002] In industrial equipment monitoring scenarios, the health status diagnosis of rotating machinery ultimately needs to be translated into actionable recommendations to guide on-site operation and maintenance. The following specific technical problems exist in industrial settings that existing diagnostic systems cannot effectively address: (1) Disconnect between diagnostic conclusions and treatment recommendations. In traditional systems, the diagnostic module and the treatment recommendation module operate independently. After the diagnostic conclusion is output, it is processed by an independent recommendation generation module, and there is no structured connection between the two. When the diagnostic conclusion changes, the treatment recommendation cannot be automatically updated, resulting in a mismatch between the recommendation and the diagnosis.

[0003] (2) The handling suggestions are not specific and actionable enough. The handling suggestions of the existing system usually adopt general templates, such as "Please check the equipment status" or "Please pay attention to trend changes", without providing actionable guidance based on the specific fault location, nature and severity. After receiving the suggestions, the on-site maintenance personnel are still unclear about what to check, how to check, and to what extent.

[0004] (3) The suggestions are too general when the confidence level is low. When the diagnostic confidence level is low, the existing system usually only outputs vague suggestions such as "further investigation is recommended", without specifying what information needs to be supplemented, which professional personnel to contact, or what confirmatory tests to be performed, resulting in low efficiency in handling low confidence alarms.

[0005] (4) Lack of feedback and linkage on treatment effects. The existing system lacks an effect tracking mechanism after the treatment suggestions are output, and it is impossible to compare and verify the treatment results with the diagnostic conclusions. As a result, the accuracy of diagnosis cannot be continuously optimized, and the incorrect diagnosis-treatment mapping relationship exists for a long time.

[0006] (5) Severity is not fully utilized in response generation. Existing systems typically generate response recommendations based only on the location and nature of the fault, without incorporating the severity rating in the diagnostic conclusions into the response strategy. This results in similar faults of different severity levels receiving the same response recommendations, lacking differentiated treatment.

[0007] In summary, how to achieve a collaborative generation scheme in the predictive maintenance scenario of industrial equipment that combines structured correlation between diagnosis and treatment, differentiated confidence suggestions, severity linkage, and treatment effect feedback linkage is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0008] In view of this, the present invention provides a diagnostic-treatment linkage generation method and system for predictive maintenance of industrial equipment, in order to solve the technical problems existing in the prior art, such as the disconnect between diagnostic conclusions and treatment suggestions, the lack of specific and executable treatment suggestions, the overly general suggestions at low confidence levels, the lack of feedback linkage on treatment effects, and the underutilization of severity levels.

[0009] This invention provides a diagnostic-treatment linkage generation method for predictive maintenance of industrial equipment, applied to predictive maintenance scenarios for industrial equipment. The method comprises: The mapping query step involves querying a preset diagnosis-treatment mapping library based on the fault location, nature, and severity in the diagnostic conclusion to obtain a basic treatment suggestion template that matches the fault location, nature, and severity. The diagnosis-treatment mapping library is a dynamically configurable structured knowledge base that contains many-to-many mapping relationships between fault location, nature, severity, and treatment actions. The confidence level differentiation step involves differentiating the basic treatment suggestion template according to the diagnostic confidence level to generate treatment suggestion content that matches the diagnostic confidence level. This differentiation includes: when the confidence level is high, filling in specific examination sites, examination items, and expected treatment actions; when the confidence level is medium, filling in priority suggestions, including priority and secondary examination items; and when the confidence level is low, filling in investigation path suggestions, including the order of investigation of possible sources of the problem and distinguishing investigation actions. The structured association output step links the diagnostic conclusions and treatment recommendations in a structured manner, so that each examination action in the treatment recommendation references the corresponding evidence identifier in the diagnostic conclusion, forming a two-way traceable chain between diagnosis and treatment; the two-way traceable chain includes a forward index from the diagnostic evidence identifier to the treatment action identifier and a reverse index from the treatment action identifier to the diagnostic evidence identifier.

[0010] Optionally, the structure of the diagnosis-treatment mapping library includes dimensions of fault location, fault nature, severity, treatment action, and priority.

[0011] Optionally, the confidence level differential filling step further includes filling the expected processing time window, interference factor elimination suggestions, and professional contact suggestions according to the confidence level.

[0012] Optionally, the structured association output step includes structured encapsulation of diagnostic conclusions, structured encapsulation of treatment recommendations, and establishment of an association index. The bidirectional association index includes a forward index from diagnostic evidence identifiers to treatment action identifiers and a reverse index from treatment action identifiers to diagnostic evidence identifiers, supporting fast bidirectional traceability queries.

