Operation and maintenance function item recognition and diagnosis method and device based on big data

Abstract

The invention relates to the technical field of data processing, and discloses an operation and maintenance function item recognition and diagnosis method and device based on big data. The method comprises the following steps: performing threshold diagnosis and difference analysis on function item technology description data and budgeting data to obtain function item diagnosis data and assessment requirement data; performing recursive calculation and adjustment processing on the function item diagnosis data and the assessment requirement data to obtain the number of function points and function item cost data, and performing tensor decomposition and optimization processing to obtain a function item adjustment scheme and investment control data; and performing comprehensive diagnosis and evaluation on the function item adjustment scheme and the investment control data to obtain a function point identification report and optimization suggestion data. Through a systematic data processing flow and a standardized diagnosis evaluation system, accurate identification, scientific diagnosis and effective optimization of operation and maintenance function items are realized, and the technical problems of inaccurate function item identification, non-standard diagnosis flow and non-strict cost control in the prior art are solved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and device for identifying and diagnosing operation and maintenance function items based on big data. Background Art

[0002] In the current context of big data technology applications, the identification and diagnosis of operation and maintenance function items has become an important part of project management. Existing technologies mainly use methods such as manual experience judgment, simple mathematical statistics, and basic data analysis to identify function items, and have accumulated certain practical experience in project establishment, budget preparation, and technical document writing. At the same time, some operation and maintenance projects have begun to introduce technical means such as machine learning and data mining, trying to improve the accuracy of function item identification through data analysis. Existing operation and maintenance function item identification methods have played an important role in project pre-planning, process management, and post-evaluation.

[0003] However, existing technologies have obvious deficiencies in practical applications. First, the functional points of operation and maintenance projects are diverse, complex, and dynamic, and traditional identification methods are difficult to accurately capture the correlation and evolution laws between functional items. Secondly, the existing data processing methods lack systematicity and consistency, resulting in frequent inconsistencies between the technical description of functional items and the preparation of budget estimates. Thirdly, due to the lack of standardized diagnostic processes and evaluation systems, it is difficult for technical personnel to effectively identify functional point anomalies and conduct cost control. In addition, existing methods lack the ability to conduct deep mining and multi-dimensional analysis of data, and cannot fully utilize the advantages of big data technology for accurate diagnosis and optimized decision-making. Summary of the invention

[0004] The present application provides a method and device for identifying and diagnosing operation and maintenance function items based on big data, which is used to achieve accurate identification, scientific diagnosis and effective optimization of operation and maintenance function items through a systematic data processing process and a standardized diagnosis and evaluation system, and to solve the technical problems of inaccurate function item identification, non-standard diagnosis process and lax cost control in the prior art.

[0005] In the first aspect, the present application provides an operation and maintenance function item identification and diagnosis method based on big data, and the operation and maintenance function item identification and diagnosis method based on big data includes: performing feature decomposition and classification mapping on the operation and maintenance project maintenance objects, maintenance contents, maintenance requirements and maintenance level data to obtain an operation and maintenance function item feature set and a function item classification rule; performing data fusion and cross-identification on the operation and maintenance function item feature set to obtain function item technical description data and budget preparation data; performing threshold diagnosis and difference analysis on the function item technical description data and budget preparation data to obtain function item diagnosis data and assessment requirement data; performing recursive calculation and adjustment processing on the function item diagnosis data and assessment requirement data to obtain function point quantity and function item cost data; performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plan and investment control data; performing comprehensive diagnosis and evaluation on the function item adjustment plan and investment control data to obtain a function point identification report and optimization suggestion data.

[0006] In a second aspect, the present application provides an operation and maintenance function item identification and diagnosis device based on big data, and the operation and maintenance function item identification and diagnosis device based on big data includes: A mapping module is used to perform feature decomposition and classification mapping on the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance project to obtain the feature set of operation and maintenance function items and the classification rules of function items; An identification module is used to perform data fusion and cross-identification on the feature set of the operation and maintenance function items to obtain technical description data of the function items and budget preparation data; An analysis module, used to perform threshold diagnosis and difference analysis on the functional item technical description data and the budget preparation data, to obtain functional item diagnosis data and assessment requirement data; A processing module, used for recursively calculating and adjusting the function item diagnosis data and the assessment requirement data to obtain the function point quantity and function item cost data; A decomposition module, used for performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plans and investment control data; The evaluation module is used to perform comprehensive diagnosis and evaluation on the functional item adjustment plan and investment control data to obtain a function point identification report and optimization suggestion data.

[0007] In the technical solution provided in the present application, by performing feature decomposition and classification mapping on the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance projects, a complete feature set of operation and maintenance function items and classification rules for function items are established, which effectively solves the problem of inaccurate function item identification; by performing data fusion and cross-identification on the feature set of operation and maintenance function items, the technical description of the function items and the budget preparation data are obtained, and the unity of technical requirements and cost estimates is achieved; threshold diagnosis and difference analysis are performed on the technical description of the function items and the budget preparation data to obtain the function item diagnosis data and assessment requirement data, and a scientific diagnosis and evaluation mechanism is established; by performing recursive calculation and adjustment processing on the function item diagnosis data and assessment requirement data, accurate function point quantity and function item cost data are obtained, and precise resource allocation is achieved; tensor decomposition and optimization processing are used to perform in-depth analysis on the function point quantity and function item cost data, and reasonable function item adjustment plans and investment control data are generated, ensuring the feasibility of the optimization plan; finally, through comprehensive diagnosis and evaluation, function point identification reports and optimization suggestion data are formed, providing a reliable basis for the continuous improvement of operation and maintenance projects. The entire method forms a complete data processing closed loop, with close logical connections between each step and smooth data flow, ensuring the controllability of the entire process from data collection, processing, analysis to result output, and significantly improving the accuracy, scientificity and practicality of operation and maintenance function item identification and diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0009] Figure 1 A schematic diagram of an embodiment of a method for identifying and diagnosing operation and maintenance function items based on big data in an embodiment of the present application; Figure 2 This is a schematic diagram of an embodiment of an operation and maintenance function item identification and diagnosis device based on big data in an embodiment of the present application. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide a method and device for identifying and diagnosing operation and maintenance function items based on big data. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0011] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the operation and maintenance function item identification and diagnosis method based on big data includes: Step S101: Decompose and classify the maintenance object, maintenance content, maintenance requirements and maintenance level data of the operation and maintenance project to obtain the operation and maintenance function item feature set and function item classification rules; Step S102: Perform data fusion and cross-identification on the feature set of the operation and maintenance function items to obtain technical description data of the function items and budget preparation data; Step S103: Perform threshold diagnosis and difference analysis on the function item technical description data and the budget preparation data to obtain function item diagnosis data and assessment requirement data; Step S104: recursively calculate and adjust the function item diagnosis data and the assessment requirement data to obtain the function point quantity and function item cost data; Step S105, performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plans and investment control data; Step S106: Comprehensively diagnose and evaluate the function item adjustment plan and investment control data to obtain a function point identification report and optimization suggestion data.

