A method and system for constructing a power plant equipment health assessment model

Abstract

The present invention relates to the technical field of equipment health assessment, and in particular to a method and system for constructing a power plant equipment health assessment model. The method comprises the following steps: obtaining power plant equipment site information data; configuring a site equipment data transmission channel for the power plant equipment site information data to generate a site equipment data transmission channel; integrating the power plant equipment site information data using the site equipment data transmission channel to generate a power plant equipment operation data integration platform; monitoring the power plant equipment operation on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; and calculating equipment multi-operating condition trend fluctuations on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data. The present invention improves the accuracy of the power plant equipment health assessment model through data integration, refined operating condition analysis, non-steady-state abnormal state transition analysis, and dynamic loss assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment health assessment, and in particular to a method and system for constructing a power plant equipment health assessment model. Background Art

[0002] Initially, power plant equipment health assessment relied on experience and periodic manual inspections. This approach suffers from poor timeliness, low accuracy, and an inability to provide early warning of potential failures. With the development of sensor technology, the Internet of Things (IoT), and data acquisition systems, real-time monitoring of power plant equipment has become possible. Sensors can collect real-time equipment operating data (such as temperature, pressure, and vibration), providing essential data for health assessment. The rapid development of machine learning and artificial intelligence technologies has provided new solutions for power plant equipment health assessment. Data-based health assessment models are gradually shifting from traditional statistical methods (such as fault tree analysis and Bayesian networks) to more intelligent algorithmic models such as deep learning, support vector machines (SVMs), and decision trees. These models can predict equipment health and diagnose faults based on historical operating data, sensor data, and fault cases. However, traditional assessment methods fail to fully utilize the changing trends of equipment under different operating conditions for accurate analysis and ignore abnormal conditions that occur during non-steady-state operation. This results in low accuracy in power plant equipment health assessment models. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for constructing a power plant equipment health assessment model to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a method for constructing a power plant equipment health assessment model is provided, the method comprising the following steps:

[0005] Step S1: Acquire power plant equipment site information data; configure a site equipment data transmission channel for the power plant equipment site information data to generate a site equipment data transmission channel; integrate the power plant equipment site information data using the site equipment data transmission channel to generate a power plant equipment operation data integration platform;

[0006] Step S2: monitoring the power plant equipment operation on the power plant equipment operation data integration platform to obtain the power plant equipment operation monitoring data; performing equipment multi-operating condition trend fluctuation calculation on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operation modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; performing equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data;

[0007] Step S3: Acquire rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate steady-state operation mode loss data of power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of power plant equipment and the steady-state operation mode loss data of power plant equipment to generate a power plant equipment health assessment model;

[0008] Step S4: Import the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate a power plant equipment health assessment report; construct an equipment repair decision based on the power plant equipment health assessment report to perform power plant equipment health operations.

[0009] By configuring data transmission channels for site equipment, this invention ensures seamless transmission and integration of operational data from power plant equipment, eliminating data silos and enabling real-time sharing of equipment information. This provides a unified data platform for subsequent monitoring, analysis, and evaluation, ensuring data accuracy, completeness, and timeliness, thereby improving the reliability of health assessments. Multi-operating-condition trend fluctuation calculations comprehensively capture equipment operational fluctuations under different operating conditions, enabling precise classification of equipment's steady-state and non-steady-state operating modes. This analysis helps identify abnormal changes or potential faults in equipment, making health assessments more refined and accurate. Furthermore, abnormal state transition analysis based on non-steady-state operating modes reveals potential equipment risks, provides early warning of faults, and improves equipment safety and reliability. Loss analysis based on steady-state operating modes provides loss data generated during equipment operation, helping to predict long-term wear and tear. Combined with non-steady-state abnormal state transition probability data, a comprehensive assessment of equipment health can be made, not just based on static rated lifespan, but dynamically considering the equipment's actual performance under different operating conditions. This provides a precise basis for equipment maintenance and replacement decisions, avoiding the economic losses caused by excessive or delayed maintenance. By combining data from steady-state and non-steady-state operating modes, a comprehensive equipment health assessment report is generated, which not only provides the current health status of the equipment, but also predicts future failure risks. This comprehensive assessment helps to optimize maintenance strategies, formulate more targeted repair decisions, reduce unnecessary downtime, and improve the operating efficiency of power plants. In addition, the equipment repair decision-making construction based on the health assessment report can formulate maintenance priorities according to the specific conditions of the equipment, thereby optimizing resource allocation, reducing maintenance costs, and extending the service life of the equipment. Therefore, the present invention improves the accuracy of the power plant equipment health assessment model through data integration, refined operating condition analysis, non-steady-state abnormal state transition analysis, and dynamic loss assessment.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Acquire power plant equipment site information data;

[0012] Step S12: performing data preprocessing on the power plant equipment site information data to generate standard power plant equipment site information data, wherein the data preprocessing includes data cleaning, data denoising, data missing value filling and data standardization;

[0013] Step S13: configuring a site equipment data transmission channel for the standard power plant equipment site information data to generate a site equipment data transmission channel;

[0014] Step S14: Use the site equipment data transmission channel to deploy the edge perception computing gateway for the standard power plant equipment site information data to generate edge perception computing gateway deployment data; integrate the standard power plant equipment site information data based on the edge perception computing gateway deployment data to generate a power plant equipment operation data integration platform.

