Big model-based metrology pipeline fault diagnosis and intelligent operation and maintenance system

Abstract

This invention discloses a fault diagnosis and intelligent operation and maintenance system for metering pipelines based on a large model, belonging to the field of metering pipeline operation and maintenance technology. It includes a cloud platform and an application platform. The cloud platform includes a fault analysis module, which analyzes monitoring data to obtain fault diagnosis data and sends it to the application platform. It also configures a corresponding application correction model for the application platform. The application platform includes a monitoring module, an application diagnosis module, and an operation and maintenance module. The monitoring module performs real-time monitoring of the metering pipeline, obtains monitoring data, and sends the monitoring data to both the application diagnosis module and the fault analysis module in the cloud platform. The application diagnosis module performs fault diagnosis analysis, acquires the fault diagnosis data sent from the cloud platform in real time, and analyzes the fault diagnosis data and monitoring data using a pre-configured application correction model to obtain application diagnosis data. The operation and maintenance module performs operation and maintenance processing on the metering pipeline based on the application diagnosis data.

Description

Technical Field

[0001] This invention belongs to the field of metering pipeline operation and maintenance technology, specifically a metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model. Background Technology

[0002] In the field of power metering verification, with the accelerated construction of smart grids, the demand for power metering equipment verification has exploded. Traditional verification lines rely on manual inspections and periodic maintenance, which suffers from prominent problems such as delayed fault detection, low diagnostic efficiency, and high operation and maintenance costs. For example, potential faults such as wear and tear of mechanical parts, aging of electrical components, or sensor drift are often only detected after causing production line shutdowns or exceeding verification accuracy tolerances, resulting in a large backlog of electricity meters, delayed order delivery, and direct economic losses of tens of thousands of yuan per hour. At the same time, manual diagnosis relies on expert experience, and different personnel have subjective differences in judging fault characteristics, which can easily lead to misdiagnosis or missed diagnosis, further exacerbating operation and maintenance risks.

[0003] In existing technologies, some calibration systems attempt to introduce automated monitoring methods, such as collecting equipment status parameters through PLCs (Programmable Logic Controllers) or using SCADA (Supervisory Control and Data Acquisition) systems to visualize basic data. However, these solutions mostly focus on single equipment or local operating conditions, lacking correlation analysis of the overall production line status. For example, abnormal vibration of a robotic arm may be coupled with an increase in gearbox oil temperature, but traditional systems struggle to identify such cross-domain fault characteristics. Furthermore, while rule-based expert systems can handle known fault modes, they cannot adapt to dynamic equipment aging or the emergence of new fault types, resulting in insufficient model generalization ability. Summary of the Invention

[0004] To address the problems of the above solutions, this invention provides a fault diagnosis and intelligent operation and maintenance system for metering pipelines based on a large model.

[0005] The objective of this invention can be achieved through the following technical solutions: A fault diagnosis and intelligent operation and maintenance system for metering pipelines based on large models, including cloud and application terminals; The cloud-based system includes a fault analysis module; The fault analysis module is used to analyze the monitoring data of the metering production line through a preset fault diagnosis model, obtain the corresponding fault diagnosis data, send the fault diagnosis data to the application diagnosis module of the corresponding application terminal, and configure the corresponding application correction model for the corresponding application terminal according to the preset fault diagnosis model.

[0006] Furthermore, based on the preset fault diagnosis model, corresponding application correction models are configured for the respective application terminals, including: Obtain the metering pipeline information of the corresponding application terminal, and generate corresponding simulation material data based on the metering pipeline information. The simulation material data includes monitoring data and standard fault diagnosis results for the metering pipeline. Determine whether a new application correction model needs to be created for the metering pipeline based on the metering pipeline information; When it is determined that no new application correction model is needed, the corresponding application correction model is matched according to the metering pipeline information, the application correction model is corrected by simulating material data, and the corrected application correction model is configured in the corresponding application terminal. When it is determined that a new application correction model needs to be created, the fault diagnosis model is simulated and analyzed using simulated material data to obtain the simulation analysis results corresponding to the simulated material data; an application correction model is established based on the simulated material data and fault diagnosis data, and the application correction model is configured in the corresponding application terminal; and the newly created application correction model is stored in the cloud.