[0012] It is understandable that the execution subject of the present application can be a big data-based operation and maintenance function item identification and diagnosis device, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0013] Specifically, the data of operation and maintenance projects are feature decomposed and classified and mapped. Specifically, the maintenance object data is collected, including maintenance equipment type, equipment parameters, operating status and other information; maintenance content data, covering daily inspection, fault repair, performance optimization and other work content; maintenance requirement data, including maintenance cycle, maintenance standards, quality requirements and other regulatory requirements; maintenance level data, recording the priority, importance, response time and other hierarchical information of maintenance tasks. These raw data are feature decomposed, key dimension indicators are extracted, and a feature matrix is ​​established. Then, the feature set and classification rules of functional items are constructed through classification mapping. In the data fusion stage, data from different sources in the feature set of functional items are uniformly processed and standardized. For example, for the maintenance project of the refrigeration system of a data center, equipment operation data (temperature, humidity, energy consumption, etc.), maintenance record data (maintenance time, maintenance content, fault type, etc.) and personnel operation data (operation record, response time, processing method, etc.) are integrated. Through data cleaning, standardization and feature selection, a unified technical description of functional items is formed, including indicators such as maintenance frequency (twice a day), maintenance difficulty (professional level), and equipment status (normal operation time 95%). At the same time, preliminary budget preparation data is generated, including specific cost items such as labor costs, material costs, and energy costs.

[0014] In the threshold diagnosis and difference analysis phase, set the threshold standards for various indicators, such as the temperature control range of 18-24℃, the energy consumption index PUE not exceeding 1.5, and the fault response time not exceeding 15 minutes. By comparing the difference between the actual operating data and the standard threshold, identify abnormal points and problem areas. For example, it is found that the energy consumption index of a refrigeration unit exceeds the threshold by 20%. Through difference analysis, it is determined that this is due to the decrease in refrigeration efficiency and uneven load distribution, and the corresponding diagnostic data and assessment requirement data are generated. In the recursive calculation and adjustment processing stage, in-depth analysis and optimization are carried out for the problems found in the diagnosis. For example, for the problem of the refrigeration unit, the optimal range of the equipment operating parameters is determined through recursive calculation, and the input-output ratio of various maintenance measures is analyzed, and finally the number of function points (need to increase 1 special inspection per day) and function item cost data (monthly maintenance cost increases by 8,000 yuan) are obtained.

[0015] In the tensor decomposition and optimization processing stage, the number of function points and cost data are analyzed in multiple dimensions to identify key influencing factors and optimization space. Through data dimensionality reduction and feature extraction, common problems and characteristic laws in refrigeration system maintenance are found, such as uneven distribution of refrigeration load mainly occurs during the peak hours in the morning and evening, and the temperature fluctuations in some areas are large. Based on this, a function item adjustment plan is generated, including specific measures such as optimizing inspection routes, adjusting maintenance frequency, and improving maintenance methods. At the same time, investment control data is formulated, such as setting sub-item cost caps and dividing budget execution nodes. By setting an evaluation indicator system, including technical feasibility, economic rationality, implementation difficulty and other dimensions, the plans are scored and ranked. For example, for the optimization plan of the refrigeration system, the energy-saving effect, investment payback period, maintenance difficulty and other aspects are evaluated, and a detailed function point identification report and optimization suggestion data are generated. The identification report clearly points out the existing problems (high energy consumption, large temperature fluctuations), optimization directions (load balancing, intelligent adjustment) and specific measures (adding temperature sensors, transforming control systems), and provides detailed suggestions such as implementation steps and expected effects.

[0016] In the embodiment of the present application, by performing feature decomposition and classification mapping on the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance project, a complete operation and maintenance function item feature set and function item classification rules are established, which effectively solves the problem of inaccurate function item identification; by performing data fusion and cross-identification on the operation and maintenance function item feature set, the function item technical description and budget preparation data are obtained, and the unity of technical requirements and cost estimation is achieved; threshold diagnosis and difference analysis are performed on the function item technical description and budget preparation data to obtain function item diagnosis data and assessment requirement data, and a scientific diagnosis and evaluation mechanism is established; by performing recursive calculation and adjustment processing on the function item diagnosis data and assessment requirement data, accurate function point quantity and function item cost data are obtained, and precise resource allocation is achieved; tensor decomposition and optimization processing are used to perform in-depth analysis on the function point quantity and function item cost data, and reasonable function item adjustment plans and investment control data are generated to ensure the feasibility of the optimization plan; finally, through comprehensive diagnosis and evaluation, function point identification reports and optimization suggestion data are formed, which provide a reliable basis for the continuous improvement of operation and maintenance projects. The entire method forms a complete data processing closed loop, with close logical connections between each step and smooth data flow, ensuring the controllability of the entire process from data collection, processing, analysis to result output, and significantly improving the accuracy, scientificity and practicality of operation and maintenance function item identification and diagnosis.

[0017] In a specific embodiment, the process of executing step S101 may specifically include the following steps: (1) Clean and standardize the original data of the operation and maintenance project to obtain standardized maintenance data, and segment the standardized maintenance data into time windows to obtain time series feature data; (2) Perform multi-scale decomposition of the time series feature data through wavelet transform to obtain the maintenance object feature vector, and perform frequency domain feature extraction on the maintenance object feature vector to obtain frequency domain feature data; (3) Calculate the correlation between the frequency domain feature data and the maintenance content data to obtain the feature correlation matrix, and perform hierarchical clustering on the feature correlation matrix to obtain the hierarchical structure of functional items; (4) Perform semantic analysis on the function item hierarchy and maintenance requirement data to obtain a semantic feature set, and perform attribute mapping on the semantic feature set to obtain a function item attribute table; (5) Extract rules from the function item attribute table and maintenance level data to obtain a rule set, and verify and screen the rule set to obtain a candidate rule base; (6) Calculate the weights of the candidate rule base to obtain a weighted rule set, and merge the weighted rule set to obtain the target classification rule; (7) Cross-validate the target classification rules to obtain the validation result data, and modify the classification rules based on the validation result data to obtain the operation and maintenance function item feature set and function item classification rules.