[0015] Through data preprocessing and standardization, this invention ensures the accuracy and consistency of power plant equipment information, providing a reliable foundation for subsequent analysis and prediction, and reducing misjudgments and errors caused by data quality issues. By precisely configuring the transmission channels for site equipment data, the stability and speed of data transmission are optimized, ensuring that power plant equipment data can be transmitted to the central processing system or edge computing nodes in real time, supporting rapid response and decision-making. The deployment of an edge-aware computing gateway allows for direct data processing at the power plant site, reducing the burden of data transmission and improving the system's real-time responsiveness. Preliminary on-site data analysis reduces latency and enables faster response to equipment anomalies. By building an information integration platform, data from diverse power plant equipment can be centrally managed and analyzed, enabling comprehensive monitoring of equipment status. This helps identify potential faults and anomalies, improves equipment operating efficiency, and reduces downtime. The combination of the integrated platform and edge computing gateway facilitates the construction of an intelligent, automated equipment monitoring system capable of automatically detecting anomalies and rapidly responding to them. This not only improves power plant equipment management efficiency but also enhances overall safety and reliability. Through the integrated data platform, the system can provide power plant managers with real-time equipment operating status, and by analyzing historical data, equipment trends, etc., it can issue early warnings to assist managers in making scientific decisions.

[0016] Preferably, the method for constructing a power plant equipment health assessment model includes:

[0017] Based on the edge perception computing gateway deployment data, the standard power plant equipment site information data is uploaded to the edge perception computing gateway for data spatiotemporal alignment to generate power plant equipment site spatiotemporal alignment data;

[0018] Perform data characteristic analysis on the spatiotemporal alignment data of power plant equipment sites to generate data characteristics of power plant equipment sites; perform weighted fusion of associated data on the spatiotemporal alignment data of power plant equipment sites based on the data characteristics of power plant equipment sites to generate associated fusion data of power plant equipment sites;

[0019] The associated integrated data of power plant equipment sites are connected to the cloud platform to generate a power plant equipment operation data integration platform.

[0020] This invention, through the deployment of an edge-aware computing gateway, enables rapid upload and on-site preliminary processing of device data, reducing data transmission latency and enabling real-time access to device health status, providing timely decision-making support for managers. Data characteristic analysis and weighted fusion steps facilitate precise assessment of device health status. By integrating different types of data, the accuracy and reliability of health assessments are improved. By establishing benchmarks and predictive models, changes in device health can be detected promptly, providing early warning of potential equipment failures or degradation trends. The combination of weighted fusion and a cloud platform facilitates comprehensive analysis of various device performance data, providing multi-dimensional support for device health assessments and ensuring the comprehensiveness and accuracy of assessment results. Data analysis based on edge computing and cloud platforms enables power plants to quickly detect equipment anomalies and implement timely intervention. This intelligent monitoring and early warning mechanism effectively reduces equipment failure rates, minimizes unplanned downtime, and improves power plant operational efficiency. Through real-time monitoring and health assessments, power plants can optimize maintenance plans based on equipment health status and implement precise preventive maintenance. This not only extends equipment life but also effectively reduces unnecessary repair costs and downtime.

[0021] Preferably, step S2 includes the following steps:

[0022] Step S21: performing power plant equipment operation monitoring on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; performing multi-dimensional graph conversion on the power plant equipment operation monitoring data to generate a multi-dimensional graph of power plant equipment operation;

[0023] Step S22: performing comparative analysis of equipment multi-operating condition related trends on the power plant equipment operation monitoring data according to the multi-dimensional graph of power plant equipment operation, and generating equipment multi-operating condition trend comparative analysis data;

[0024] Step S23: performing trend fluctuation calculation on the equipment multi-operating condition trend comparison analysis data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operating modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation mode and equipment unsteady-state operation mode;

[0025] Step S24: Based on the non-steady-state operation mode of the equipment, the multi-dimensional map of the power plant equipment operation is monitored and early warning data is screened to obtain non-steady-state early warning data of the power plant equipment; abnormal state transition analysis is performed on the non-steady-state early warning data of the power plant equipment to generate non-steady-state abnormal state transition probability data of the power plant equipment.

[0026] Through multi-dimensional graphs and trend fluctuation analysis, this invention enables power plants to understand the operating status of their equipment in real time, promptly identify potential faults, and issue early warnings. This process not only improves early fault identification capabilities but also prevents large-scale equipment failures during operation, reducing unplanned downtime and repair costs. Comparative analysis of equipment trends across multiple operating conditions provides important support for equipment operation optimization, enabling rational scheduling of equipment operations and load distribution based on equipment performance under different operating conditions, thereby extending equipment lifespan and improving operational efficiency. By classifying equipment's non-steady-state operating modes and analyzing abnormal state transitions, power plants gain a deeper understanding of equipment health and promptly detect anomalies. Abnormal state transition probability data helps managers scientifically predict equipment problems, enabling timely adjustments and maintenance, and reducing the risk of equipment failure and downtime. Through early warning, precise fault detection, and intelligent state transition analysis, power plants can optimize maintenance plans, reduce excessive maintenance and repair costs, and improve equipment operational safety. Comprehensive equipment monitoring, data analysis, and intelligent early warning enable more refined power plant management, avoiding human errors and the inefficiencies of traditional maintenance methods, thereby improving overall operational efficiency. By optimizing operating modes and adjusting equipment loads, power plants can maximize the use of existing equipment resources and improve production efficiency.