[0013] Optionally, it also includes a step for dynamically adjusting treatment recommendations, automatically updating treatment recommendations and generating change descriptions when the diagnostic conclusion changes.

[0014] Optionally, it also includes a feedback and linkage step for handling effect, which records the handling execution results, calculates the effect matching degree, and triggers diagnostic review or priority optimization.

[0015] Optionally, the diagnosis-treatment mapping library supports both manual review and dynamic maintenance.

[0016] Optionally, it also includes an urgency assessment step, which calculates a comprehensive urgency score based on multi-dimensional factors and outputs it in a graded manner.

[0017] Optionally, the output format is a three-segment output, including a monitoring display segment, a conclusion segment, and a recommendation segment.

[0018] Optionally, the treatment actions in the diagnostic-treatment mapping library may also include conditional treatment recommendations related to the equipment's operating conditions.

[0019] Furthermore, the present invention also provides a diagnostic-treatment linkage generation system for predictive maintenance of industrial equipment, characterized in that it includes: a mapping query module, a confidence degree differential filling module, and a structured association output module.

[0020] Furthermore, the present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the method described in any of the preceding claims.

[0021] The beneficial effects of this invention are as follows: By establishing a structured association between fault location, fault nature, severity, and handling actions through a diagnosis-handling mapping library, it solves the technical problem of the disconnect between diagnostic conclusions and handling recommendations. Through a confidence-differentiated filling mechanism, it achieves precise matching of handling recommendations at different confidence levels, solving the technical problems of insufficiently specific and executable handling recommendations and overly general recommendations at low confidence levels. By introducing a severity dimension, it achieves linked matching between handling recommendations and fault severity, enabling differentiated handling strategies for similar faults of different severity levels. Through a bidirectional traceable chain in the structured association output steps, it achieves complete association and traceability between diagnostic conclusions and handling recommendations. Through the handling effect feedback linkage steps, it establishes a continuous optimization closed loop for diagnostic accuracy. This invention provides core output generation support with "diagnosis-handling linkage-effect feedback" capabilities for predictive maintenance systems for industrial equipment based on an intelligent agent architecture, systematically solving the technical problem of the disconnect between diagnosis and handling in industrial equipment monitoring scenarios, with an overall effect greater than the sum of its parts. Attached Figure Description

[0022] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings: Figure 1 A flowchart of a diagnostic-treatment linkage generation method for predictive maintenance of industrial equipment according to Embodiment 1 of the present invention is shown; Figure 2 A schematic diagram of the diagnosis-treatment mapping library in Embodiment 1 of the present invention is shown; Figure 3 This diagram illustrates the confidence-differential filling strategy in Embodiment 1 of the present invention. Figure 4 This diagram illustrates the data association relationships in the structured association output of Embodiment 1 of the present invention. Figure 5 This illustrates a flowchart of the dynamic adjustment of treatment suggestions and the linkage of effect feedback in Embodiment 1 of the present invention; Figure 6 A schematic diagram of the three-segment output format in Embodiment 1 of the present invention is shown; Figure 7 The diagram shows a structural diagram of a diagnostic-treatment linkage generation system for predictive maintenance of industrial equipment according to Embodiment 2 of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] In this embodiment of the invention, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0026] In this embodiment of the invention, the term "multiple" refers to two or more, and other quantifiers are similar. Example 1

[0027] This embodiment provides a method for generating a diagnosis-treatment linkage in predictive maintenance of industrial equipment, such as... Figure 1 As shown, the method includes: S1: Mapping query step, based on the fault location, fault nature and severity in the diagnostic conclusion, query the preset diagnosis-treatment mapping library to obtain the basic treatment suggestion template; S2: Confidence level differentiation step, which differentiates the basic treatment suggestion template according to the diagnostic confidence level to generate treatment suggestion content; S3: The structured correlation output step outputs the diagnostic conclusions and treatment recommendations in a structured way, forming a two-way traceable chain of diagnosis and treatment.

[0028] In this embodiment, a structured association between faults and treatments is achieved through a diagnosis-treatment mapping library, accurate matching of suggestions is achieved through confidence-differential filling, and complete traceability is achieved through structured association output.

[0029] Each step of this embodiment will be described in detail below.

[0030] In this embodiment S1, the system queries a preset diagnosis-treatment mapping library based on the fault location, nature, and severity in the diagnostic conclusion.