[0018] Specifically, the original data of the operation and maintenance project is processed. The original data includes equipment operation data (operation status, performance parameters, alarm information), maintenance record data (maintenance time, maintenance content, processing method) and personnel operation data (operation record, response time, processing result). During the data cleaning process, the abnormal values ​​and missing values ​​in the original data are processed. For example, the records of temperature data exceeding the normal range are eliminated, and the missing operation parameters are supplemented by interpolation method. Standardization processing unifies data of different types and dimensions to the same scale to make the data comparable. In time window segmentation, the continuous data stream is divided into multiple data segments according to fixed time intervals. For example, the equipment operation data is divided into 24 time windows according to the hour unit to form a data set, and each window contains all the monitoring parameters in the period. Multi-scale decomposition through wavelet transform can separate the high-frequency and low-frequency components in the data. The high-frequency component reflects the short-term fluctuation characteristics, and the low-frequency component reflects the long-term change trend.

[0019] The maintenance object feature vector is extracted through frequency domain feature extraction to obtain frequency domain feature data, including the change frequency, fluctuation amplitude, energy distribution and other features of various parameters. For example, the temperature data of an air conditioning system extracts features such as temperature change rate, temperature fluctuation range, and temperature stability. The correlation between these frequency domain feature data and the maintenance content data is calculated to construct a feature correlation matrix to reflect the correlation strength between different features. The correlation matrix is ​​analyzed by the hierarchical clustering method, and features with high similarity are classified into one category to form a hierarchical functional item structure. For example, different types of maintenance activities such as daily inspections, regular maintenance, and fault repair are hierarchically classified according to feature similarity. Semantic analysis is performed on this hierarchical structure to extract key descriptive information in maintenance activities, such as maintenance cycle, equipment type, technical requirements, etc., to form a semantic feature set.

[0020] For example, in an industrial air conditioning maintenance project, the original data includes operating parameters such as temperature, humidity, and energy consumption, which are collected every 5 minutes. In the data cleaning stage, outliers with temperatures outside the range of -10°C to 50°C are deleted, and the missing humidity data is supplemented by the average value of adjacent time points. Standardization processing unifies the data of different units such as temperature, humidity, and energy consumption into the range of 0-1. Segmentation is performed according to the window of 1 hour to obtain data segments of 24 consecutive time windows. Features such as temperature average, temperature change rate, and humidity stability are extracted in each time window. At the same time, combined with the content in the maintenance record, such as "Monthly inspection found abnormal temperature fluctuations, replace the thermostat", etc., the correlation between features is calculated. Through hierarchical clustering, similar maintenance features are classified to form a functional item structure including daily inspection (temperature inspection, equipment cleaning), preventive maintenance (regular maintenance, parts replacement), fault repair (fault diagnosis, repair processing), etc. Semantic analysis is performed on maintenance records to extract time requirements such as "weekly inspection", "quarterly maintenance", and "fault response", and performance requirements such as "temperature control accuracy ±1°C" and "energy consumption index ≤2.0", and map them into a function item attribute table.

[0021] Formulate rules based on maintenance level data. For example, the importance of equipment is divided into three levels: A / B / C. The corresponding maintenance response time requirements are 15 minutes / 30 minutes / 60 minutes, and the fault repair time requirements are 2 hours / 4 hours / 8 hours. Through rule verification and screening, a candidate rule library containing time requirements, performance indicators, quality standards, etc. is established. The rules are weighted and merged to form the final classification rules, and the accuracy and reliability of the rules are ensured through cross-validation. The scattered operation and maintenance data is converted into the system's functional item feature set and classification rules, providing data support for the refined management of operation and maintenance projects. The accuracy and reliability of functional item identification are ensured through standardized processing procedures.

[0022] In a specific embodiment, the process of executing step S102 may specifically include the following steps: (1) Perform data standardization on the feature set of operation and maintenance function items to obtain a standardized function item data matrix, and complete the missing values ​​of the standardized function item data matrix to obtain a complete feature matrix; (2) Perform correlation analysis on the complete feature matrix to obtain the feature correlation coefficient matrix, and perform feature selection on the feature correlation coefficient matrix to obtain the core feature data; (3) Perform multi-source data fusion on the core feature data to obtain a fused feature vector, and cross-validate the fused feature vector to obtain a validation feature set; (4) Extract the technical description of the verification feature set to obtain the technical description elements, and perform semantic analysis on the technical description elements to obtain a semantic feature table; (5) Mapping the semantic feature table to budget items to obtain budget-related data, and performing cost accounting on the budget-related data to obtain the cost baseline value; (6) Decompose the cost baseline value to obtain itemized cost data, and then combine and compile the itemized cost data to obtain the budget data set; (7) Perform consistency check on the budget data set to obtain the test result data, and perform data calibration based on the test result data to obtain the functional item technical description data and budget preparation data.

[0023] Specifically, standardization unifies data of different dimensions and ranges to the same scale. The processing objects include multidimensional data such as maintenance frequency, response time, and operating parameters. After data standardization, a standardized function item data matrix is ​​formed. The missing values ​​in the matrix are filled by multiple interpolation methods. Considering the time series characteristics and correlation of the data, the missing data are estimated by neighbor value averaging, linear regression and other methods to obtain a complete feature matrix. In the correlation analysis stage, the relationship between each feature in the complete feature matrix is ​​quantitatively calculated to obtain a feature correlation coefficient matrix. The correlation coefficient reflects the degree of association between features. By setting the correlation threshold, a combination of strongly correlated features is screened out. In the feature selection process, the recursive feature elimination method is used to gradually remove redundant and irrelevant features and retain the core feature data that has the greatest impact on function item identification.

[0024] Multi-source data fusion integrates the core feature data from different maintenance links and different acquisition devices. The fusion process adopts a hierarchical fusion strategy, first locally fuses the same type of data, and then globally fuses across types to generate a fusion feature vector. The fused feature vector is cross-validated, and the stability and reliability of the fusion result are evaluated by the K-fold cross-validation method to obtain a verification feature set. In the technical description extraction stage, the verification feature set is deeply analyzed to extract the key technical elements of the functional items, including maintenance object attributes, maintenance operation specifications, quality requirements, etc. Semantic analysis is to structure these technical description elements, identify keywords, semantic relationships, and technical specifications, and form a standardized semantic feature table.

[0025] Budget item mapping establishes a corresponding relationship between the technical elements in the semantic feature table and the cost items to obtain budget-related data. In the cost accounting process, multiple dimensions such as labor cost, material cost, and equipment cost are considered, and the benchmark value of each cost is calculated by combining market prices and historical data. The cost benchmark value is decomposed, and the total cost is decomposed into various sub-items to obtain detailed sub-item cost data. The sub-item cost data is statistically summarized to form a complete budget data set.

[0026] For example, in a refrigeration equipment maintenance project, the original function item feature set includes data such as temperature control accuracy (ranging from ±0.5℃ to ±2℃), maintenance frequency (2-4 times a day), and response time (15-60 minutes). After standardization, these data are uniformly mapped to the 0-1 interval. There are some missing response time records in the data matrix. By analyzing historical data, it is found that the response time is correlated with the fault type and the period of occurrence. Based on this, the conditional average method is used to supplement it. Correlation analysis found that the correlation coefficient between temperature control accuracy and maintenance frequency is 0.85, indicating that the two are highly correlated. Feature selection retains 8 core features including control accuracy, failure rate, and energy consumption indicators. Data fusion integrates equipment operation data (temperature, pressure, flow, etc.), maintenance records (operation items, processing methods, materials used, etc.) and cost data (man-hours, material consumption, energy use, etc.).

