Big data-based equipment maintenance intelligent technology architecture construction method

Abstract

The present application relates to the field of equipment maintenance intelligent technology, in particular to an equipment maintenance intelligent technology architecture construction method based on big data, comprising the following steps: constructing a multidimensional data processing platform; constructing a mechanics discipline simulation operation platform; constructing a prediction and state evaluation algorithm platform; constructing an equipment maintenance intelligent control center; and constructing a maintenance decision system. The present application provides an equipment maintenance intelligent technology architecture construction method based on big data, accurately evaluates the health state of the equipment, and quantifies the health state to obtain an evaluation result set integrated / associated / encapsulated on a principle model unit node, supports real-time simulation verification results to be automatically backfilled according to classified nodes. The present application can intelligently analyze and judge the equipment maintenance quality, plan the maintenance support resources, intelligently decide and push the best maintenance scheme, and greatly improve the accuracy of equipment fault judgment and the timeliness of maintenance.

Description

Technical Field

[0001] This invention relates to the field of intelligent equipment maintenance technology, specifically to a method for constructing an intelligent equipment maintenance technology architecture based on big data. Background Technology

[0002] In December 2018, with the support of the National Natural Science Foundation of China and led by Harbin Institute of Technology, a seminar on "Opportunities and Challenges of Mechanics Driven by Big Data" was held in Shenzhen. The conference invited experts from interdisciplinary fields such as mechanics, mechanics informatics, materials genome engineering, equipment structural performance testing, structural health monitoring, dynamic systems and control, computational mechanics, experimental mechanics, multi-field coupled mechanics, manned spaceflight, and philosophy of science to report on the application progress of big data methods and intelligent methods. The conference generally agreed that the intersection of big data and mechanics will create new growth points for academic directions.

[0003] Based on big data in equipment maintenance, a simulation platform for mechanics is constructed using cross-disciplinary mechanics models. This platform is linked to powerful algorithm mechanisms and combines standard criteria and intelligent computing technology to accurately evaluate the health status of equipment. The quantified health status results are integrated, correlated, and encapsulated on the principle model unit nodes. Real-time simulation verification results are automatically backfilled according to classification nodes. At the same time, historical data is compared and analyzed to intelligently evaluate the quality of equipment maintenance and push the best maintenance plan. This forms an intelligent technical architecture for equipment maintenance based on big data, which is of great significance. Summary of the Invention

[0004] To address the aforementioned technical challenges, this invention proposes a method for constructing an intelligent technical architecture for equipment maintenance based on big data. By sequentially establishing models for each level of equipment maintenance, constructing a simulation platform integrating interdisciplinary mechanics, acquiring multidimensional data, and linking it with powerful algorithm mechanisms, the invention enables intelligent analysis and evaluation of equipment maintenance quality, comprehensive planning of maintenance support resources, intelligent decision-making, and the delivery of optimal maintenance solutions, thereby improving the accuracy of equipment fault diagnosis and the timeliness of maintenance.

[0005] The technical problem to be solved by this invention is achieved by the following technical solution:

[0006] The method for constructing an intelligent technology architecture for equipment maintenance based on big data includes the following steps:

[0007] Step (1) Construct a multidimensional data processing platform:

[0008] Step (11) Select a data-driven processing method based on performance variables;

[0009] Step (12) Select a data-driven processing method based on degradation state identification;

[0010] Step (13) integrates steps (11) and (12), the data-driven model and the physical model to achieve continuous analysis, optimization and correction of the simulation model of the entire maintenance and support process;

[0011] Step (2) Constructing a simulation platform for mechanics:

[0012] Step (21) Analyze typical operating conditions of the equipment and construct a geometric model of the entire equipment;

[0013] Step (22) Based on the analysis results of typical working conditions of the equipment, identify the key components and parts of the equipment and construct their geometric models;

[0014] Step (23) Create a static strength simulation verification method for key structural components of the equipment;

[0015] Step (24) Create a fatigue life simulation verification method for key structural components of equipment;

[0016] Step (25) Create a simulation verification method for the vibration intensity of key structural components of the equipment;