[0007] Furthermore, based on the metering pipeline information, it is determined whether a new application correction model needs to be created, including: The fault diagnosis model is simulated and verified using simulated material data to obtain simulation analysis results. Based on the simulation analysis results and simulated material data, the diagnostic deviation rate corresponding to various fault types is statistically analyzed, and the corresponding comprehensive diagnostic deviation rate is calculated based on the diagnostic deviation rate of various fault types. When the overall diagnostic deviation rate is greater than the threshold X1, it is determined that there is no need to create a new application correction model. When the overall diagnostic deviation rate is not greater than the threshold X1, it is determined that a new application correction model needs to be created.

[0008] Furthermore, based on the metering pipeline information, it is determined whether a new application correction model needs to be created, including: The process involves: acquiring verification records for establishing a large-scale fault diagnosis model; identifying various metering pipeline information corresponding to the verification records; marking the metering pipeline information as baseline pipeline information; calculating the diagnostic deviation rate corresponding to various fault types based on the verification records and baseline pipeline information; calculating the comprehensive diagnostic deviation rate corresponding to the corresponding baseline pipeline information based on the diagnostic deviation rate of various fault types; and establishing an application evaluation model based on the comprehensive diagnostic deviation rate corresponding to each baseline pipeline information. The metering pipeline information is matched with the information of each benchmark pipeline based on similarity. The benchmark pipeline information with the highest similarity is marked as the reference metering pipeline information. The comprehensive diagnostic deviation rate corresponding to the reference metering pipeline information is identified. The metering pipeline information, reference metering pipeline information, and comprehensive diagnostic deviation rate are integrated into the input data and input into the application evaluation model for analysis to obtain the corresponding application evaluation value. Determine whether a new application correction model is needed based on the application evaluation values.

[0009] Furthermore, the expression for the applied evaluation model is: ; In the formula: (s, s', PG) are the input data, s represents the measurement pipeline information, s' represents the reference measurement pipeline information, and PG is the comprehensive diagnostic deviation rate of the reference measurement pipeline; the output data is the application evaluation value YP(s, s', PG), and the application evaluation value is 1 or 0.

[0010] Furthermore, the calculation of the overall diagnostic deviation rate includes: Preset weighting coefficients for various fault types and label the weighting coefficients as λ. i i = 1, 2, ..., n, where n is the number of fault types; The corresponding comprehensive diagnostic deviation rate is calculated based on the evaluation formula, which is as follows: ; In the formula: PG is the overall diagnostic deviation rate; ZPi represents the diagnostic deviation rate of the corresponding fault type.

[0011] The application module includes a monitoring module, an application diagnostic module, and an operation and maintenance module. The monitoring module is used to monitor the metering production line in real time, obtain monitoring data, and send the monitoring data to the application diagnostic module and the cloud-based fault analysis module respectively. The application diagnostic module is used to perform fault diagnosis analysis, acquire fault diagnosis data sent from the cloud in real time, and analyze the fault diagnosis data and monitoring data through a pre-configured application correction model to obtain application diagnostic data.

[0012] Furthermore, the fault diagnosis data and monitoring data are analyzed using a pre-configured application correction model, including: Based on the simulated material data, identify the types of faults, perform statistical analysis on the simulated material data and fault diagnosis data according to various fault types, determine the diagnostic deviation of different fault types under different monitoring data, and determine the corresponding correction range based on the diagnostic deviation of fault types under different monitoring data. Analyze the monitoring data and fault diagnosis data according to the correction range of the corresponding fault type to obtain optimized monitoring data and optimized fault diagnosis data; The application correction model is used to analyze the optimized monitoring data and optimized fault diagnosis data to obtain the corresponding application diagnostic data.

[0013] Furthermore, based on the diagnostic deviations of the fault type under different monitoring data, the corresponding correction range is determined, including: Generate corresponding diagnostic deviation curves based on the diagnostic deviations of different monitoring data according to the fault type; Identify the correction range for the fault type based on the diagnostic deviation curve, and mark the corresponding correction range on the diagnostic deviation curve.