[0031] The diagnosis-treatment mapping library is a dynamically configurable structured knowledge base. Its initial data can come from historical treatment data (mature scenarios), expert rules (new scenarios), or a combination of both. In specific implementations, the structured knowledge base can be a structured knowledge graph, a relational database, a vector database, or other organized data storage formats. The structure of the diagnosis-treatment mapping library includes five dimensions: The fault location dimension records the equipment component level at which a fault may occur, including system level (such as pump system, gearbox system), component level (such as bearing, coupling, impeller), and part level (such as bearing outer ring, rolling elements, cage). The fault location uses a hierarchical coding method, supporting step-by-step location from system level to part level.

[0032] The failure nature dimension records the type and stage of the failure, including types such as wear, fatigue, loosening, misalignment, imbalance, and cavitation, as well as stages such as early, middle, and late. Failure nature uses a type-stage combination coding, such as "wear-early" and "fatigue-middle".

[0033] The severity dimension records the severity level of the fault, typically using a 1-5 level scale. Level 1 indicates a minor anomaly, while level 5 indicates a severe fault requiring immediate shutdown. The severity rating is output by the upstream diagnostic module (such as a fault diagnosis module based on heterogeneous evidence cross-validation) and serves as one of the key inputs for the mapping query in this embodiment.

[0034] The action dimension records the set of executable actions corresponding to the combination of fault location, fault nature, and severity. Each action includes action type (inspection, maintenance, replacement, adjustment), execution object (specific component or location), execution standard (criteria for judging normal / abnormal), and expected effect (the indicator improvement target to be achieved after execution).

[0035] The priority dimension records the recommended priority of each response action under the same combination of fault location, fault nature, and severity. Priority is calculated based on a combination of historical success rate and on-site feasibility; the higher the historical success rate and the easier it is to execute on-site, the higher the priority of the response action.

[0036] During mapping queries, the system uses the fault location, fault nature, and severity as the joint query key to retrieve a set of matching treatment actions from the diagnosis-treatment mapping library. Because there is a many-to-many mapping relationship between fault location, fault nature, and severity, the system returns all matching combinations, sorted by priority.

[0037] For example, if the diagnosis is "early wear of the outer ring of the pump end bearing, severity level 2", the system query will retrieve the following set of actions: Priority 1: Check the bearing lubrication condition and confirm the amount and quality of lubricating oil; Priority 2: On-site auscultation confirms abnormal noise during bearing operation; Priority 3: Monitor the bearing temperature trend to confirm whether the temperature rise is within the normal range; Priority 4: Plan to disassemble and inspect the condition of the bearing outer ring during the next shutdown.

[0038] If the severity level is 4 (relatively severe), the priority ranking may be adjusted as follows: Priority 1: Plan to shut down the machine soon to check the condition of the bearing outer ring; Priority 2: Confirm abnormal bearing operating noise through on-site auscultation, paying particular attention to metallic friction sounds; Priority 3: Monitor bearing temperature trends and report any abnormal temperature rise immediately; Priority 4: Check the bearing lubrication condition to confirm whether there is insufficient lubrication leading to accelerated deterioration.

[0039] In this embodiment S2, the system differentiates the basic treatment suggestion template according to the diagnostic confidence level.

[0040] When the diagnostic confidence level is high, the system fills in the specific examination site, examination items, and expected treatment actions. The treatment recommendations include: Specific inspection areas: Clearly indicate the equipment components and locations that need to be inspected, such as "outer ring of the bearing at the pump coupling end"; Specific inspection items: List the specific inspection items that need to be confirmed, such as "lubrication status, abnormal noise, and temperature trend"; Expected action: Provide suggested measures, such as "add lubricating oil to the standard level" or "if the abnormal noise is obvious, plan to replace the bearing"; Estimated processing time window: The suggested processing time range, such as "It is recommended to complete the inspection within the next 2 weeks"; Recommended maintenance plan: Provide maintenance suggestions based on the equipment operation plan, such as "It is recommended to focus on inspection during the next planned shutdown".

[0041] When the diagnostic confidence level is medium, the system populates priority suggestions, including priority and secondary checks. The treatment suggestions include: Priority inspection items: List the most likely relevant inspection items, such as "It is recommended to prioritize checking the lubrication condition of the bearing outer ring"; Secondary inspection items: List the secondary and potentially related inspection items, such as "If the lubrication is normal, check the alignment of the coupling"; Supplementary testing recommendations: Confirmatory tests are recommended, such as "It is recommended to supplement the measurement of horizontal vibration to confirm the consistency of the trend"; Interference factors to be excluded: List factors that may affect the accuracy of the diagnosis, such as "Please confirm whether there have been any speed regulation operations or load changes recently".