[0017] Step (26) Create a performance degradation analysis verification model under the working condition load;

[0018] Step (27) Develop and package the simulation and comprehensive computing software, and formulate the development and packaging process;

[0019] Step (28) Interdisciplinary simulation calculation and verification of mechanics to provide a control scheme for the intelligent maintenance technology architecture;

[0020] Step (3) Construct a prediction and state evaluation algorithm platform:

[0021] Step (31) encapsulates the diagnostic and prediction algorithms using a configuration module;

[0022] Step (32) integrates the equipment failure probability prediction method, the equipment performance degradation failure time prediction method, the system reliability and risk prediction method, and the system risk assessment output compliance judgment method.

[0023] Step (33) selects a reasonable algorithm by matching matrix, and based on steps (1) and (2), realizes the identification of equipment status and the estimation of service life;

[0024] Step (4) Construct an intelligent control center for equipment maintenance:

[0025] Step (41) Construct an equipment maintenance simulation verification and health assessment platform;

[0026] Step (42) Construct a health baseline based on the solver model, extract performance degradation features, and construct the feature space of the equipment in a healthy state as the "health baseline" for health assessment;

[0027] Step (43) Import a multi-dimensional database, embed a simulation software library, and build an intelligent model for the entire maintenance and support process;

[0028] Step (5) Construct a maintenance decision system:

[0029] Step (51) Based on the idea of ​​the analytic hierarchy process, establish the hierarchical models of system maintenance in sequence;

[0030] Step (52) uses multidimensional data processing as input, cross-disciplinary mechanics simulation as control, correlation prediction and state assessment algorithm mechanism, and equipment maintenance intelligent control center and maintenance decision output to output compliance evaluation results.

[0031] Preferably, the specific process of step (22) is as follows: construct the equipment model structure tree, customize and develop the model parsing tool to obtain the overall equipment model, and extract the models of each key component of the equipment according to the model structure tree, and determine the "pre-processing" solution of the three-dimensional model in the comprehensive simulation.

[0032] Preferably, step (23) is as follows: a three-dimensional model is established based on the design finalized drawings. Under the same ideal standard three-dimensional model, different boundary conditions that conform to the actual situation are set according to the maintenance characteristics to carry out static strength simulation verification.

[0033] Preferably, step (24) is as follows: the fatigue stress amplitude of the evaluation part is obtained through structural finite element simulation analysis. Based on the material properties, fatigue stress spectrum and structural design parameters, the fatigue limit of the structure can be given, thereby obtaining the fatigue safety margin of the equipment and components, and finally obtaining the evaluation result of the fatigue life level of the structure.

[0034] Preferably, step (25) is specifically performed as follows: using the modal analysis module in the simulation software (ANSYS Workbench) to perform finite element modal analysis on the equipment, setting the material parameters of the parts, discretizing the computational entity model, applying fixed installation constraints, and calculating the natural frequencies and mode shapes of the equipment.

[0035] Preferably, step (26) specifically involves: analyzing the performance degradation law of key components of the equipment under the coupling of multiple fields such as flow field, pressure field, temperature field, and stress field under load, using stochastic process to characterize the equipment failure occurrence and development process, obtaining simulation analysis data, and laying the foundation for the health status characterization of the equipment.

[0036] Preferably, the specific process of step (27) is as follows: secondary development is carried out using finite element analysis software as the platform, customized finite element analysis automated analysis script is customized, the script is encapsulated using program software, and batch processing calls the platform batch processing command are performed.

[0037] Preferably, the prediction and state assessment algorithm platform in step (3) includes a user layer, a user interface layer, a system interface layer, a data processing layer, and a model management layer.

[0038] Preferably, in step (42), the "health baseline" is used as the evaluation benchmark. The distance between the feature points of the data to be evaluated and the "health baseline" is calculated and represented by the Mahalanobis distance (MD). The distance is then normalized to a CV value to represent the current health status of the equipment. The normalization calculation formula for the Mahalanobis distance is as follows:

[0039]

[0040] In the formula, a is the normalization parameter; CV is a value between 0 and 1.