[0027] The technical description extraction identifies specific maintenance requirements such as "replace the temperature control valve when the temperature fluctuation exceeds the standard", and the semantic analysis deconstructs it into elements such as "trigger condition: temperature fluctuation exceeds the standard", "operation item: replacement", and "maintenance object: temperature control valve". Budget mapping establishes the correspondence between "replacement of temperature control valve" and the corresponding labor cost (2 working hours × 200 yuan / working hour) and material cost (1500 yuan / temperature control valve). Cost accounting results in a single maintenance cost baseline of 1900 yuan, and the cost is decomposed into 400 yuan for labor and 1500 yuan for materials. Consistency inspection ensures that technical requirements match the budget, such as the cost estimate of high-level temperature control valves (3000 yuan / piece) corresponding to high-precision temperature control requirements (±0.5℃). The final generated functional item technical description and budget preparation data fully reflect the maintenance requirements and resource investment, providing a reliable basis for operation and maintenance management.

[0028] In a specific embodiment, the process of executing step S103 may specifically include the following steps: (1) Extract features from the technical description data of the functional items and the budget preparation data to obtain the technical description feature vector, and normalize the technical description feature vector to obtain standardized feature data; (2) Perform threshold calculation on the standardized feature data to obtain a threshold decision matrix, and mark the abnormal points of the threshold decision matrix to obtain an abnormal feature set; (3) Perform cost deviation analysis on the abnormal feature set to obtain cost difference data, and perform hierarchical decomposition on the cost difference data to obtain a hierarchical difference table; (4) Performing correlation analysis on the hierarchical difference table to obtain a feature correlation map, and extracting diagnostic elements from the feature correlation map to obtain a diagnostic feature set; (5) Mapping the diagnostic feature set to the assessment index to obtain an assessment mapping table, and calibrating the parameters of the assessment mapping table to obtain an assessment parameter set; (6) Perform threshold constraint calculation on the assessment parameter set to obtain a constraint condition set, and perform consistency verification on the constraint condition set to obtain a verification result table; (7) Perform parameter calibration on the verification result table to obtain a calibration data set, and merge the results based on the calibration data set to obtain functional item diagnostic data and assessment requirement data.

[0029] Specifically, the processing of the technical description data of the functional items and the budget preparation data begins with the feature extraction stage. Feature extraction mainly extracts key indicators such as maintenance frequency, response time, and technical parameters to form a multidimensional vector that describes the characteristics of the functional items. These feature vectors are normalized to unify data of different dimensions and ranges to a standard scale, so that the processed data is more comparable and consistent. During the threshold calculation process, a reasonable threshold interval is set for the standardized feature data, and a threshold judgment matrix is ​​constructed by comparing it with the standard threshold. The threshold judgment matrix reflects whether each indicator exceeds the preset range, and then marks the abnormal points to form an abnormal feature set. The abnormal feature set contains various indicators that exceed the normal range and their degree of deviation.

[0030] Cost deviation analysis quantifies the cost of indicator deviations in the abnormal feature set, calculates the cost impact caused by each abnormality, and generates cost difference data. The cost difference data is hierarchically decomposed, and the overall cost difference is decomposed into different levels such as labor cost, material cost, and equipment cost to form a hierarchical difference table to clearly display the cost deviation of each level. In the correlation analysis stage, the correlation relationship between each level in the hierarchical difference table is quantitatively analyzed, and a feature correlation map is constructed to show the mutual influence between different cost items. Diagnostic elements are extracted from the correlation map, including key influencing factors, abnormal transmission paths, cost sensitive points, etc., to form a diagnostic feature set.

[0031] Assessment indicator mapping establishes a corresponding relationship between the diagnostic feature set and the assessment system, generates an assessment mapping table, and clarifies the assessment indicator corresponding to each abnormal feature. Parameter calibration is performed on the assessment mapping table, and the benchmark value, tolerance range, and weight coefficient of each assessment indicator are set to form an assessment parameter set. Threshold constraint calculation sets reasonable constraints on the assessment parameter set, including upper and lower limits of indicators, variation range, correlation requirements, etc., to generate a constraint condition set. The constraint condition set is verified for consistency, and whether there is a conflict between the constraints is checked to form a verification result table.

[0032] For example, in a UPS equipment maintenance project in a data center, the technical description data includes indicators such as maintenance frequency (3 inspections per week), response time (fault response within 15 minutes), and operating parameters (output voltage stability ±1%). The budget preparation data includes labor costs (4,000 yuan / month), spare parts costs (20,000 yuan / year), etc. After feature extraction, a feature vector containing 12 dimensions is obtained, and all data are mapped to the 0-1 interval through normalization. The threshold judgment criteria are set: the maintenance frequency deviation does not exceed ±1 time / week, the response time does not exceed 20 minutes, and the voltage stability does not exceed ±2%. Through threshold judgment, it is found that the maintenance frequency in a certain month dropped to 1 time / week, and the response time was extended to 25 minutes, which was marked as an abnormal point. Cost analysis shows that the reduction in maintenance frequency leads to an increase in equipment failure rate, an increase in maintenance costs of 3,000 yuan, and the extension of response time causes equipment downtime losses of 5,000 yuan.

[0033] Hierarchical decomposition decomposes the total cost difference of 8,000 yuan into direct costs (maintenance costs of 3,000 yuan) and indirect costs (downtime losses of 5,000 yuan). Correlation analysis found that there is a correlation coefficient of 0.8 between maintenance frequency and failure rate, indicating that the two are highly correlated. Diagnostic features include key issues such as insufficient maintenance frequency and delayed response mechanism. Establish assessment indicators: maintenance execution rate (actual frequency / planned frequency), response timeliness rate (number of responses within the specified time / total number of responses), etc., and set the benchmark values ​​​​to 95% and 90% respectively. Constraints include maintenance execution rate not less than 90% and response timeliness rate not less than 85%. Consistency verification ensures that various indicators and constraints are coordinated with each other.

[0034] Through parameter calibration, the maintenance frequency was adjusted to 4 times per week to ensure fault prevention, and the staff scheduling was optimized to shorten the response time to less than 15 minutes, with a corresponding increase in labor costs of 1,000 yuan per month. The final functional item diagnostic data and assessment requirement data provided specific improvement directions and assessment basis for operation and maintenance management.