[0042] This prioritization suggestion method not only provides clear directions for priority inspections, but also offers alternative investigation paths for on-site maintenance personnel. It avoids the problem of suggestions being too general or too specific when the confidence level is medium, and achieves a practical strategy of "prioritizing and investigating step by step".

[0043] When the diagnostic confidence level is low, the system populates suggested troubleshooting paths, including the order in which to investigate possible sources of the problem and the corresponding troubleshooting actions. The suggested actions include: Troubleshooting order: Arrange the sources of problems from highest to lowest probability, such as "The problem may originate from the bearing area or coupling misalignment. It is recommended to check in the following order:"; Differentiate and investigate actions: Provide specific actions to distinguish different possible sources, such as "1. First check the bearing lubrication status. If the lubrication is normal, then lubrication is not a factor; 2. On-site auscultation to confirm the location of the abnormal noise source. If the abnormal noise comes from the coupling area, then focus on checking the alignment; 3. Supplement with horizontal vibration measurements and compare the vertical data to confirm the consistency of the trend." Recommended professionals to contact: such as "It is recommended to contact a vibration analyst for manual review"; The types of information that need to be supplemented include: such as "Please provide the current operating speed, bearing model, and on-site operating conditions." Conservative monitoring recommendations: such as "It is recommended to increase the frequency of daily inspections of the equipment and closely monitor trend changes."

[0044] This troubleshooting path suggestion method completely solves the problem of user confusion caused by vague suggestions such as "further investigation is recommended" when dealing with low confidence levels. By providing a clear troubleshooting sequence and distinguishing actions, even with low diagnostic confidence, on-site maintenance personnel can obtain specific and actionable troubleshooting guidance, knowing "what to check first, what to check next, and how to distinguish," significantly improving the processing efficiency of low confidence alarms and user satisfaction. This is one of the key innovations that distinguishes this invention from existing technologies.

[0045] The differentiated input mechanism ensures accurate matching of treatment recommendations at different confidence levels, avoiding the problem of overly general recommendations at high confidence levels or overly vague recommendations at low confidence levels. In particular, the investigation path recommendation mechanism at low confidence levels upgrades the traditional binary choice of "whether to give recommendations" to a fine-grained control of "how to give recommendations," maintaining both the prudence of diagnosis and ensuring the practical value of the product.

[0046] In this embodiment S3, the system outputs a structured association between the diagnostic conclusion and the treatment suggestion.

[0047] The structured encapsulation of diagnostic conclusions includes: fault location, fault nature, severity, confidence level, list of key evidence, and analysis process. Each piece of evidence in the list of key evidence is accompanied by a unique evidence identifier, such as "EVD-001" or "EVD-002".

[0048] The structured encapsulation of treatment recommendations includes: a set of treatment actions, action priorities, execution criteria, and referenced evidence identifiers. Each treatment recommendation explicitly references the corresponding evidence identifier from the diagnostic conclusion, forming a one-way traceability chain of "treatment recommendation → evidence".

[0049] During the association index establishment phase, the system establishes a bidirectional association index between the diagnostic conclusion data package and the treatment suggestion data package. This bidirectional association index includes a forward index from diagnostic evidence identifiers to treatment action identifiers and a reverse index from treatment action identifiers to diagnostic evidence identifiers. Each piece of evidence in the diagnostic conclusion can be queried to all treatment suggestions referencing that evidence; each treatment suggestion can be traced back to all diagnostic evidence upon which it is based. The bidirectional association index supports rapid retrieval and audit traceability.

[0050] In addition, this embodiment also includes a step for dynamically adjusting the treatment recommendations: When the diagnostic conclusion changes during the status transition, for example, from "early wear of bearing outer ring, severity level 2" to "early wear of bearing inner ring, severity level 2", the system re-queries the diagnosis-treatment mapping library based on the changed diagnostic conclusion to generate updated treatment suggestions.

[0051] The system compares the handling suggestions before and after the update, identifying any added, deleted, or modified actions. It generates a handling change description and pushes it to the user interface, explaining the reason for and content of the change. For example: "The diagnostic conclusion has been updated to early wear of the bearing inner ring, and the handling suggestions have been adjusted accordingly: the original suggestion 'check the lubrication status of the bearing outer ring' has been updated to 'check the fit clearance of the bearing inner ring,' and a new suggestion 'monitor the temperature trend of the inner ring' has been added." The system records the change history of handling recommendations, including the reason for the change, the time of the change, the content of the change, and the operator's identification, for subsequent auditing and experience accumulation.