[0041] Preferably, the maintenance decision system in step (5) includes maintenance plan, maintenance scheme, maintenance task, maintenance implementation and results.

[0042] The beneficial effects of this invention are:

[0043] This invention provides a method for constructing an intelligent technical architecture for equipment maintenance based on big data. It organically organizes various maintenance processes and multi-dimensional data resources. Based on maintenance big data, it uses cross-disciplinary simulation calculations of various levels of equipment maintenance models, coupled with powerful algorithm mechanisms, to accurately evaluate the health status of the equipment. The quantified health status assessment results are integrated / associated / encapsulated on the principle model unit nodes, supporting real-time simulation verification results automatically backfilled according to classification nodes. It can intelligently analyze and evaluate the quality of equipment maintenance, coordinate and plan maintenance support resources, make intelligent decisions, and push the best maintenance solutions, greatly improving the accuracy of equipment fault diagnosis and the timeliness of maintenance. It has broad application prospects and significant military, economic, and social benefits. Attached Figure Description

[0044] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0045] Figure 1 This is a schematic diagram of the structure of the present invention;

[0046] Figure 2 for Figure 1 (a) Schematic diagram;

[0047] Figure 3 for Figure 1 Schematic diagram (b)

[0048] Figure 4 for Figure 1 Schematic diagram (c);

[0049] Figure 5This is a diagram showing the overall architecture of the prediction and state evaluation algorithm platform in this invention.

[0050] Figure 6 for Figure 1 Schematic diagram (d)

[0051] Figure 7 This is a software implementation architecture diagram of the equipment maintenance simulation verification and health assessment platform in this invention;

[0052] Figure 8 This is a flowchart illustrating the equipment health assessment implementation process based on the health baseline and Mahalanobis distance in this invention.

[0053] Figure 9 for Figure 1 Schematic diagram (e) in the middle. Detailed Implementation

[0054] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0055] A method for constructing an intelligent technology architecture for equipment maintenance based on big data, such as Figures 1 to 4 , Figure 6 , Figure 9 As shown, this includes methods for constructing multi-dimensional data processing platforms, simulation and computation platforms for mechanics, prediction and state assessment algorithm platforms, intelligent control centers for equipment maintenance, and maintenance decision-making systems.

[0056] The specific steps are as follows:

[0057] Step (1) Construct a multidimensional data processing platform.

[0058] Step (11) selects performance parameters with wear characteristics from the performance test data as analysis samples for product performance degradation trends. Taking into account equipment quality information data, health factors are extracted. Combined with expert experience and knowledge data, the distribution characteristics of performance are analyzed, and the overall performance degradation trend is inferred based on the sample distribution to predict the remaining useful life (RUL). This is a prediction method described using direct data-driven methods, that is, a data-driven processing method based on performance variables.

[0059] Step (12) Equipment performance is usually related to the structure and fit of parts. Select the fit dimensions of parts that are related to the structure and can represent the changes in size wear under long-term use conditions (i.e., structural size detection data), as well as typical fault technical information data. Combine expert experience and knowledge data to describe the indirect data-driven prediction method that reflects the nonlinearity and instability of failure mode evolution, i.e., the data-driven processing method based on degradation state identification.

[0060] Step (13) integrates steps (11) and (12), the data-driven model, and the physical model, including available online data and historical offline data, as well as technical quality information data and expert experience knowledge data, to carry out optimization iterations, including noise reduction, smoothing, missing value supplementation, and training of data-driven prediction algorithms. This enables continuous analysis, optimization, and correction of the simulation model for the entire maintenance and support process.

[0061] Step (2) Construct a simulation platform for mechanics.

[0062] Step (21) Analyze typical operating conditions of the equipment and construct a geometric model of the entire equipment.

[0063] Step (22) Based on the analysis results of typical working conditions of the equipment, identify the key components and parts of the equipment and construct their geometric models.

[0064] Perform steps 21 and 22 to construct the equipment model structure tree, customize and develop a model parsing tool to obtain the overall equipment model, and extract the models of each key component of the equipment based on the model structure tree. Determine the "pre-processing" solution for the 3D model during comprehensive simulation, namely model construction, key dimension definition, material definition and assignment, etc.