[0014] The operation and maintenance module is used to perform operation and maintenance processing on the metering pipeline based on application diagnostic data.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention achieves a highly efficient upgrade of the operation and maintenance model in the field of power metering and verification by constructing an intelligent diagnostic service system based on a unified platform. Addressing the core problems of traditional solutions, such as delayed fault detection, low diagnostic efficiency, and insufficient model generalization ability, this invention innovatively deploys a multimodal large-scale model on a cloud platform. Through a centralized training and distributed inference architecture design, it simultaneously resolves the contradiction between resource utilization and diagnostic accuracy: On the one hand, the platform relies on massive equipment data to construct a fault diagnosis model with strong generalization capabilities, avoiding the waste of computing power caused by repeated construction by various users and reducing the operation and maintenance cost per device; on the other hand, dynamic correction is performed through application-based correction models, so that only application-based correction models need to be built for different users, greatly reducing the modeling difficulty. This forms a closed-loop optimization system of "cloud intelligence + edge calibration," effectively solving the model failure problem caused by dynamic equipment aging and new fault modes. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a block diagram illustrating the principle of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, the metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model includes the cloud and the application end; The cloud platform is used to provide services to numerous users, including State Grid. It can be configured by a higher-level department or through delegation, or it can be configured independently by a single entity but shared with other users. Ideally, it should be configured by an entity within the system for higher credibility and data security. Otherwise, it needs to have reliable data encryption and security protection technologies to prevent data leakage.

[0020] The cloud-based system includes a fault analysis module; The fault analysis module is used to analyze the received monitoring data of the metering pipeline through a preset fault diagnosis model, obtain the corresponding fault diagnosis data, send the fault diagnosis data to the corresponding application terminal, and configure the corresponding application correction model for the corresponding application terminal according to the preset fault diagnosis model.

[0021] The modified model is applied to eliminate potential errors caused by differences in the metering pipeline at the actual user's location compared to the large fault diagnosis model.

[0022] In one embodiment, the fault diagnosis big model is: based on existing big model technology, a corresponding fault diagnosis big model is established for the fault diagnosis of various metering production lines. Although there are differences in equipment models, process flows, sensor configurations, etc., the core fault modes of metering and verification production lines have common characteristics, so a common big model can be set up.

[0023] In one embodiment, a corresponding application correction model is configured for the corresponding application based on a preset fault diagnosis model, including: Obtain the metering pipeline information of the corresponding application terminal, and generate corresponding simulation material data based on the metering pipeline information. The simulation material data includes monitoring data and standard fault diagnosis results for the metering pipeline, and can be generated based on historical fault data of similar metering pipelines. Determine whether a new application correction model is needed based on the metering pipeline information; When it is determined that no new application correction model is needed, the corresponding application correction model is matched according to the metering pipeline information. The application correction model is then corrected by simulating material data, i.e., verification and adaptive adjustment are performed. Appropriate fine-tuning is then done, and the corrected application correction model is configured in the corresponding application. When it is determined that a new application correction model needs to be created, the fault diagnosis big model is simulated and analyzed using simulated material data to obtain the simulation analysis results corresponding to the simulated material data. That is, the fault diagnosis big model analyzes the monitoring data corresponding to the simulated material data to obtain the corresponding fault diagnosis data, which is marked as the simulation analysis results. A corresponding application correction model is established based on the simulated material data and the fault diagnosis data. Techniques such as pruning, quantization, and knowledge distillation can also be applied to perform lightweight processing on the application correction model. The application correction model is configured in the corresponding application terminal. The newly created application correction model is stored in the cloud.