[0035] In a specific embodiment, the process of executing step S104 may specifically include the following steps: (1) Calculate the feature weights of the functional item diagnostic data to obtain a weight vector, where:

[0036] in, represents the weight vector, represents the number of features, represents the number of samples, represents the eigenvalue of the i-th function item, represents the eigenvalue of the kth function item, Represents the importance coefficient of the jth feature, with a value range of [0,1], represents the mean of the jth feature, represents the standard deviation of the jth feature, represents the natural exponential function; (2) Normalize the weight vector to obtain normalized weight data; (3) Perform recursive iterative calculations on the normalized weight data and assessment requirement data to obtain an iterative result matrix, where:

[0037] in, represents the iteration result matrix, represents the result matrix of the t-th iteration, p represents the dimension of the iteration parameter, e Indicates the number of functional items, r Represents the dimension of the feature, represents the weight coefficient of the kth feature, represents the bias parameter of the kth feature, Represents the learning rate, which is used to control the iteration step size, and its value range is (0,1). Indicates i Function item j eigenvalues, represents the mean of all eigenvalues; (4) Verify the convergence of the iterative result matrix to obtain a stable feature set; (5) Decompose the stable feature set into function points to obtain a function point feature table, and perform quantitative statistics on the function point feature table to obtain initial quantitative data; (6) Check the constraints on the initial quantity data to obtain a valid quantity set, and adjust the parameters of the valid quantity set to obtain an adjustment parameter table; (7) Calculate the cost of the adjustment parameter table to obtain cost estimation data, and perform cost decomposition on the cost estimation data to obtain a sub-item cost set; (8) Perform data verification on the itemized cost set to obtain a verification result table, and perform consistency analysis on the verification result table to obtain consistent data; (9) Integrate the consistency data to obtain the number of function points and function item cost data.

[0038] Specifically, the feature weight calculation process considers multiple factors such as the feature value of the function item, the feature importance coefficient, the feature mean and standard deviation. When calculating the feature weight, first obtain the feature value of each function item, then weight it according to the feature importance coefficient, and consider the distribution characteristics of the data (mean and standard deviation), and obtain the weight vector through exponential operation. This calculation method fully considers the distribution characteristics and importance differences of the data. The weight vector is standardized, and each weight value is adjusted to a uniform scale range to obtain normalized weight data. Standardization ensures the comparability of different feature weights. Recursively iterate the normalized weight data and the assessment requirement data, set a suitable learning rate to control the iteration step, and obtain the iteration result matrix through multiple rounds of iterative optimization. During the iterative calculation process, each iteration will be adjusted in combination with the feature weight coefficient and bias parameter, while considering the overall distribution of all eigenvalues.

[0039] Verify the convergence of the iteration result matrix to determine whether the iteration result has reached a stable state, and extract stable feature combinations to form a stable feature set. The stable feature set is decomposed into function points to decompose complex functional features into basic function points, and the initial function point quantity data is obtained by statistics. The constraint condition test verifies the rationality of the initial quantity data to ensure that the number of function points meets business requirements and resource constraints, and selects the effective quantity set. Adjust the parameters of the effective quantity set, fine-tune each parameter according to the actual situation, and generate an adjustment parameter table. Calculate the cost based on the adjustment parameter table, estimate the implementation cost of each function point, obtain the cost estimation data, and then decompose the total cost into labor costs, material costs, equipment costs and other sub-item costs.

[0040] For example, in the operation and maintenance project of a certain industrial automation production line, the functional item diagnostic data includes 10 feature dimensions such as equipment status monitoring, preventive maintenance, and fault diagnosis. After feature weight calculation, the importance coefficient of equipment status monitoring is 0.9 (highest), preventive maintenance is 0.7, and fault diagnosis is 0.8. After the calculated weight vector is standardized, the weights of the three features are 0.375, 0.292, and 0.333, respectively. The learning rate is set to 0.1 during iterative calculation. After 50 rounds of iteration, the result tends to be stable, and a stable feature combination is obtained. Function point decomposition shows that equipment status monitoring includes three basic function points: temperature monitoring, vibration monitoring, and current monitoring. The initial quantity data is 2 inspections per shift. The constraint condition requires that the machine equipment be inspected at least once every 4 hours, so the number of inspections is adjusted to 3 times per shift.

[0041] Cost calculations show that the labor cost of each function point is 200 yuan / time, the equipment cost is 500 yuan / month, and the consumables cost is 300 yuan / month. After cost decomposition, the monthly sub-item costs are obtained: labor costs are 18,000 yuan (3 times / shift × 3 shifts / day × 30 days × 200 yuan / time), equipment costs are 1,500 yuan (3 monitoring points × 500 yuan / month), and consumables costs are 900 yuan (3 monitoring points × 300 yuan / month), totaling 20,400 yuan / month. Data verification confirms that the cost calculations are reasonable and consistent with industry standards. Consistency analysis shows that the number of function points matches the staffing and equipment capabilities. The final integrated function point quantity and function item cost data provide an accurate quantitative basis for operation and maintenance management.

[0042] In a specific embodiment, the process of executing step S105 may specifically include the following steps: (1) Construct a feature tensor for the function point quantity data to obtain a three-dimensional feature data set, and perform dimension mapping on the function item cost data to obtain a cost mapping table; (2) Perform tensor decomposition on the three-dimensional feature data set to obtain a core feature table, and perform dimensionality reduction processing on the cost mapping table to obtain a reduced dimensionality data set; (3) Perform data fusion on the core feature table and the dimension reduction data set to obtain a fused feature set, and perform functional item clustering on the fused feature set to obtain a functional item category table; (4) Analyze the constraint conditions of the function item category table to obtain a constraint parameter table, and adjust the constraint parameter table to obtain an adjustment coefficient set; (5) Generate an investment plan for the adjustment coefficient set to obtain an investment plan table, and perform cost analysis on the investment plan table to obtain a cost data set; (6) Screen the cost data set to obtain a set of alternative solutions, and conduct risk assessment on the set of alternative solutions to obtain a risk control table; (7) Integrate the data of the risk control table to obtain the functional item adjustment plan and investment control data.

[0043] Specifically, the function point quantity data is constructed into a three-dimensional feature data set according to the function type, time dimension and space dimension. At the same time, the function item cost data is dimensionally mapped to establish the corresponding relationship between the cost and the function point, forming a cost mapping table. The three-dimensional feature data set reflects the distribution characteristics of the function points in different dimensions, and the cost mapping table reflects the cost structure of each function item. Tensor decomposition is a key step in extracting core information from the three-dimensional feature data set. The complex three-dimensional data is converted into a more easily processed two-dimensional form through decomposition operations to obtain a core feature table. The cost mapping table is subjected to dimensionality reduction processing to compress the multi-dimensional cost data into the main dimension to obtain a dimensionality reduction data set. These two data sets retain the key information of the original data while reducing the complexity of data processing.