[0052] In addition, this embodiment also includes a feedback linkage step for the treatment effect: When a user reports that a certain action has been taken, the system records the execution time, the personnel involved, and the result. The result includes the action described by the user and the data changes monitored by the system.

[0053] The system compares the treatment effect with the expected effect in the diagnostic conclusion and calculates the matching degree of the treatment effect. For example, if the diagnostic conclusion expects "the vibration velocity should be reduced to below 1.0 mm / s after adding lubricating oil", and the actual monitoring data shows that the vibration velocity drops to 0.8 mm / s, the matching degree is 100%; if the actual vibration velocity rises to 1.5 mm / s, the matching degree is 0%.

[0054] When the matching degree of the treatment effect is lower than a preset threshold, the diagnostic review process is triggered to reassess the accuracy of the diagnostic conclusion. The system marks the diagnostic conclusion as pending review and pushes a reminder to the diagnostic engineer, explaining the specific reasons why the treatment effect does not meet expectations.

[0055] When the matching degree of the treatment effect is higher than a preset threshold, the priority weight of the corresponding diagnosis-treatment mapping relationship is increased. The system records successful cases of this mapping relationship, increases its historical success rate statistics, and increases the recommended priority of this treatment action in subsequent similar diagnoses.

[0056] In addition, this embodiment also includes an urgency assessment step: The system calculates a comprehensive urgency score based on the importance level of the fault location, the severity of the fault, the severity rating in the diagnostic conclusion, and the confidence level. The importance level of the fault location is determined by the criticality of the equipment in the production process, with critical equipment receiving a higher score and auxiliary equipment receiving a lower score. The severity of the fault is determined by its impact on the safe operation of the equipment, with late-stage severe faults receiving a higher score and early-stage minor faults receiving a lower score. The severity rating is a direct input factor in the score calculation, with a higher score indicating a higher severity level. The confidence level serves as a correction factor; at high confidence levels, the urgency score is calculated as is, while at low confidence levels, the urgency score is appropriately reduced.

[0057] The urgency level of the response recommendations is classified according to the overall urgency score: Immediate Action: If the score exceeds the preset emergency threshold, immediate action is required, and the system will push an emergency notification to the responsible person and management. Planned maintenance: If the score is between the preset emergency threshold and the preset general threshold, it will be included in the recent maintenance plan, and the system will push a plan arrangement suggestion; Continuous monitoring: If the score is lower than the preset general threshold, no immediate action is required. The system will push monitoring suggestions and include them in the routine inspection.

[0058] The urgency level classification results are output as an additional attribute of the handling recommendations to guide users in making decisions on handling priorities.

[0059] Furthermore, the output format in this embodiment adopts a three-segment output: The first section is the monitoring display section, describing the current status and trend of the monitoring data. For example: "Monitoring shows that the vertical velocity at the coupling end of the P446-01A circulating water pump is slowly increasing, currently at 1.25 mm / s, exceeding the attention threshold of 1.0 mm / s. The 227 Hz peak in the spectrum matches the characteristic frequency of the bearing outer ring, indicating a significant exceedance of the high-frequency envelope energy." The second paragraph is the conclusion, which includes the location, nature, and severity of the fault. For example: "Conclusion: Early wear may exist on the outer ring of the pump coupling end bearing; insufficient lubrication leading to initial deterioration cannot be ruled out." The third paragraph is the recommendation section, containing action suggestions that match the confidence level. For example: "Recommendation: 1. Inspect the lubrication status of the pump coupling bearing on-site to confirm whether the lubricating oil quantity and quality are normal; 2. Listen to the bearing on-site to confirm whether there are any abnormal noises during operation; 3. It is recommended to complete the inspection within the next two weeks, and if the abnormal noise is obvious, plan to replace the bearing." The three segments are structurally linked through preset association markers, supporting segmented retrieval and independent consumption. Downstream systems can consume the conclusion segment separately for status management, the suggestion segment separately for work order generation, or consume the entire segment for user push notifications.

[0060] Furthermore, the diagnosis-treatment mapping library in this embodiment supports both manual review and dynamic maintenance: Administrators can view the historical usage count, success rate, and user satisfaction rating for each mapping relationship in the mapping review interface. It supports filtering and sorting by multiple dimensions, including fault location, fault nature, severity, and type of handling action.

[0061] Administrators can manually add, modify, or delete mapping relationships. When adding a mapping relationship, fields such as fault location, fault nature, severity, handling action, execution standard, and expected effect must be filled in, and it will take effect after manual confirmation. When modifying a mapping relationship, the system records a comparison before and after the modification, and supports version rollback.