[0024] In one embodiment, determining whether a new application correction model needs to be created based on metering pipeline information includes: The fault diagnosis model is simulated and verified using simulated material data to obtain simulation analysis results. Based on the simulation analysis results and simulated material data, the diagnostic deviation rate corresponding to various fault types is statistically analyzed. Preset weighting coefficients for various fault types and label the weighting coefficients as λ. i i = 1, 2, ..., n, where n is the number of fault types; The corresponding comprehensive diagnostic deviation rate is calculated based on the evaluation formula, which is as follows: ; In the formula: PG is the overall diagnostic deviation rate; ZPi represents the diagnostic deviation rate of the corresponding fault type; When the overall diagnostic deviation rate is greater than the threshold X1, it is determined that there is no need to create a new application correction model. When the overall diagnostic deviation rate is not greater than the threshold X1, it is determined that a new application correction model needs to be created.

[0025] The threshold X1 is set by the platform. For example, the threshold X1 is set based on the value at which the comprehensive diagnostic deviation rate would fail to meet the user's application needs or exceed the preset correction level.

[0026] In the above embodiments, judgments can also be made based on the simulation results of the simulated material data in other ways.

[0027] In one embodiment, determining whether a new application correction model needs to be created based on metering pipeline information includes: Obtain the validation records of the established fault diagnosis model. These records are the data used to validate the fault diagnosis model using the validation set after its training. Based on these records, determine the comprehensive diagnostic deviation rate of the fault diagnosis model for various metering pipeline information. For metering pipelines trained by the fault diagnosis model, its fault diagnosis should meet application requirements. For untrained metering pipelines, a preliminary assessment based on the comprehensive diagnostic deviation rate is required. To differentiate it from metering pipeline information requiring evaluation, the metering pipeline information corresponding to the fault diagnosis model is marked as the baseline pipeline information. An application evaluation model is established based on the comprehensive diagnostic deviation rate corresponding to each baseline pipeline information. The expression of the application evaluation model is as follows: ; In the formula: (s, s', PG) are the input data, where s represents the corresponding metrology pipeline information, s' represents the corresponding reference metrology pipeline information, and PG represents the comprehensive diagnostic deviation rate of the corresponding reference metrology pipeline; the output data is the application evaluation value YP(s, s', PG), which is 1 or 0; the pipeline difference information is determined based on the metrology pipeline information and the reference metrology pipeline information; based on the pipeline difference information and the comprehensive diagnostic deviation, it is estimated whether the difference will cause the comprehensive diagnostic deviation of the metrology pipeline to be greater than the threshold X1. If it is estimated to be greater than the threshold X1, then s does not meet the application requirements; otherwise, s meets the application requirements. The corresponding training set is set up based on the above method for training. The similarity of the metering pipeline information with that of each benchmark pipeline information is matched, and the benchmark pipeline information with the highest similarity is marked as the reference metering pipeline information; the comprehensive diagnostic deviation rate corresponding to the reference metering pipeline information is identified; the similarity is calculated according to the existing similarity calculation method, such as cosine similarity, etc. The metering pipeline information, reference metering pipeline information, and comprehensive diagnostic deviation rate are integrated into the input data and input into the application evaluation model for analysis to obtain the corresponding application evaluation value. When the application evaluation value is 1, it is determined that no new application correction model is needed. When the application evaluation value is 0, it is determined that a new application correction model needs to be created.

[0028] In one embodiment, a corresponding application correction model is established based on the simulated material data and fault diagnosis data. This model can be established based on machine learning, deep learning algorithms, etc., and trained using the simulated material data and fault diagnosis data labeled with the corresponding training set. Any other method can be used to accurately correct the fault diagnosis results of the large fault diagnosis model analysis.

[0029] The application module includes a monitoring module, an application diagnostic module, and an operation and maintenance module. The monitoring module is used to monitor the metering production line in real time, obtain corresponding monitoring data, and send the monitoring data to the application diagnostic module and the fault analysis module in the cloud.

[0030] The application diagnostic module is used to perform fault diagnosis analysis, acquire fault diagnosis data sent from the cloud in real time, analyze the fault diagnosis data and monitoring data through a pre-configured application correction model, obtain the fault diagnosis data after metering pipeline correction, and mark the application diagnostic data.

[0031] In one embodiment, fault diagnosis data and monitoring data are analyzed using a pre-configured application correction model, allowing for complete analysis and correction directly through the application correction model.