[0027] Preferably, performing multi-dimensional graph conversion on power plant equipment operation monitoring data includes:

[0028] Perform power plant equipment operation signal conversion on power plant equipment operation monitoring data to generate power plant equipment operation signals; perform waveform conversion on power plant equipment operation signals to generate power plant equipment operation signal waveforms; perform time domain analysis on power plant equipment operation signal waveforms to generate power plant equipment time domain signal waveforms;

[0029] Perform order ratio analysis on the time domain signal waveform of power plant equipment to generate an order ratio diagram of power plant equipment;

[0030] Perform signal waveform extraction on the time domain signal waveform of power plant equipment to obtain the time domain signal waveform curve of power plant equipment; perform FFT transformation on the time domain signal waveform curve of power plant equipment to generate the frequency domain waveform curve of power plant equipment; perform power spectrum analysis on the frequency domain waveform curve of power plant equipment to generate the power spectrum diagram of power plant equipment;

[0031] Extract equipment influencing parameters from the power spectrum of power plant equipment to obtain power influencing parameters of power plant equipment; perform waterfall chart analysis on the power spectrum of power plant equipment based on the power influencing parameters of power plant equipment to generate a waterfall chart of power plant equipment;

[0032] The time domain signal waveform diagram of power plant equipment, the order ratio diagram of power plant equipment, the power spectrum diagram of power plant equipment and the waterfall diagram of power plant equipment are integrated to generate a multi-dimensional diagram of power plant equipment operation.

[0033] By comprehensively analyzing data from different dimensions, including time domain, frequency domain, and power spectrum, this method enables equipment fault diagnosis from multiple perspectives. For example, time domain signal analysis identifies equipment fluctuations, frequency domain analysis reveals potential mechanical failures or electrical issues, and waterfall chart analysis predicts long-term equipment operational issues. By integrating multi-dimensional graphs, power plants can comprehensively assess equipment performance, including key indicators such as vibration, load, and energy distribution. This provides an accurate basis for long-term equipment health management, performance optimization, and load scheduling. Frequency domain and power spectrum analysis of power plant equipment can reveal early signs of failure, such as frequency anomalies and mechanical wear, providing data support for equipment maintenance decisions. Waterfall chart analysis can proactively identify potential equipment problems under different loads, enabling preemptive troubleshooting or repairs, reducing the risk of sudden failures. By integrating the graphs, power plant managers can obtain a clear and concise overview of equipment health status, facilitating rapid decision-making. The graphs not only reflect the current status but also predict the likelihood of future failures through analysis of historical data, making power plant operations more intelligent and scientific. Based on accurate fault diagnosis and performance analysis from multi-dimensional graphs, power plants can develop more rational maintenance strategies, avoiding over-maintenance and misdiagnosis. By optimizing maintenance plans, equipment downtime is reduced, operational efficiency is improved, and ultimately maintenance costs and production interruption losses are reduced.

[0034] Preferably, dividing the power plant equipment operation monitoring data into equipment operation modes according to the equipment operating condition trend fluctuation data includes:

[0035] Segmenting the equipment operating condition trend fluctuation data into adjacent fluctuation curve segments based on a preset time range to obtain a first equipment operating condition trend fluctuation curve segment and a second equipment operating condition trend fluctuation curve segment; and calculating trend fluctuation peak values ​​for the first equipment operating condition trend fluctuation curve segment and the second equipment operating condition trend fluctuation curve segment to obtain a first fluctuation curve peak value and a second fluctuation curve peak value.

[0036] Calculating the frequency domain signal ratio of the first fluctuation curve peak value and the second fluctuation curve peak value to obtain first fluctuation curve frequency domain signal ratio data and second fluctuation curve frequency domain signal ratio data, wherein the frequency domain signal includes a low-frequency signal and a high-frequency signal; performing a frequency characteristic analysis based on the first fluctuation curve frequency domain signal ratio data and the first fluctuation curve peak value to generate first frequency fluctuation characteristic data; performing a frequency characteristic analysis based on the second fluctuation curve frequency domain signal ratio data and the second fluctuation curve peak value to generate second frequency fluctuation characteristic data;

[0037] The first frequency fluctuation characteristic data and the second frequency fluctuation characteristic data are subjected to similarity calculation to obtain frequency fluctuation characteristic similarity data; the frequency fluctuation characteristic similarity data is compared with a preset standard similarity threshold value; when the frequency fluctuation characteristic similarity data is greater than or equal to the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a steady-state operation mode; when the frequency fluctuation characteristic similarity data is less than the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a non-steady-state operation mode.

[0038] By segmenting fluctuation curves and calculating fluctuation peaks and frequency domain signal proportions, the present invention accurately captures subtle changes during equipment operation. This detailed trend analysis helps precisely identify the equipment's operating status. Similarity calculation based on frequency characteristic analysis effectively distinguishes whether the equipment is in steady-state or non-steady-state mode. Steady-state mode indicates stable equipment operation, while non-steady-state mode indicates potential equipment failure or anomalies. By comparing frequency fluctuation characteristic similarity data with standard similarity thresholds, the system can promptly identify equipment with abnormal operation and issue warnings, helping operators take preemptive measures to prevent dangerous events such as equipment failure or fire. Combined with real-time equipment operation data, this system can provide intelligent decision-making support for power plant management. After identifying non-steady-state operating modes, operators can perform targeted maintenance or adjustments to optimize equipment operating efficiency. By timely monitoring equipment operating status and accurately classifying operating modes, sudden failures caused by unclear equipment operating status are avoided, improving the long-term reliability and stability of power plant equipment.