[0062] The system periodically optimizes the priority ranking of mapping relationships based on feedback data regarding treatment effectiveness. The automatic optimization algorithm comprehensively considers historical success rates, treatment effectiveness matching degrees, and user satisfaction to dynamically adjust the ranking of mapping relationships. Administrators can set the enabling / disabling status of automatic optimization and the adjustment range limits.

[0063] Furthermore, in this embodiment, the treatment actions in the diagnosis-treatment mapping library also include conditional treatment suggestions related to the equipment's operating conditions. When the equipment is in a specific operating condition, the corresponding conditional treatment suggestion is automatically triggered.

[0064] For example, when the diagnosis is "pump-end cavitation" and the equipment is currently operating under "high load," the system automatically triggers a conditional handling suggestion: "The equipment is currently operating under high load, increasing the risk of cavitation. It is recommended to appropriately reduce the load or adjust the inlet pressure, and at the same time check whether there is any blockage in the inlet pipeline." The triggering conditions for the conditional handling suggestion include operating parameter ranges such as load range, speed range, and medium temperature range.

[0065] The complete diagnosis-treatment linkage generation method described in this embodiment has the following beneficial effects: (1) Solved the technical problem of the disconnect between diagnosis conclusions and treatment recommendations: A structured association between fault location, fault nature, severity and treatment actions was established through the diagnosis-treatment mapping library. The complete traceability of diagnosis and treatment was realized through bidirectional association index. When the diagnosis conclusion changes, the treatment recommendations are automatically updated synchronously.

[0066] (2) Solved the technical problem that the disposal suggestions were not specific and executable enough: Through the disposal action dimension in the diagnosis-disposal mapping library, each disposal action includes the specific execution object, execution standard and expected effect. After receiving the suggestions, the on-site operation and maintenance personnel clearly know what to check, how to check and to what extent.

[0067] (3) It solves the technical problem of overly general suggestions when the confidence level is low: Through the confidence level differential filling mechanism, especially the investigation path suggestion mechanism when the confidence level is low, it not only gives the prompt of "suggest further investigation", but also specifies the investigation order, distinguishing actions and supplementary detection items. It upgrades "whether to give suggestions" to the fine control of "how to give suggestions", which significantly improves the processing efficiency of low confidence level alarms and user satisfaction. This "confidence level-based dynamic suggestion granularity control" solution is technically more refined and creative than the traditional binary choice of "give or not give", and at the same time perfectly solves the pain point of "users need guidance" in the product field.

[0068] (4) Solved the technical problem of underutilization of severity: By introducing the severity dimension as one of the joint keys of the mapping query, differentiated handling strategies are obtained for similar faults with different severity levels. For example, for the same "bearing outer ring wear", the early stage (level 2) focuses on observation and planned maintenance, while the late stage (level 4) focuses on stopping the machine for inspection as soon as possible, thus achieving a precise match between the handling strategy and the severity of the fault.

[0069] (5) Solved the technical problem of lack of feedback linkage of treatment effect: Through the treatment effect feedback linkage step, a comparison and verification mechanism between treatment execution results and diagnosis conclusions was established. Incorrect diagnosis-treatment mapping relationship can be identified and corrected in a timely manner, and correct mapping relationship can be continuously strengthened, realizing a continuous optimization closed loop of diagnostic accuracy.

[0070] (6) Achieving a complete closed loop of diagnosis-treatment-effect: This invention realizes the transformation from diagnosis to treatment through a diagnosis-treatment mapping library, achieves accurate matching of suggestions through confidence-differentiation filling, achieves strategy differentiation through severity linkage, achieves complete traceability through structured correlation output, and achieves continuous optimization of effects through treatment effect feedback. This closed-loop capability of "diagnosis-treatment-feedback-optimization" makes the output behavior of the intelligent agent highly anthropomorphic—the more accurate the diagnosis, the more precise the suggestion, and the better the effect, fundamentally different from the one-way output mode of traditional diagnostic systems. In particular, the investigation path suggestion mechanism under low confidence conditions enables the intelligent agent to provide valuable action guidance to users even in uncertain situations, reflecting the product philosophy of "conservative but not giving up", and perfectly balancing technical prudence and product practicality. Example 2

[0071] A diagnostic-treatment linkage generation system for predictive maintenance of industrial equipment, such as Figure 7 As shown, the system includes a mapping query module, a confidence-based differential filling module, and a structured association output module: The mapping query module is used to query the diagnosis-treatment mapping library based on the fault location, fault nature and severity in the diagnosis conclusion, and obtain a basic treatment suggestion template. The confidence level differentiation filling module is used to differentiate the basic treatment suggestion template according to the diagnostic confidence level and generate treatment suggestion content; The structured correlation output module is used to output the diagnostic conclusions and treatment suggestions in a structured way, forming a two-way traceable chain of diagnosis and treatment.