[0065] Step (23) Create a static strength simulation verification method for key structural components of equipment.

[0066] First, establish a three-dimensional model (ideal standard model) based on the finalized design drawings, i.e., simulation verification under design parameter conditions; second, consider the wear and tear of components during the equipment's service life and the changes in structural dimensions caused by maintenance, i.e., under the same ideal standard three-dimensional model, set different boundary conditions that conform to the actual situation according to maintenance characteristics to carry out static strength simulation verification.

[0067] Step (24) Create a fatigue life simulation verification method for key structural components of equipment.

[0068] Based on the static strength assessment, fatigue strength assessment of key structural components of the equipment is carried out. The fatigue stress amplitude of the assessment area is obtained through finite element simulation analysis of the structure. According to the material properties, fatigue stress spectrum and structural design parameters, the fatigue limit of the structure can be given, thereby obtaining the fatigue safety margin of the equipment and components, and finally obtaining the assessment result of the fatigue life level of the structure.

[0069] Step (25) Create a simulation verification method for the vibration intensity of key structural components of the equipment.

[0070] When a product's natural frequency is close to its operating frequency, it can trigger resonance, leading to structural damage and jeopardizing operational safety. Specifically, the modal analysis module in simulation software (ANSYS Workbench) is used to perform finite element modal analysis on the equipment. Material parameters for the parts are set, the computational solid model is discretized, and fixed installation constraints are applied. The natural frequencies and mode shapes of the equipment are then calculated. The vibration excitation is based on the durability vibration spectrum proposed during equipment design (or the vibration frequencies or mode shapes detected by sensors in the actual system environment, or the vibration frequencies or mode shapes interpreted by flight parameters). This vibration excitation is transmitted from the constraints at the fixed lugs of the equipment to the entire product, thus analyzing the vibration excitation response perpendicular to each axis of the equipment. As the excitation vibration frequency gets closer to the equipment's natural frequency (or resonant frequency), the amplitude of the response increases. When the frequencies are equal, the resonance amplitude reaches its maximum, resulting in the greatest deformation, stress, and destructive force. This allows for the verification of the equipment's vibration intensity.

[0071] Explanation: If a product or system has n degrees of freedom, it also has n eigenvalues, which are described by the deformation shape of the structure in terms of the nth modal vibration, i.e., vibration modes (also called mode shapes). Eigenvalues ​​and structural vibration modes describe the vibration characteristics and frequency features of the structure under free vibration. The mode shapes and frequencies obtained through modal decomposition can easily yield the response of any linear structure. Moreover, in practical problems, it is usually sufficient to consider only the first few mode shapes to obtain a fairly accurate solution. Complex vibrations can generally be decomposed into combinations of simple vibrations, and are independent of the pattern of external excitation, depending only on the properties and constraints of the object itself.

[0072] Step (26) Create a performance degradation analysis verification model under working load.

[0073] This study analyzes the performance degradation patterns of key equipment components under load, including fatigue crack degradation, aging degradation of seals, and erosion degradation of valve cores, through the coupling of multiple fields such as flow field, pressure field, temperature field, and stress field. The study employs stochastic processes to characterize the occurrence and development of equipment failures, obtaining simulation analysis data to lay the foundation for equipment health status characterization.

[0074] Step (27) Develop and package the simulation and comprehensive computing software and formulate the development and packaging process.

[0075] By leveraging finite element analysis software (such as ANSYS Workbench) as a platform for secondary development, customized automated finite element analysis scripts are created. These scripts are then encapsulated using programming software (such as Python) and batch-processed to call platform (such as Workbench) batch commands. This results in a user-friendly, convenient, and easy-to-use human-computer interface. The interface is simple to use, enabling rapid and accurate parametric modeling of structures, precise algorithm calls via hyperlinks for finite element analysis, and automatic generation of analysis results for key structures. This transforms finite element analysis from a professional CAE analyst activity into a simple and efficient process.

[0076] Step (28) considers the prediction conclusion of step (1), performs simulation calculations and verifications in the interdisciplinary field of mechanics, and provides a control scheme for the intelligent maintenance technology architecture.