[0039] Preferably, performing abnormal state transition analysis on the non-steady-state warning data of power plant equipment includes:

[0040] Extract warning parameters from the non-steady-state warning data of power plant equipment to obtain non-steady-state warning parameters of power plant equipment; confirm warning timestamps of non-steady-state warning parameters of power plant equipment to obtain non-steady-state warning timestamps of power plant equipment;

[0041] Dynamically screen the warning parameters of the power plant equipment's non-steady-state warning parameters with adjacent timestamps using the non-steady-state warning timestamps to obtain warning correlation parameters; perform abnormal state correlation analysis on the non-steady-state warning parameters of the power plant equipment using the warning correlation parameters to generate non-steady-state abnormal state correlation data for the power plant equipment;

[0042] A Markov chain is constructed on the correlation data of the non-steady abnormal state of power plant equipment to generate a power plant equipment state transition graph; and edge distance calculation is performed on the power plant equipment state transition graph to generate the non-steady abnormal state transition probability data of power plant equipment.

[0043] The present invention extracts parameters and dynamically filters the non-steady-state warning data of power plant equipment, and can extract key warning parameters closely related to equipment failures from a large amount of monitoring data. This screening can reduce redundant information and improve the accuracy of analysis. By confirming the warning timestamp and dynamically filtering the warning parameters, abnormal changes in the equipment within a certain time period can be clearly identified, and relevant parameters can be timely associated to help quickly identify problems with the equipment. For example, multiple time windows of abnormal states can provide a more accurate time range for fault prediction. By performing abnormal state association analysis on the warning association parameters, the inherent connection between different states of the equipment can be revealed, and potential systemic failure modes can be identified. In this way, the working state of the equipment can be understood as a whole, avoiding misdiagnosis caused by a single factor. By using the Markov chain model to model the state transition process of the equipment, the transition law of the equipment state can be accurately described. By generating a power plant equipment state transition diagram, the future state changes of the equipment can be predicted, providing data support for subsequent maintenance and management. By calculating the edge distance of the equipment state transition diagram, the transition probability of the non-steady-state abnormal state of the power plant equipment can be obtained. This enables the future state of the equipment to be described probabilistically, providing a scientific basis for formulating equipment maintenance and optimization plans, and further improving the safety and reliability of the equipment.

[0044] Preferably, step S3 includes the following steps:

[0045] Step S31: Acquire rated life data of power plant equipment;

[0046] Step S32: performing a power plant equipment operating load analysis on the multi-dimensional graph of power plant equipment operation based on the steady-state operating mode of the equipment to generate steady-state operating power plant equipment operating load data; performing an equipment loss analysis on the rated life data of the power plant equipment based on the steady-state operating power plant equipment operating load data to generate steady-state operating mode loss data of the power plant equipment;

[0047] Step S33: dividing the data sets of the power plant equipment's non-steady-state abnormal state transition probability data and the power plant equipment's steady-state operation mode loss data into a model training set and a model test set;

[0048] Step S34: Perform model training on the model training set using a feedforward neural network algorithm to generate a power plant equipment health assessment pre-model; perform model optimization iteration on the power plant equipment health assessment pre-model using the model test set to generate a power plant equipment health assessment model.

[0049] By performing operating load analysis based on steady-state equipment operation, the present invention comprehensively assesses the actual operating conditions of power plant equipment, including load conditions and operating efficiency, helping to identify whether the equipment is operating optimally. Analyzing the equipment's operating load and rated life in steady-state operation accurately assesses equipment wear and tear, providing a reference for future maintenance, overhaul, and replacement. Generating wear and tear data in steady-state operation helps reveal issues such as physical wear, fatigue damage, and reduced efficiency during long-term operation, helping managers understand the current health of the equipment. This analysis provides data support for equipment lifecycle prediction, enabling more accurate maintenance and replacement plans for power plants. By rationally partitioning the non-steady-state abnormal state transition probability data and steady-state operating mode wear and tear data to generate training and test sets, the representativeness and accuracy of the training data are ensured. This step avoids overfitting and underfitting and improves the model's predictive capabilities. Using a feedforward neural network algorithm for training not only leverages the advantages of deep learning to process large amounts of data but also learns complex equipment health assessment models. Optimizing and iterating the model using the test set further improves the accuracy and stability of the power plant equipment health assessment model. The iterative optimization process helps capture subtle variations in equipment health, thereby improving the model's accuracy and practical application value. This method, combined with steady-state operation analysis and data on abnormal transitions to non-steady states, provides comprehensive data support for power plant equipment health assessments. Through neural network model training, potential equipment failures can be intelligently identified, providing early warnings to prevent sudden equipment failures and enhancing the intelligence of equipment management.

[0050] Preferably, step S4 includes the following steps:

[0051] Step S41: Importing the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate power plant equipment health assessment data;

[0052] Step S42: Visualize the power plant equipment health assessment data to generate a power plant equipment health assessment report; construct equipment repair decisions based on the power plant equipment health assessment report to perform power plant equipment health operations.