[0072] In this embodiment, a mapping query module achieves precise matching between faults and treatments, a confidence level differentiation filling module enables adaptive confidence levels for recommendations, and a structured association output module ensures complete traceability. The system also includes interfaces with the diagnostic engine, device archives, user interaction, and effect feedback, forming a complete closed-loop diagnostic-treatment process.

[0073] The functions of each part in the diagnosis-treatment linkage generation system described in this embodiment are the same as the methods and steps described in Embodiment 1. Therefore, for details not covered in this embodiment, please refer to Embodiment 1 and... Figures 1 to 6 The specific details will not be elaborated here. Example 3

[0074] This embodiment also provides a diagnostic-treatment linkage generation device for predictive maintenance of industrial equipment, including a processor, a memory, and a computer program stored in the memory and run on the processor. When the computer program runs, it implements the method steps in the diagnostic-treatment linkage generation method of Embodiment 1.

[0075] By using a computer program stored in memory and running on a processor, it is possible to perform mapping queries, confidence level differential filling, and structured correlation output, effectively solving the technical problem of the disconnect between diagnosis and treatment in industrial equipment monitoring scenarios.

[0076] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the computer device, connecting all parts of the computer device through various interfaces and lines.

[0077] Memory can be used to store computer programs and / or models. The processor performs various functions of the computer device by running or executing the computer programs and / or models stored in the memory, and by accessing data stored in the memory. Memory can primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on usage. Furthermore, memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital cards (SD cards), flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0078] It should be understood that each block of a flowchart and / or block diagram, and combinations of blocks in a flowchart and / or block diagram, can be implemented by a computer program. These computer programs can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that instructions executable by the processor of the computer or other programmable data processing device generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0079] These computer programs may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0080] These computer programs may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0081] This embodiment also provides a computer-readable storage medium, which includes at least one instruction that, when executed by a computer, implements the method steps in the diagnosis-treatment linkage generation method of Embodiment 1.

[0082] By executing a computer-readable storage medium containing at least one instruction, it is possible to achieve mapping queries, confidence level differential filling, and structured correlation output, effectively solving the technical problem of the disconnect between diagnosis and treatment in industrial equipment monitoring scenarios.

[0083] Similarly, for details not covered in this embodiment, please refer to Embodiment 1, Embodiment 2, and... Figures 1 to 7 The specific details will not be elaborated here.

[0084] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A diagnostic-treatment linkage generation method for predictive maintenance of industrial equipment, applied to predictive maintenance scenarios of industrial equipment, characterized in that, include: The mapping query step involves querying a preset diagnosis-treatment mapping library based on the fault location, nature, and severity in the diagnostic conclusion to obtain a basic treatment suggestion template that matches the fault location, nature, and severity. The diagnosis-treatment mapping library is a dynamically configurable structured knowledge base that contains many-to-many mapping relationships between fault location, nature, severity, and treatment actions. The confidence level differentiation step involves differentiating the basic treatment suggestion template according to the diagnostic confidence level to generate treatment suggestion content that matches the diagnostic confidence level. This differentiation includes: when the confidence level is high, filling in specific examination sites, examination items, and expected treatment actions; when the confidence level is medium, filling in priority suggestions, including priority and secondary examination items; and when the confidence level is low, filling in investigation path suggestions, including the order of investigation of possible sources of the problem and distinguishing investigation actions. The structured association output step links the diagnostic conclusions and treatment recommendations in a structured manner, so that each examination action in the treatment recommendation references the corresponding evidence identifier in the diagnostic conclusion, forming a two-way traceable chain between diagnosis and treatment; the two-way traceable chain includes a forward index from the diagnostic evidence identifier to the treatment action identifier and a reverse index from the treatment action identifier to the diagnostic evidence identifier.

2. The method according to claim 1, characterized in that, The structure of the diagnosis-treatment mapping library includes: The fault location dimension records the equipment component level at which faults can occur, including system level, component level, and part level; The failure nature dimension records the type and stage of the failure, including wear, fatigue, loosening, misalignment, imbalance, cavitation and its early / mid / late stages; The severity dimension records the severity level of the fault; The action dimension records the set of executable actions corresponding to the combination of fault location, fault nature, and severity. Each action includes action type, execution target, execution standard, and expected effect. The priority dimension records the recommended priority of each action under the same combination of fault location, fault nature, and severity. The priority is calculated based on a combination of historical success rate and on-site feasibility.