[0044] In the data fusion stage, the core feature table and the dimension reduction data set are integrated. The fusion process takes into account the relationship between functional features and cost factors to form a fused feature set. Function items are clustered in the fused feature set, and functional items are grouped according to feature similarity to form a functional item category table, which clearly shows the characteristics and cost attributes of functional items of different categories. Constraint condition analysis systematically identifies constraints for the functional item category table, including technical constraints, resource constraints, and cost constraints, and generates a constraint parameter table. The constraint parameter table is adjusted and calculated, and each constraint parameter is optimized according to the actual situation to obtain an adjustment coefficient set. The adjustment coefficient set reflects the feasible solution space under various constraint conditions.

[0045] An investment plan is generated based on the adjustment coefficient set. The plan includes specific resource allocation and cost budget to form an investment plan table. Cost analysis is performed on the investment plan table to calculate the return on investment and cost-effectiveness of each plan to obtain a cost data set. During the plan screening process, the cost data set is evaluated based on indicators such as cost-effectiveness to screen out a set of feasible alternative plans.

[0046] For example, in the operation and maintenance project of the power supply system of a certain industrial park, the data on the number of function points includes three dimensions: function type (such as power monitoring, equipment maintenance, fault handling, etc.), time dimension (monthly, quarterly, annual) and space dimension (substation, distribution room, power equipment, etc.). Function item cost data includes multiple dimensions such as labor cost, equipment cost, material cost, etc. Through three-dimensional feature construction, a data set containing 12 function types, 4 time periods, and 8 spatial nodes was formed. The cost mapping shows that power monitoring costs 800,000 yuan per month, equipment maintenance costs 600,000 yuan, and fault handling costs 400,000 yuan. After tensor decomposition, a two-dimensional table reflecting the core functional characteristics is obtained, which contains 30 key indicators. After the cost data is reduced in dimension, the two main dimensions of labor cost and equipment cost are retained.

[0047] After data fusion, it was found that the power monitoring function and the equipment maintenance function have a high correlation, and cluster analysis classified them into the preventive maintenance category. Constraint analysis shows that the staffing must not be less than 50 people, the equipment investment must not be less than 2 million yuan, and the annual total budget must be controlled within 20 million yuan. According to these constraints, the adjustment coefficient is set: 10% increase in staff and 15% increase in equipment investment. The generated investment plan includes three levels: basic maintenance plan (annual investment of 15 million yuan), standard maintenance plan (annual investment of 18 million yuan) and advanced maintenance plan (annual investment of 20 million yuan). Cost analysis shows that the standard maintenance plan has the highest return on investment, and risk assessment believes that the plan is the most reasonable in terms of technical feasibility and cost control. The final functional item adjustment plan recommends the use of a standard maintenance plan, a maintenance team of 60 people, an equipment investment of 2.3 million yuan, and an annual budget of 18 million yuan. The investment control data clarifies the specific amount and adjustment space of each expenditure, providing a clear guiding basis for the implementation of operation and maintenance projects.

[0048] In a specific embodiment, the process of executing step S106 may specifically include the following steps: (1) Perform diagnostic data standardization on the functional item adjustment plan and investment control data to obtain a standard data set, and construct an indicator system for the standard data set to obtain a diagnostic indicator table; (2) Perform hierarchical decomposition on the diagnostic index table to obtain a hierarchical index set, and assign weights to the hierarchical index set to obtain an index weight table; (3) Perform weighted calculation on the indicator weight table and the standard data set to obtain a comprehensive score table, and perform threshold analysis on the comprehensive score table to obtain an evaluation level set; (4) Generate diagnostic rules for the evaluation level set to obtain a diagnostic rule table, and perform scenario matching on the diagnostic rule table to obtain a matching result set; (5) Identify problems on the matching result set to obtain a problem list, and analyze the causes of the problem list to obtain a cause data set; (6) Analyze the optimization direction of the cause data set to obtain an optimization strategy table, and conduct a feasibility evaluation on the optimization strategy table to obtain an evaluation result set; (7) Merge the evaluation result set to obtain the function point identification report and optimization suggestion data.

[0049] Specifically, the data normalization method is used to convert data of different dimensions into a unified scale, including data of multiple dimensions such as operation and maintenance indicators, cost indicators, and performance indicators. The standardized data set is constructed through the indicator system to form a diagnostic indicator table containing technical indicators, economic indicators, and management indicators. The hierarchical decomposition process of the diagnostic indicator table adopts a top-down hierarchical method to decompose the top-level indicators step by step into specific evaluation indicators to form a hierarchical indicator set. The hierarchical indicators include primary indicators (such as maintenance quality, operating efficiency, and cost control), secondary indicators (such as failure rate, response time, and labor cost), and tertiary indicators (such as equipment failure rate, fault handling time, and maintenance personnel man-hour cost). The hierarchical indicator set is weighted, and the relative importance of each layer of indicators is determined by the hierarchical analysis method to obtain the indicator weight table.

[0050] The indicator weight table is weighted and calculated with the standard data set to obtain the scoring results reflecting the comprehensive performance of each evaluation object, forming a comprehensive scoring table. Threshold analysis is performed on the comprehensive scoring table, and evaluation standards of different levels are set. The evaluation results are divided into excellent, good, qualified, and to be improved levels to form an evaluation level set. According to the evaluation level set, the corresponding diagnostic rules are generated. The rule content includes trigger conditions, judgment criteria, and processing suggestions to form a diagnostic rule table. The diagnostic rule table is matched with scenarios, and the rules are matched with actual operation and maintenance scenarios to obtain a matching result set. The matching result set reflects the applicability and effectiveness of each rule in a specific scenario.

[0051] The problem identification phase analyzes the anomalies and deviations in the matching result set, identifies the key issues that need to be improved, and forms a problem list. The causes of each problem in the problem list are analyzed, and the cause and effect analysis method is used to find the root cause of the problem to obtain the cause data set. The cause data set shows the deep-seated causes and influencing factors of the problem. For example, in the equipment operation and maintenance project of a chemical plant, the functional item adjustment plan includes equipment detection, preventive maintenance, emergency maintenance and other functions, and the investment control data includes an annual budget of 3 million yuan, a staff of 20 people, and an equipment investment of 1 million yuan. After standardization, these data are unified into the 0-1 range for comprehensive evaluation.