[0032] In one embodiment, fault diagnosis data and monitoring data are analyzed using a pre-configured application correction model, including: The simulation data identifies the fault types, with the same fault considered as different fault types for different equipment targets. Statistical analysis is performed on the simulation data and fault diagnosis data based on various fault types to determine the diagnostic deviation of different fault types under different monitoring data, generating corresponding diagnostic deviation curves. The vertical axis represents the diagnostic deviation, and the horizontal axis represents the representative sequence number of the corresponding monitoring data (e.g., 1 corresponds to one monitoring data point). The monitoring data are sorted according to their corresponding data range in a preset order, such as temperature from low to high, noise level from low to high, etc. A comprehensive sort is then calculated based on the weight coefficients of each monitoring item, and the results are then distributed on the horizontal axis. Alternatively, diagnostic deviation curves can be generated using other methods. Determine the correction range for the corresponding fault type based on the diagnostic deviation curve, that is, the monitoring data range corresponding to the diagnostic deviation not meeting the requirements; mark the corresponding correction range on the diagnostic deviation curve. Based on the analysis of each diagnostic deviation curve, the monitoring data and fault diagnosis data are obtained to obtain the monitoring data and fault diagnosis data that need to be corrected, namely the monitoring data within the correction range and the fault diagnosis data corresponding to the monitoring data in the fault type, which are marked as optimized monitoring data and optimized fault diagnosis data, respectively. The application correction model is used to analyze the optimized monitoring data and optimized fault diagnosis data to obtain the corresponding application diagnostic data.

[0033] In one embodiment, the correction range can also be determined directly based on the diagnostic deviation of the fault type under different monitoring data, but the diagnostic deviation curve is more intuitive; subsequently, the optimized monitoring data and optimized fault diagnosis data are determined according to the correction range.

[0034] The operation and maintenance module is used to perform operation and maintenance on the metering pipeline based on application diagnostic data. Since there are currently many operation and maintenance processes based on fault diagnostic data, they will not be described in detail. If a potential fault is predicted at a certain time, the corresponding operation and maintenance process will be performed based on the result, or the operation and maintenance process will be performed according to the equipment status corresponding to the application diagnostic data. When it is determined that there is no fault, the corresponding health status score can also be determined by the fault diagnosis big model, and then the corresponding operation and maintenance process will be performed.

[0035] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0036] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A fault diagnosis and intelligent operation and maintenance system for metering pipelines based on a large model, characterized in that, Including both cloud and application-side; The cloud platform includes a fault analysis module; the application platform includes a monitoring module, an application diagnostic module, and an operation and maintenance module. The fault analysis module is used to analyze the monitoring data of the metering line through a preset fault diagnosis model, obtain the corresponding fault diagnosis data, send the fault diagnosis data to the application diagnosis module of the corresponding application terminal, and configure the corresponding application correction model for the corresponding application terminal according to the preset fault diagnosis model. The monitoring module is used to monitor the metering production line in real time, obtain monitoring data, and send the monitoring data to the application diagnostic module and the cloud-based fault analysis module respectively. The application diagnostic module is used to perform fault diagnosis analysis, acquire fault diagnosis data sent from the cloud in real time, and analyze the fault diagnosis data and monitoring data through a pre-configured application correction model to obtain application diagnostic data. The operation and maintenance module is used to perform operation and maintenance processing on the metering pipeline based on application diagnostic data.

2. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model as described in claim 1, characterized in that, Based on the preset fault diagnosis model, configure the corresponding application correction model for the relevant application, including: Obtain the metering pipeline information of the corresponding application terminal, and generate corresponding simulation material data based on the metering pipeline information. The simulation material data includes monitoring data and standard fault diagnosis results for the metering pipeline. Determine whether a new application correction model needs to be created for the metering pipeline based on the metering pipeline information; When it is determined that no new application correction model is needed, the corresponding application correction model is matched according to the metering pipeline information, the application correction model is corrected by simulating material data, and the corrected application correction model is configured in the corresponding application terminal. When it is determined that a new application correction model needs to be created, the fault diagnosis model is simulated and analyzed using simulated material data to obtain the simulation analysis results corresponding to the simulated material data; an application correction model is established based on the simulated material data and fault diagnosis data, and the application correction model is configured in the corresponding application terminal; and the newly created application correction model is stored in the cloud.

3. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model as described in claim 2, characterized in that, Determine whether a new application correction model needs to be created based on the metering pipeline information, including: The fault diagnosis model is simulated and verified using simulated material data to obtain simulation analysis results. Based on the simulation analysis results and simulated material data, the diagnostic deviation rate corresponding to various fault types is statistically analyzed, and the corresponding comprehensive diagnostic deviation rate is calculated based on the diagnostic deviation rate of various fault types. When the overall diagnostic deviation rate is greater than the threshold X1, it is determined that there is no need to create a new application correction model. When the overall diagnostic deviation rate is not greater than the threshold X1, it is determined that a new application correction model needs to be created.

4. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model as described in claim 2, characterized in that, Determine whether a new application correction model needs to be created based on the metering pipeline information, including: The process involves: acquiring verification records for establishing a large-scale fault diagnosis model; identifying various metering pipeline information corresponding to the verification records; marking the metering pipeline information as baseline pipeline information; calculating the diagnostic deviation rate corresponding to various fault types based on the verification records and baseline pipeline information; calculating the comprehensive diagnostic deviation rate corresponding to the corresponding baseline pipeline information based on the diagnostic deviation rate of various fault types; and establishing an application evaluation model based on the comprehensive diagnostic deviation rate corresponding to each baseline pipeline information. The metering pipeline information is matched with the information of each benchmark pipeline based on similarity. The benchmark pipeline information with the highest similarity is marked as the reference metering pipeline information. The comprehensive diagnostic deviation rate corresponding to the reference metering pipeline information is identified. The metering pipeline information, reference metering pipeline information, and comprehensive diagnostic deviation rate are integrated into the input data and input into the application evaluation model for analysis to obtain the corresponding application evaluation value. Determine whether a new application correction model is needed based on the application evaluation values.

5. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model according to claim 4, characterized in that, The expression for the application evaluation model is: ； In the formula: (s, s', PG) are the input data, s represents the measurement pipeline information, s' represents the reference measurement pipeline information, and PG is the comprehensive diagnostic deviation rate of the reference measurement pipeline; the output data is the application evaluation value YP(s, s', PG), and the application evaluation value is 1 or 0.

6. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model according to claim 3 or 4, characterized in that, The calculation of the overall diagnostic deviation rate includes: Preset weighting coefficients for various fault types and label the weighting coefficients as λ. i i = 1, 2, ..., n, where n is the number of fault types; The corresponding comprehensive diagnostic deviation rate is calculated based on the evaluation formula, which is as follows: ； In the formula: PG is the overall diagnostic deviation rate; ZPi represents the diagnostic deviation rate of the corresponding fault type.

7. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model as described in claim 1, characterized in that, The fault diagnosis data and monitoring data are analyzed using a pre-configured application correction model, including: Based on the simulated material data, identify the types of faults, perform statistical analysis on the simulated material data and fault diagnosis data according to various fault types, determine the diagnostic deviation of different fault types under different monitoring data, and determine the corresponding correction range based on the diagnostic deviation of fault types under different monitoring data. Analyze the monitoring data and fault diagnosis data according to the correction range of the corresponding fault type to obtain optimized monitoring data and optimized fault diagnosis data; The application correction model is used to analyze the optimized monitoring data and optimized fault diagnosis data to obtain the corresponding application diagnostic data.

8. The metering pipeline fault diagnosis and intelligent operation and maintenance system based on a large model according to claim 7, characterized in that, The corresponding correction range is determined based on the diagnostic deviation of the fault type under different monitoring data, including: Generate corresponding diagnostic deviation curves based on the diagnostic deviations of different monitoring data according to the fault type; Identify the correction range for the fault type based on the diagnostic deviation curve, and mark the corresponding correction range on the diagnostic deviation curve.

Images