[0053] By inputting data on the probability of transitions from non-steady-state abnormal states and loss data from steady-state operating modes into a health assessment model, the present invention comprehensively assesses the health status of power plant equipment. This comprehensive analysis method considers changes in equipment health under different operating modes, providing more accurate health assessment data. Comprehensive health assessment data not only reflects the current operating status of the equipment but also predicts future failures or damage, helping managers make proactive maintenance or replacement decisions and mitigate the impact of equipment failures. Visualizing power plant equipment health assessment data transforms complex technical data into intuitive, easy-to-understand charts and reports, helping managers, engineers, and decision makers more quickly and accurately understand the equipment's health status, risk levels, and potential issues. Visual reports not only display the health assessment results but also highlight key risk points and early warning information, helping power plant personnel quickly respond to emerging equipment issues and reduce the probability of failures. Health assessment reports provide a scientific basis for equipment repair or replacement. An intelligent decision support system can develop optimized repair plans based on equipment health status and risk assessments, improving the relevance and efficiency of repair efforts. Automating equipment repair decisions reduces human error and improves the accuracy of repair decisions. At the same time, properly arranging maintenance and replacement time can help reduce equipment downtime and avoid impact on power plant production.

[0054] In this specification, a system for constructing a power plant equipment health assessment model is provided, which is used to execute the above-mentioned method for constructing a power plant equipment health assessment model. The system for constructing a power plant equipment health assessment model includes:

[0055] The data integration module is used to obtain the site information data of the power plant equipment; configure the site equipment data transmission channel for the power plant equipment site information data to generate the site equipment data transmission channel; use the site equipment data transmission channel to integrate the power plant equipment site information data to generate the power plant equipment operation data integration platform;

[0056] The operating condition analysis module is used to monitor the operation of power plant equipment on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; perform equipment multi-operating condition trend fluctuation calculations on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; classify the power plant equipment operation monitoring data into equipment operating modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; perform equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data;

[0057] The health prediction module is used to obtain the rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate the steady-state operation mode loss data of the power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment to generate the power plant equipment health assessment model;

[0058] The evaluation and decision-making module is used to import the non-steady-state abnormal state transition probability data of power plant equipment and the loss data of the steady-state operation mode of power plant equipment into the power plant equipment health assessment model to conduct a comprehensive health assessment and generate a power plant equipment health assessment report; based on the power plant equipment health assessment report, equipment repair decision-making is constructed to execute power plant equipment health operations.

[0059] The beneficial effects of the present invention lie in the efficient integration and real-time data transmission of multi-site equipment information through the configuration of site equipment data transmission channels and the establishment of a power plant equipment operation data integration platform, ensuring data integrity and accuracy. The use of this integrated platform simplifies the data management process, provides unified data support for real-time monitoring and subsequent analysis of equipment status, and improves overall operational efficiency. Equipment operation monitoring, combined with multi-operating condition trend fluctuation calculations, comprehensively captures dynamic changes in equipment operating status and accurately identifies steady-state and unsteady-state operating modes. The classification of operating modes effectively distinguishes between normal and abnormal equipment operating states, providing a foundation for subsequent abnormality analysis and health assessment. Abnormal state transition probability analysis, using unsteady-state operating mode data, quantifies the probability of equipment transitioning from one state to another. Markov chain analysis models provide a scientific basis for predicting equipment operation, helping managers identify potential risks in advance and take preventive measures. Based on equipment steady-state operating modes and unsteady abnormal state transition probability data, and taking into account equipment losses and abnormal conditions, a health assessment model is constructed. This assessment method can comprehensively quantify the health status of equipment. The construction of this health assessment model provides a scientific equipment status diagnostic tool that can accurately assess the current operating status and remaining life of the equipment. Health assessment reports visually display the operational and health status of equipment, providing a clear scientific basis for repair or replacement. Repair decisions, combined with health assessment data, can dynamically adjust maintenance plans, avoid blind repairs, and optimize resource allocation. Therefore, this invention improves the accuracy of power plant equipment health assessment models through data integration, refined operating condition analysis, analysis of non-steady-state abnormal state transitions, and dynamic loss assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic flow chart of steps for constructing a power plant equipment health assessment model;

[0061] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0062] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0063] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0064] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0065] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0066] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0067] To achieve this, please refer to Figures 1 to 3 A method for constructing a power plant equipment health assessment model comprises the following steps:

[0068] Step S1: Acquire power plant equipment site information data; configure a site equipment data transmission channel for the power plant equipment site information data to generate a site equipment data transmission channel; integrate the power plant equipment site information data using the site equipment data transmission channel to generate a power plant equipment operation data integration platform;

[0069] Step S2: monitoring the power plant equipment operation on the power plant equipment operation data integration platform to obtain the power plant equipment operation monitoring data; performing equipment multi-operating condition trend fluctuation calculation on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operation modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; performing equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data;

[0070] Step S3: Acquire rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate steady-state operation mode loss data of power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of power plant equipment and the steady-state operation mode loss data of power plant equipment to generate a power plant equipment health assessment model;

[0071] Step S4: Import the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate a power plant equipment health assessment report; construct an equipment repair decision based on the power plant equipment health assessment report to perform power plant equipment health operations.