3. The method according to claim 1 or 2, characterized in that, The confidence level differential imputation step further includes: When the confidence level is high, the handling recommendations also include the expected handling time window and the recommended maintenance schedule. When the confidence level is medium, the recommended actions also include identifying interfering factors and confirmatory testing items that need to be excluded. When the confidence level is low, the recommended actions also include the recommended professionals to contact and the types of additional information that need to be provided.

4. The method according to any one of claims 1-3, characterized in that, The structured association output step includes: The diagnostic conclusions are structured and encapsulated, including the location of the fault, the nature of the fault, the severity, the confidence level, and a list of key evidence, into a diagnostic conclusion data package. The disposal recommendations are structured and encapsulated, which encapsulates the set of disposal actions, action priorities, execution criteria, and referenced evidence identifiers into a disposal recommendation data package; A bidirectional association index is established between the diagnostic conclusion data package and the treatment suggestion data package, so that each treatment action can be traced back to the diagnostic evidence on which it is based, and each diagnostic evidence can be queried to its corresponding treatment action; the bidirectional association index includes a forward index from the diagnostic evidence identifier to the treatment action identifier and a reverse index from the treatment action identifier to the diagnostic evidence identifier.

5. The method according to any one of claims 1-4, characterized in that, It also includes steps for dynamically adjusting the handling recommendations: When the diagnostic conclusion changes during the status transition, the diagnosis-treatment mapping library is queried again based on the changed diagnostic conclusion to generate updated treatment suggestions; The differences between the handling suggestions before and after the update are compared, a handling change explanation is generated, and pushed to the user interaction terminal; Record the history of changes to proposed actions, including the reasons for the changes, the time of the changes, and the content of the changes, for use in subsequent audits and for the accumulation of experience.

6. The method according to any one of claims 1-5, characterized in that, It also includes a feedback and linkage mechanism for handling effectiveness: When a user reports that a certain action has been performed, the system records the execution time, the person who performed the action, and the result. The treatment effect is compared with the expected effect in the diagnosis conclusion, and the matching degree of treatment effect is calculated. When the matching degree of the treatment effect is lower than the preset threshold, the diagnosis review process is triggered to re-evaluate the accuracy of the diagnosis conclusion; When the matching degree of treatment effect is higher than the preset threshold, the priority weight of the corresponding diagnosis-treatment mapping relationship is increased.

7. The method according to any one of claims 1-6, characterized in that, The diagnostic-treatment mapping library supports both manual review and dynamic maintenance. Administrators can view the historical usage count, success rate, and user satisfaction of each mapping relationship in the mapping review interface; It supports manually adding, modifying, or deleting mapping relationships. Modified mapping relationships take effect after manual confirmation. The system periodically optimizes the priority ranking of mapping relationships based on feedback data on the treatment effects.

8. The method according to any one of claims 1-7, characterized in that, It also includes an urgency assessment step: A comprehensive urgency score is calculated based on the importance level of the fault location, the severity of the fault nature, the severity rating and confidence level in the diagnostic conclusion; The urgency level of the response recommendations is classified according to the comprehensive urgency score, including immediate action, planned maintenance, and continuous monitoring; The urgency level classification results are output as an additional attribute of the handling recommendations to guide users in making decisions on handling priorities.

9. The method according to any one of claims 1-8, characterized in that, The output format is a three-segment output: The first section is the monitoring display section, which describes the current status and trend of the monitoring data; The second paragraph is the conclusion, which includes the location, nature, and severity of the fault. The third paragraph is the recommendation paragraph, which contains handling suggestions that match the confidence level; The three segments are linked in a structured manner through preset association markers, supporting segmented retrieval and independent consumption.

10. The method according to any one of claims 1-9, characterized in that, The treatment actions in the diagnosis-treatment mapping library also include conditional treatment suggestions related to the equipment's operating conditions. When the equipment is in a specific operating condition, the corresponding conditional treatment suggestion is automatically triggered.

11. A diagnostic-treatment linkage generation system for predictive maintenance of industrial equipment, characterized in that, include: The mapping query module is used to query the diagnosis-treatment mapping library based on the fault location, nature and severity in the diagnosis conclusion, and obtain a basic treatment suggestion template. The confidence level differential filling module is used to differentially fill the basic treatment suggestion template according to the diagnostic confidence level; The structured correlation output module is used to output the diagnostic conclusions and treatment recommendations in a structured way, forming a two-way traceable chain between diagnosis and treatment.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-10.

Images