[0077] Step (3) Construct a prediction and state assessment algorithm platform.

[0078] Step (31) encapsulates the diagnostic and prediction algorithms using a configuration module.

[0079] Step (32) integrates the equipment failure probability prediction method, the equipment performance degradation failure time prediction method, the system reliability and risk prediction method, and the system risk assessment output compliance judgment method.

[0080] Step (33) selects a reasonable algorithm by matching matrix and, based on steps (1) and (2), realizes the identification of equipment status and the estimation of service life.

[0081] The overall architecture of the prediction and state assessment algorithm platform is as follows: Figure 5 As shown, there are five layers: the user layer, the user interface layer, the system interface layer, the data processing layer, and the model management layer. The user layer provides a user-friendly window environment; the user interface layer provides interfaces for users and the external devices under test; the system interface layer provides interfaces for future system modeling and external device interface management; the data processing layer contains various diagnostic and prediction software modules; and the model management layer manages various system models and databases. This overall architecture design professionalizes and localizes the most advanced simulation software and big data analysis software from both domestic and international sources, integrates and seamlessly connects various professional field computational analysis software, and constructs wizard-driven simulation analysis templates. Specific methods include:

[0082] A hierarchical recursive model solution method is employed. The platform decomposes large systems into system-level, subsystem-level, component-level, and part-level structures, facilitating the combination and decomposition of models at different levels. The use of hierarchical recursive techniques for solving models at different levels effectively reduces the difficulty of developing fault diagnosis and prediction models for large systems and improves the computational efficiency of complex models.

[0083] An open system model library construction method is employed. The platform effectively encapsulates domestically and internationally validated fault diagnosis models for basic components, forming a fault diagnosis and prediction model library. Large-scale system models can be easily constructed using the basic component models in the library. Import templates and interfaces for these models are also provided, allowing designers to easily expand the model library.

[0084] The model is easy to port. The platform can export the developed and tested fault diagnosis and prediction model as a model file and provides a dynamic link library, which makes it easy to port the model to systems developed in other programming languages ​​(such as C).

[0085] The overall architecture also includes the platform operating environment, with server configurations no less than: CPU model: Xeon-E5 2.2GHz; storage space no less than 2Tb; storage capacity: 32GB DDR3; operating system: Win7.

[0086] Step (4) Construct an intelligent control center for equipment maintenance.

[0087] Step (41) Construct an equipment maintenance simulation verification and health assessment platform.

[0088] The equipment geometric model and key parameters established in a platform software environment (such as Solidworks) are synchronized to a finite element simulation verification environment (such as Comsol Multiphysics). Boundary conditions for finite element simulation verification are determined through working condition analysis. The main simulation results are displayed in the simulation integration platform display window. Simultaneously, a system modeling window is added to realize boundary simulation functions for finalized drawings and real-time scenarios. The software implementation architecture is as follows: Figure 7 As shown.

[0089] Step (42) Construct a health baseline based on the solver model, extract performance degradation features, and construct the feature space of the equipment in a healthy state as the "health baseline" for health assessment.

[0090] The algorithm is as follows: Residual data of the equipment system in a healthy state is acquired, and time-domain features are extracted from the residual data. A Mahalanobis space is constructed using the feature points in the healthy state, which serves as the "health baseline" of the equipment system under that operating condition. Using the "health baseline" as the evaluation benchmark, the distance between the feature points of the data to be evaluated and the "health baseline" is calculated, represented by the Mahalanobis distance (MD), and normalized to a CV value to represent the current health state of the equipment. The equipment's health level is obtained by measuring the degree to which its health state deviates from the health baseline. For each health assessment, only one calculation is needed to achieve real-time evaluation. Finally, the CV curve is formed from the results of each individual assessment to describe the health performance degradation trend of the equipment system over a period of time. The implementation process of equipment health level evaluation based on the health baseline and Mahalanobis distance is as follows: Figure 8 As shown.

[0091] The normalized formula for Mahalanobis distance is as follows:

[0092]

[0093] In the formula, a is a normalization parameter. Adjusting the value of parameter a will change the degradation sensitivity of the CV value to different stages of use.