[0072] By configuring data transmission channels for site equipment, this invention ensures seamless transmission and integration of operational data from power plant equipment, eliminating data silos and enabling real-time sharing of equipment information. This provides a unified data platform for subsequent monitoring, analysis, and evaluation, ensuring data accuracy, completeness, and timeliness, thereby improving the reliability of health assessments. Multi-operating-condition trend fluctuation calculations comprehensively capture equipment operational fluctuations under different operating conditions, enabling precise classification of equipment's steady-state and non-steady-state operating modes. This analysis helps identify abnormal changes or potential faults in equipment, making health assessments more refined and accurate. Furthermore, abnormal state transition analysis based on non-steady-state operating modes reveals potential equipment risks, provides early warning of faults, and improves equipment safety and reliability. Loss analysis based on steady-state operating modes provides loss data generated during equipment operation, helping to predict long-term wear and tear. Combined with non-steady-state abnormal state transition probability data, a comprehensive assessment of equipment health can be made, not just based on static rated lifespan, but dynamically considering the equipment's actual performance under different operating conditions. This provides a precise basis for equipment maintenance and replacement decisions, avoiding the economic losses caused by excessive or delayed maintenance. By combining data from steady-state and non-steady-state operating modes, a comprehensive equipment health assessment report is generated, which not only provides the current health status of the equipment, but also predicts future failure risks. This comprehensive assessment helps to optimize maintenance strategies, formulate more targeted repair decisions, reduce unnecessary downtime, and improve the operating efficiency of power plants. In addition, the equipment repair decision-making construction based on the health assessment report can formulate maintenance priorities according to the specific conditions of the equipment, thereby optimizing resource allocation, reducing maintenance costs, and extending the service life of the equipment. Therefore, the present invention improves the accuracy of the power plant equipment health assessment model through data integration, refined operating condition analysis, non-steady-state abnormal state transition analysis, and dynamic loss assessment.

[0073] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart showing the steps of a method for constructing a power plant equipment health assessment model according to the present invention. In this example, the method for constructing a power plant equipment health assessment model includes the following steps:

[0074] Step S1: Acquire power plant equipment site information data; configure a site equipment data transmission channel for the power plant equipment site information data to generate a site equipment data transmission channel; integrate the power plant equipment site information data using the site equipment data transmission channel to generate a power plant equipment operation data integration platform;

[0075] In an embodiment of the present invention, real-time device data, including but not limited to temperature, pressure, flow rate, device operating status, fault information, and maintenance records, is acquired from various equipment sites in a power plant. This data is typically collected through sensors and device interfaces and stored in a local database or a remote cloud platform. Data can be interfaced with a power plant management system (e.g., a SCADA system) or an equipment automation system (e.g., a PLC system) to ensure real-time or periodic data acquisition. Appropriate communication protocols and data transmission methods are designed based on the power plant's network environment and device types. For example, protocols such as OPC (OLE for Process Control), Modbus, MQTT, and REST API can be used to establish a real-time data transmission channel between the device and the central platform. Parameters of the data transmission channel, such as data transmission frequency, bandwidth, error checking, and retransmission mechanisms, are configured to ensure data stability and security. A VPN (virtual private network) or encrypted channel may be established to enhance data transmission security. Data collected by different devices should be consistent in format, using a unified format (e.g., JSON, XML, or CSV) for data transmission to facilitate subsequent data processing and integration. Through the configured data transmission channel, device operating data from each site is acquired in real time and imported into the data integration platform. The data integration platform can automatically identify and convert data formats from different sites and devices, ensuring that data from different sources can be displayed and analyzed uniformly on a single platform. Collected device data is integrated through algorithms, such as time series analysis to address data timeliness, data cleaning techniques to remove noise and invalid data, and missing value filling methods to supplement missing data. Based on this integrated data, a power plant equipment operation data integration platform can be established. This platform can use data visualization technology to display data from different devices in real time, supporting various display methods such as charts, reports, and trend analysis, helping operations and maintenance personnel promptly identify equipment failures or potential issues. The integration platform should possess intelligent analytical capabilities, using machine learning and data mining techniques to analyze device data, such as for failure prediction, remaining life estimation, and performance optimization recommendations. The platform should support data exchange and interfacing with other management systems (such as ERP, CMMS, and EAM) to form a complete equipment management system.

[0076] Step S2: monitoring the power plant equipment operation on the power plant equipment operation data integration platform to obtain the power plant equipment operation monitoring data; performing equipment multi-operating condition trend fluctuation calculation on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operation modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; performing equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data;