[0052] The indicator system is constructed into a three-layer structure: the first-level indicators include operation and maintenance quality (weight 0.4), operation efficiency (weight 0.3), and cost control (weight 0.3); the second-level indicators include 12 indicators such as equipment availability, fault response rate, and maintenance cost rate; the third-level indicators include 36 specific assessment parameters. After weighted calculation, the comprehensive score is 85 points, corresponding to the "good" level. The diagnostic rules generate rules such as "when the equipment availability rate is lower than 90%, the detection frequency needs to be increased" and "when the fault response time exceeds 30 minutes, the staff configuration needs to be optimized". Scenario matching found that the equipment failure rate is high in high temperature environments, and the maintenance cost exceeds the budget by 15%. Problem identification identified three key problems: insufficient equipment detection frequency, unreasonable maintenance personnel configuration, and lagging spare parts inventory management. Cause analysis shows that the insufficient detection frequency is due to the old detection equipment and low detection efficiency; the unreasonable personnel configuration is due to the single skill structure that cannot meet the diversified maintenance needs; the lagging inventory management is due to the lack of intelligent management system and the long replenishment cycle. Based on this, an optimization strategy was formulated: updating testing equipment (investment of 500,000 yuan), optimizing personnel structure (investment of 200,000 yuan in training), and building intelligent warehousing (investment of 300,000 yuan). The evaluation results showed that the optimization plan met the requirements in terms of technical feasibility and economic rationality. It is expected that after implementation, the equipment availability rate can be increased to 95%, the fault response time can be shortened to 15 minutes, and the annual maintenance cost can be reduced by 10%. The final function point identification report records the problem analysis process and improvement suggestions in detail, providing a reliable basis for the continuous optimization of operation and maintenance management.

[0053] The above describes the operation and maintenance function item identification and diagnosis method based on big data in the embodiment of the present application. The following describes the operation and maintenance function item identification and diagnosis device based on big data in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the operation and maintenance function item identification and diagnosis device based on big data includes: A mapping module 201 is used to perform feature decomposition and classification mapping on the maintenance object, maintenance content, maintenance requirements and maintenance level data of the operation and maintenance project to obtain a feature set of operation and maintenance function items and a classification rule for function items; Identification module 202, used to perform data fusion and cross-identification on the operation and maintenance function item feature set to obtain function item technical description data and budget preparation data; The analysis module 203 is used to perform threshold diagnosis and difference analysis on the function item technical description data and the budget preparation data to obtain function item diagnosis data and assessment requirement data; Processing module 204, used for recursively calculating and adjusting the function item diagnosis data and the assessment requirement data to obtain the function point quantity and function item cost data; Decomposition module 205, used for performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plan and investment control data; The evaluation module 206 is used to perform comprehensive diagnosis and evaluation on the functional item adjustment plan and investment control data to obtain a function point identification report and optimization suggestion data.

[0054] Through the collaborative cooperation of the above-mentioned components, by decomposing and classifying the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance projects, a complete operation and maintenance function item feature set and function item classification rules were established, effectively solving the problem of inaccurate function item identification; by fusion and cross-identification of the operation and maintenance function item feature set, the technical description of the function item and the budget preparation data were obtained, achieving the unity of technical requirements and cost estimation; threshold diagnosis and difference analysis were performed on the technical description of the function item and the budget preparation data, the function item diagnosis data and assessment requirement data were obtained, and a scientific diagnosis and evaluation mechanism was established; by recursively calculating and adjusting the function item diagnosis data and assessment requirement data, accurate function point quantity and function item cost data were obtained, achieving precise resource allocation; tensor decomposition and optimization processing were used to conduct in-depth analysis of the function point quantity and function item cost data, generating reasonable function item adjustment plans and investment control data, ensuring the feasibility of the optimization plan; finally, through comprehensive diagnosis and evaluation, function point identification reports and optimization suggestion data were formed, providing a reliable basis for the continuous improvement of the operation and maintenance projects. The entire method forms a complete data processing closed loop, with close logical connections between each step and smooth data flow, ensuring the controllability of the entire process from data collection, processing, analysis to result output, and significantly improving the accuracy, scientificity and practicality of operation and maintenance function item identification and diagnosis.

[0055] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying and diagnosing operation and maintenance function items based on big data, characterized in that: The operation and maintenance function item identification and diagnosis method based on big data includes: Perform feature decomposition and classification mapping on the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance projects to obtain the feature set of operation and maintenance function items and the classification rules of function items; Performing data fusion and cross-identification on the feature set of the operation and maintenance function items to obtain technical description data of the function items and budget preparation data; Perform threshold diagnosis and difference analysis on the technical description data of the functional items and the budget preparation data to obtain functional item diagnosis data and assessment requirement data; Recursively calculate and adjust the function item diagnosis data and assessment requirement data to obtain function point quantity and function item cost data; Performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plans and investment control data; Comprehensive diagnosis and evaluation are performed on the functional item adjustment plan and investment control data to obtain a functional point identification report and optimization suggestion data.

2. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1 is characterized in that: The feature decomposition and classification mapping of the maintenance object, maintenance content, maintenance requirements and maintenance level data of the operation and maintenance project to obtain the operation and maintenance function item feature set and function item classification rules include: Performing data cleaning and standardization on the original data of the operation and maintenance project to obtain standardized maintenance data, and segmenting the standardized maintenance data into time windows to obtain time series feature data; Performing multi-scale decomposition on the time series feature data through wavelet transform to obtain a maintenance object feature vector, and performing frequency domain feature extraction on the maintenance object feature vector to obtain frequency domain feature data; Calculating the degree of association between the frequency domain feature data and the maintenance content data to obtain a feature association matrix, and performing hierarchical clustering on the feature association matrix to obtain a functional item hierarchy structure; Performing semantic analysis on the function item hierarchy and maintenance requirement data to obtain a semantic feature set, and performing attribute mapping on the semantic feature set to obtain a function item attribute table; Extracting rules from the function item attribute table and maintenance level data to obtain a rule set, and verifying and screening the rule set to obtain a candidate rule base; Performing weight calculation on the candidate rule base to obtain a weighted rule set, and merging rules on the weighted rule set to obtain a target classification rule; The target classification rules are cross-validated to obtain verification result data, and the classification rules are modified according to the verification result data to obtain an operation and maintenance function item feature set and a function item classification rule.

3. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1, characterized in that: The data fusion and cross-identification of the operation and maintenance function item feature set are performed to obtain function item technical description data and budget preparation data, including: Performing data standardization processing on the operation and maintenance function item feature set to obtain a standardized function item data matrix, and completing missing values ​​on the standardized function item data matrix to obtain a complete feature matrix; Performing correlation analysis on the complete feature matrix to obtain a feature correlation coefficient matrix, and performing feature selection on the feature correlation coefficient matrix to obtain core feature data; Performing multi-source data fusion on the core feature data to obtain a fused feature vector, and cross-validating the fused feature vector to obtain a validation feature set; Extracting technical descriptions from the verification feature set to obtain technical description elements, and performing semantic analysis on the technical description elements to obtain a semantic feature table; Performing budget item mapping on the semantic feature table to obtain budget-related data, and performing cost accounting on the budget-related data to obtain a cost benchmark value; Performing cost decomposition on the cost benchmark value to obtain itemized cost data, and combining and statistically analyzing the itemized cost data to obtain a budget data set; The budget data set is subjected to consistency check to obtain check result data, and data calibration is performed based on the check result data to obtain functional item technical description data and budget preparation data.

4. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1, characterized in that: The threshold diagnosis and difference analysis of the function item technical description data and the budget preparation data are performed to obtain function item diagnosis data and assessment requirement data, including: Extracting features from the functional item technical description data and the budget preparation data to obtain a technical description feature vector, and normalizing the technical description feature vector to obtain standardized feature data; Performing threshold calculation on the standardized feature data to obtain a threshold determination matrix, and marking abnormal points on the threshold determination matrix to obtain an abnormal feature set; Performing cost deviation analysis on the abnormal feature set to obtain cost difference data, and performing hierarchical decomposition on the cost difference data to obtain a hierarchical difference table; Performing correlation analysis on the hierarchical difference table to obtain a feature correlation map, and extracting diagnostic elements from the feature correlation map to obtain a diagnostic feature set; Performing assessment index mapping on the diagnostic feature set to obtain an assessment mapping table, and performing parameter calibration on the assessment mapping table to obtain an assessment parameter set; Performing threshold constraint calculation on the assessment parameter set to obtain a constraint condition set, and performing consistency verification on the constraint condition set to obtain a verification result table; Parameter calibration is performed on the verification result table to obtain a calibration data set, and results are merged based on the calibration data set to obtain functional item diagnosis data and assessment requirement data.

5. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1, characterized in that: The recursive calculation and adjustment processing of the function item diagnosis data and the assessment requirement data to obtain the function point quantity and function item cost data includes: The feature weights of the functional item diagnostic data are calculated to obtain a weight vector, where: ， in, represents the weight vector, represents the number of features, represents the number of samples, represents the eigenvalue of the i-th function item, represents the eigenvalue of the kth function item, Represents the importance coefficient of the jth feature, with a value range of [0,1], represents the mean of the jth feature, represents the standard deviation of the jth feature, represents the natural exponential function; Performing normalization processing on the weight vector to obtain normalized weight data; The normalized weight data and the assessment requirement data are recursively iterated to obtain an iterative result matrix, where: ， in, represents the iteration result matrix, represents the result matrix of the t-th iteration, p represents the dimension of the iteration parameter, e Indicates the number of functional items, r Represents the dimension of the feature, represents the weight coefficient of the kth feature, represents the bias parameter of the kth feature, Represents the learning rate, which is used to control the iteration step size, and its value range is (0,1). Indicates i Function item j eigenvalues, represents the mean of all eigenvalues; Verifying the convergence of the iterative result matrix to obtain a stable feature set; Decomposing the stable feature set into function points to obtain a function point feature table, and performing quantitative statistics on the function point feature table to obtain initial quantitative data; Performing constraint condition check on the initial quantity data to obtain a valid quantity set, and performing parameter adjustment on the valid quantity set to obtain an adjustment parameter table; Performing cost calculation on the adjustment parameter table to obtain cost estimation data, and performing cost decomposition on the cost estimation data to obtain a sub-item cost set; Performing data verification on the itemized cost set to obtain a verification result table, and performing consistency analysis on the verification result table to obtain consistent data; The consistency data is integrated to obtain the function point quantity and function item cost data.

6. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1, characterized in that: The tensor decomposition and optimization processing of the function point quantity and function item cost data to obtain the function item adjustment plan and investment control data includes: Constructing a feature tensor for the function point quantity data to obtain a three-dimensional feature data set, and performing dimension mapping on the function item cost data to obtain a cost mapping table; Performing tensor decomposition on the three-dimensional feature data set to obtain a core feature table, and performing dimensionality reduction processing on the cost mapping table to obtain a dimensionality reduction data set; Performing data fusion on the core feature table and the dimension reduction data set to obtain a fused feature set, and performing function item clustering on the fused feature set to obtain a function item category table; Performing constraint condition analysis on the function item category table to obtain a constraint parameter table, and performing adjustment calculation on the constraint parameter table to obtain an adjustment coefficient set; Generating an investment plan for the adjustment coefficient set to obtain an investment plan table, and performing cost analysis on the investment plan table to obtain a cost data set; Screening the cost data set to obtain an alternative solution set, and performing risk assessment on the alternative solution set to obtain a risk control table; The risk control table is integrated to obtain a functional item adjustment plan and investment control data.

7. The method for identifying and diagnosing operation and maintenance function items based on big data according to claim 1, characterized in that: The comprehensive diagnosis and evaluation of the functional item adjustment plan and investment control data to obtain a functional point identification report and optimization suggestion data include: Performing diagnostic data standardization processing on the functional item adjustment plan and investment control data to obtain a standard data set, and constructing an indicator system for the standard data set to obtain a diagnostic indicator table; Decomposing the diagnostic index table hierarchically to obtain a hierarchical index set, and assigning weights to the hierarchical index set to obtain an index weight table; Performing weighted calculation on the indicator weight table and the standard data set to obtain a comprehensive score table, and performing threshold analysis on the comprehensive score table to obtain an evaluation grade set; Generating diagnostic rules for the evaluation level set to obtain a diagnostic rule table, and performing scene matching on the diagnostic rule table to obtain a matching result set; Identify problems on the matching result set to obtain a problem list, and analyze causes of the problem list to obtain a cause data set; Performing optimization direction analysis on the cause data set to obtain an optimization strategy table, and performing feasibility evaluation on the optimization strategy table to obtain an evaluation result set; The evaluation result set is merged to obtain a function point identification report and optimization suggestion data.

8. A device for identifying and diagnosing operation and maintenance function items based on big data, used to implement the method for identifying and diagnosing operation and maintenance function items based on big data as described in any one of claims 1 to 7, characterized in that: The operation and maintenance function item identification and diagnosis device based on big data includes: A mapping module is used to perform feature decomposition and classification mapping on the maintenance objects, maintenance contents, maintenance requirements and maintenance level data of the operation and maintenance project to obtain the feature set of operation and maintenance function items and the classification rules of function items; An identification module is used to perform data fusion and cross-identification on the feature set of the operation and maintenance function items to obtain technical description data of the function items and budget preparation data; An analysis module, used to perform threshold diagnosis and difference analysis on the functional item technical description data and the budget preparation data, to obtain functional item diagnosis data and assessment requirement data; A processing module, used for recursively calculating and adjusting the function item diagnosis data and the assessment requirement data to obtain the function point quantity and function item cost data; A decomposition module, used for performing tensor decomposition and optimization processing on the function point quantity and function item cost data to obtain function item adjustment plans and investment control data; The evaluation module is used to perform comprehensive diagnosis and evaluation on the functional item adjustment plan and investment control data to obtain a function point identification report and optimization suggestion data.

Images