[0094] CV is a value between 0 and 1. A higher CV value indicates better system performance, while a lower CV value indicates severe performance degradation.

[0095] Note: Mahalanobis distance is essentially a method for mapping multidimensional space to low-dimensional space. Proposed by Indian statistician Mahalanobis, it is a covariance distance that measures the distance from a sample point to the sample space. As a statistical distance, Mahalanobis distance considers the intrinsic relationships between sample points and is relatively simple and fast to calculate, unaffected by dimensions.

[0096] Step (43) Import a multi-dimensional database, embed a simulation software library, and construct an intelligent model for the entire maintenance and support process. Connect with a powerful algorithm mechanism, conduct big data regression analysis for key stress-bearing components, and integrate and fuse different methods such as cross-disciplinary mechanics model simulation and multi-field coupling simulation calculations. Utilize matching matrix switching to adapt to the selection, matching, calling, and management of evaluation models under varying working conditions, as well as to comprehensively plan maintenance and support resources.

[0097] Step (5) Build a maintenance decision system.

[0098] Step (51) Based on the idea of ​​the analytic hierarchy process, establish the hierarchical models of system maintenance in sequence.

[0099] Step (52) uses multidimensional data processing as input, cross-disciplinary mechanics simulation as control, correlation prediction and state assessment algorithm mechanism, and equipment maintenance intelligent control center and maintenance decision output to output compliance evaluation results.

[0100] The output variable is defined as Y, which represents the result of the equipment maintenance system and mechanism. This part is summarized as follows: maintenance plan (Y1), maintenance scheme (Y2), maintenance task (Y3), maintenance implementation and results (Y4), etc. It includes the following specific steps:

[0101] Based on maintenance needs, the maintenance plan (Y1) takes into account factors such as maintenance inputs and maintenance control. After information analysis, processing and handling, it determines the maintenance object, maintenance content, maintenance level, maintenance time period, and assesses the feasibility of the required maintenance personnel support, supply and consumption of materials, availability of required equipment and tools, allocation of required funds, reserve of maintenance technology, and constraints of environmental laws and regulations.

[0102] The maintenance plan (Y2) is a detailed design of the maintenance activities based on the maintenance plan, with the core being the configuration of maintenance technologies. Through comprehensive consideration of maintenance inputs, mechanisms, and controls, and in practice, factors such as maintenance time, methods, costs, and funding are taken into account, resulting in a series of maintenance plans. Then, the best feasible plan is selected from these.

[0103] Maintenance task (Y3) refers to the allocation and scheduling of human resources for maintenance activities under the constraints of maintenance control. In other words, assigning a maintenance task involves determining which personnel will implement a specific maintenance plan.

[0104] Maintenance Implementation and Results (Y4): Based on the aforementioned maintenance tasks, maintenance personnel perform equipment maintenance according to different time sequences to achieve the expected results. This section places great emphasis on maintenance process control, such as maintenance skills training, material requisition, equipment and tool allocation, and work execution. Through hierarchical maintenance characteristic simulation verification, the simulation model of the entire maintenance support process is continuously analyzed, optimized, and corrected to achieve a continuous optimization decision-making system with the goal of minimizing maintenance costs and downtime losses.