[0077] In an embodiment of the present invention, a power plant equipment operation data integration platform first identifies key monitoring indicators, such as equipment temperature, pressure, speed, vibration, and power consumption. This data is collected by real-time sensors and transmitted to the integration platform via data transmission channels. The platform utilizes real-time data stream processing technology to continuously monitor the operating data of various types of equipment. For example, a time series database (such as InfluxDB) is used to store equipment operation data, and a real-time monitoring system (such as Grafana) is used to display and generate alerts. When equipment parameters exceed preset normal ranges, the platform triggers an alert to notify operations and maintenance personnel. Machine learning models can be combined to identify early signs of equipment anomalies. Equipment data performance varies when operating under different operating conditions. For example, operating parameters such as load and speed directly affect the equipment's operating status. Based on historical equipment data, clustering algorithms (such as K-means or DBSCAN) are used to identify multiple operating condition patterns and classify them into different operating condition categories. For each operating condition category, time series analysis (such as Fourier transform and wavelet analysis) is used to calculate the fluctuation trend of the equipment operation data. The following methods can be used: trend analysis methods (such as the Hodrick-Prescott filter) can be used to remove short-term fluctuations while retaining long-term trends. Volatility analysis methods (such as standard deviation and root mean square error) can be used to quantify the fluctuation amplitude of equipment under different operating conditions. Fluctuation data under different operating conditions can be used to generate equipment operating trend fluctuation data, which will be used for subsequent mode classification and state prediction. Based on the equipment operating trend fluctuation data, pattern recognition algorithms can be used to classify the equipment's operating state. Generally speaking, a steady-state mode refers to stable equipment operation with minimal fluctuations, while an unsteady-state mode refers to equipment experiencing large fluctuations or abnormal behavior. Under normal operating conditions, equipment performance fluctuates minimally, operates stably, and meets design parameters. Methods such as mean-standard deviation analysis and root mean square error can be used to determine whether the equipment is in an abnormal state or has experienced a fault during operation, resulting in unstable operation. Periods of significant fluctuation can be identified by setting fluctuation thresholds or using trend analysis. Classification algorithms such as support vector machines (SVMs) or decision trees can be used to classify equipment operating data into steady-state and unsteady-state modes based on fluctuation trends. When a device is in a non-steady-state mode, its operating state may change, such as from normal to faulty, or from a minor fault to a major fault. First, a state transition model for the device is constructed based on historical data. This can be achieved using a Markov process to model the device's state transitions. Different device states are defined, such as "normal operation," "minor fault," and "major fault." A state transition matrix is ​​used to describe the transition probabilities between these states. Based on the device's operating data, the transition probabilities between each state are calculated. For example, the transition probability from "normal operation" to "minor fault" or from "minor fault" to "major fault" is calculated.Based on historical equipment failure data, operating parameter fluctuations, and other factors, probabilistic statistical methods (such as maximum likelihood estimation) are used to calculate the transition probabilities between various states. A hidden Markov model (HMM) can be used to implement dynamic transition analysis of equipment abnormal states. Through the equipment state transition model, combined with operating condition trend fluctuation data and historical operating data, transition probability data for non-steady-state abnormal states of power plant equipment is generated. This data represents the transition probabilities between different abnormal states of the equipment in non-steady-state mode. A matrix is ​​generated, with each row and column representing an equipment operating state, and each element in the matrix represents the transition probability from one state to another. Based on this transition probability data, operations and maintenance personnel can predict equipment failure trends, providing a basis for maintenance decisions. Machine learning models can also combine this data to predict abnormal states and provide early warnings.

[0078] Step S3: Acquire rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate steady-state operation mode loss data of power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of power plant equipment and the steady-state operation mode loss data of power plant equipment to generate a power plant equipment health assessment model;

[0079] In embodiments of the present invention, rated life data for power plant equipment is obtained. This data is typically provided by the equipment manufacturer or calculated based on the equipment's operating history and design specifications. Rated life generally refers to the maximum expected time a device can operate without failure under specific operating conditions. The acquired rated life data typically includes: the rated life of various equipment, such as generators, transformers, and boilers; the operating time (e.g., hours or years) of the equipment under design conditions; the load value of the equipment under rated conditions; and the regular maintenance requirements for the equipment during its rated life. In steady-state operation, equipment operating parameters vary little, and performance is relatively stable. Therefore, in this mode, equipment losses primarily result from factors such as friction, wear, and aging during normal operation. By modeling the equipment's operating data in steady-state operation, the losses of various equipment components under long-term stable operating conditions can be analyzed. This can be achieved by: For example, the energy loss of a generator during steady-state operation can be estimated by calculating the difference between the device's input and output power. Mechanical component losses caused by vibration, friction, and other factors are typically assessed by analyzing the equipment's operating parameters (e.g., temperature, pressure, and speed). The electrical losses of equipment such as transformers and generators are typically determined by parameters such as resistance, inductance, and capacity, and can be calculated using electrical models. Based on information such as equipment operating time, load conditions, and operating conditions, the cumulative losses of the equipment in steady-state operation are calculated. For example, by integrating various losses, the loss value of the equipment over a specified period is obtained. Based on the above analysis results, steady-state loss data is generated for the equipment, including: Cumulative loss: The total loss value of the equipment in steady-state operation up to the current moment. Loss rate: The rate of loss per unit time in steady-state operation (e.g., energy loss per hour). The purpose of the health assessment model is to assess the current health status and remaining life of the equipment by combining abnormal transition probability data of the equipment in non-steady-state mode with loss data in steady-state mode. This model can be constructed using the following steps: The equipment's health index is calculated based on the equipment's steady-state losses and the probability of transitions to abnormal states in non-steady-state mode. The health index assesses the current health status of the equipment by combining steady-state losses and the risk of abnormal state transitions. The remaining useful life of the equipment is calculated based on the equipment's loss data and state transition probabilities. Machine learning models (such as regression models and neural networks) can be used to predict the remaining life of the equipment. A Bayesian network model is used to represent the health status and state transition probabilities of the device. This model, combined with the device's steady-state loss, allows for inference and assessment of the device's health status. Using the device's steady-state loss data and unsteady-state transition data, regression models (such as linear regression and support vector regression) are used to estimate the device's health index and remaining lifespan.Based on the deep learning neural network model, a health assessment model is trained through the equipment's historical operation data, loss data, and state transition data. It automatically calculates the equipment's health index and remaining lifespan, and generates regular assessment reports on the equipment's health status. The reports include the equipment's health index, remaining lifespan, and failure modes that require attention.

[0080] Step S4: Import the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate a power plant equipment health assessment report; construct an equipment repair decision based on the power plant equipment health assessment report to perform power plant equipment health operations.