[0105] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely prisms of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing an intelligent technical architecture for equipment maintenance based on big data, characterized by: Includes the following steps: Step (1) Construct a multidimensional data processing platform: Step (11) Select a data-driven processing method based on performance variables; Step (12) Select a data-driven processing method based on degradation state identification; Step (13) integrates steps (11) and (12), the data-driven model and the physical model to achieve continuous analysis, optimization and correction of the simulation model of the entire maintenance and support process; Step (2) Constructing a simulation platform for mechanics: Step (21) Analyze typical operating conditions of the equipment and construct a geometric model of the entire equipment; Step (22) Based on the analysis results of typical operating conditions of the equipment, identify the key components and parts of the equipment, and construct their geometric models: The specific process is as follows: construct an equipment model structure tree, customize and develop a model parsing tool to obtain the overall equipment model, and extract the models of each key component of the equipment based on the model structure tree to determine the "pre-processing" solution for the 3D model during comprehensive simulation. Step (23) Create a static strength simulation verification method for key structural components of the equipment: The specific process is as follows: a three-dimensional model is established based on the design drawings. Under the same ideal standard three-dimensional model, different boundary conditions that conform to the actual situation are set according to the maintenance characteristics to carry out static strength simulation verification. Step (24) Create a fatigue life simulation verification method for key structural components of the equipment: The specific process is as follows: the fatigue stress amplitude of the evaluation part is obtained through structural finite element simulation analysis. Based on the material properties, fatigue stress spectrum and structural design parameters, the fatigue limit of the structure can be given, thereby obtaining the fatigue safety margin of the equipment and components, and finally obtaining the evaluation result of the fatigue life level of the structure. Step (25) Create a simulation verification method for the vibration intensity of key structural components of the equipment: The specific process is as follows: use the modal analysis module in the simulation software ANSYS Workbench to perform finite element modal analysis on the equipment, set the material parameters of the parts, discretize the computational solid model, apply fixed installation constraints, and obtain the natural frequencies and mode shapes of the equipment through calculation. Step (26) Create a performance degradation analysis verification model under the load condition: The specific process is as follows: analyze the performance degradation law of key components of equipment under the multi-field coupling of flow field, pressure field, temperature field and stress field under load, use stochastic process to characterize the failure occurrence and development process of equipment, obtain simulation analysis data, and lay the foundation for equipment health status characterization. Step (27) Develop and package the simulation and comprehensive computing software, and formulate the development and packaging process: The specific process is as follows: secondary development is carried out on the platform using finite element analysis software, customized automated analysis scripts for finite element analysis are created, the scripts are packaged using program software, and batch processing calls are made to the platform's batch processing commands. Step (28) Interdisciplinary simulation calculation and verification of mechanics to provide a control scheme for the intelligent maintenance technology architecture; Step (3) Construct a prediction and state evaluation algorithm platform: Step (31) encapsulates the diagnostic and prediction algorithms using a configuration module; Step (32) integrates the equipment failure probability prediction method, the equipment performance degradation failure time prediction method, the system reliability and risk prediction method, and the system risk assessment output compliance judgment method. Step (33) selects a reasonable algorithm by matching matrix, and based on steps (1) and (2), realizes the identification of equipment status and the estimation of service life; Step (4) Construct an intelligent control center for equipment maintenance: Step (41) Construct an equipment maintenance simulation verification and health assessment platform; Step (42) Construct a health baseline based on the solver model, extract performance degradation features, and construct a feature space for the equipment in a healthy state as the "health baseline" for health assessment: In step (42), the "health baseline" is used as the evaluation benchmark. The distance between the feature points of the data to be evaluated and the "health baseline" is calculated and represented by the Mahalanobis distance (MD). The distance is then normalized to the CV value to represent the current health status of the equipment. The normalization calculation formula for the Mahalanobis distance is as follows: ； In the formula, a is the normalization parameter; CV is a value between 0 and 1; Step (43) Import a multi-dimensional database, embed a simulation software library, and build an intelligent model for the entire maintenance and support process; Step (5) Construct a maintenance decision system: Step (51) Based on the idea of ​​the analytic hierarchy process, establish the hierarchical models of system maintenance in sequence; Step (52) uses multidimensional data processing as input, cross-disciplinary mechanics simulation as control, correlation prediction and state assessment algorithm mechanism, and equipment maintenance intelligent control center and maintenance decision output to output compliance evaluation results.

2. The method for constructing an intelligent equipment maintenance technology architecture based on big data as described in claim 1, characterized in that: The prediction and state assessment algorithm platform in step (3) includes a user layer, a user interface layer, a system interface layer, a data processing layer, and a model management layer.

3. The method for constructing an intelligent equipment maintenance technology architecture based on big data as described in claim 1, characterized in that: The maintenance decision system in step (5) includes maintenance plan, maintenance scheme, maintenance task, maintenance implementation and results.

Images