[0081] In this embodiment of the present invention, by converting non-steady-state abnormal state transition probability data and steady-state loss data according to a predetermined format, they ensure seamless integration within the health assessment model. Data cleaning is performed to remove noise and outliers, and standardization is performed to ensure that the data meets the input requirements of the assessment model. An interface compatible with the power plant equipment health assessment model is developed to import the aforementioned data into the model. This interface can be a database interface, an API interface, or an automated data transmission system based on a cloud platform. The imported data (non-steady-state abnormal state transition probability data and steady-state loss data) is input into a pre-set health assessment model (such as a Bayesian network, neural network, support vector machine, etc.). Based on this data, the model calculates the equipment's health index and predicts the equipment's remaining useful life (RUL). Based on the non-steady-state abnormal state transition probability data, the model assesses the probability and risk of equipment abnormalities occurring in the future and combines this with the steady-state loss data for a comprehensive analysis. The steady-state loss data is used to assess the equipment's wear progression under normal operating conditions, which in turn affects the prediction of the health index and RUL. Based on the comprehensive assessment results, a health assessment report for the equipment is generated.

[0082] Preferably, step S1 includes the following steps:

[0083] Step S11: Acquire power plant equipment site information data;

[0084] Step S12: performing data preprocessing on the power plant equipment site information data to generate standard power plant equipment site information data, wherein the data preprocessing includes data cleaning, data denoising, data missing value filling and data standardization;

[0085] Step S13: configuring a site equipment data transmission channel for the standard power plant equipment site information data to generate a site equipment data transmission channel;

[0086] Step S14: Use the site equipment data transmission channel to deploy the edge perception computing gateway for the standard power plant equipment site information data to generate edge perception computing gateway deployment data; integrate the standard power plant equipment site information data based on the edge perception computing gateway deployment data to generate a power plant equipment operation data integration platform.

[0087] In this embodiment of the present invention, information is collected from various equipment sites within a power plant, including data on equipment type, location, operating status, maintenance logs, and energy consumption. Automated means such as sensors, monitoring systems, and IoT devices collect data from these sites in real time, ensuring data integrity and timeliness. Data from different sites and equipment is integrated to ensure a unified format and standard for subsequent processing. Duplicate, erroneous, or invalid data is removed to ensure data quality. Significantly abnormal values ​​are marked or deleted to prevent interference with analysis. Smoothing algorithms (such as moving averages or Kalman filters) are used to remove noise from sensor data. Signal processing techniques are employed to enhance data accuracy. Missing equipment data is filled using mean filling, interpolation, or predictive models to ensure data integrity. Data of different dimensions (such as temperature, humidity, and current) is standardized to ensure consistent units of measurement, facilitating subsequent analysis. Appropriate transmission channels are designed based on the equipment site's geographic location, network conditions, and equipment type. Wireless, wired, or hybrid networks are used to ensure efficient data transmission. Protocols suitable for data transmission at the equipment site, such as MQTT, HTTP / HTTPS, and Modbus, are configured. Set transmission parameters such as frequency and bandwidth to ensure smooth data transmission and avoid congestion. Test the configured transmission channels to ensure stability, reliability, and efficiency. Adjust the transmission plan based on test results to ensure data real-time and integrity. Select appropriate edge computing gateway devices based on the data volume, real-time requirements, and computing needs of the power plant equipment. Deploy an edge computing gateway at each equipment site to ensure the gateway can receive and perform preliminary processing of equipment data in real time. Configure the gateway to perform local data processing, such as data filtering, preprocessing, caching, and real-time analysis, to reduce data transmission pressure and improve efficiency. Configure the gateway's data storage, backup, and recovery mechanisms to ensure data reliability and security. Integrate data generated by different sites and equipment, performing multi-level data fusion (e.g., temporal and spatial fusion). Ensure compatibility of data from different sources through data cleansing and format conversion. Design and build an integrated platform that can receive data from each edge computing gateway. The platform should support multiple data sources and real-time data stream processing, and be highly scalable and flexible. Implement functions such as real-time data monitoring, visualization, alarm mechanisms, and historical data analysis. Provide intelligent functions such as equipment failure prediction, operation optimization, and energy consumption analysis to improve power plant operating efficiency. Test the integrated platform at different sites to ensure data accuracy and platform stability. Optimize the platform based on test feedback to improve data processing capabilities and response speed.

[0088] Preferably, the method for constructing a power plant equipment health assessment model includes:

[0089] Based on the edge perception computing gateway deployment data, the standard power plant equipment site information data is uploaded to the edge perception computing gateway for data spatiotemporal alignment to generate power plant equipment site spatiotemporal alignment data;

[0090] Perform data characteristic analysis on the spatiotemporal alignment data of power plant equipment sites to generate data characteristics of power plant equipment sites; perform weighted fusion of associated data on the spatiotemporal alignment data of power plant equipment sites based on the data characteristics of power plant equipment sites to generate associated fusion data of power plant equipment sites;

[0091] The associated integrated data of power plant equipment sites are connected to the cloud platform to generate a power plant equipment operation data integration platform.

[0092] In this embodiment of the present invention, real-time operating data (such as temperature, pressure, vibration, and power) is collected from various power plant equipment sites. This data typically has different timestamps and spatial distributions. The collected power plant equipment site information is uploaded to an edge computing gateway via a configured transmission channel. The edge computing gateway is responsible for local data processing and preliminary analysis. Because different devices have different collection times and different spatial locations for data collection, it is necessary to perform spatiotemporal alignment of data from different devices. This ensures that the data is compared and analyzed according to unified temporal and spatial standards. For device data collected at different times, the data is interpolated or synchronized using timestamps to ensure that the data from each device has the same timestamp. Using Geographic Information System (GIS) technology, the device locations are calibrated, and the device data from different sites is mapped to a unified spatial coordinate system to ensure spatial data consistency. Statistical analysis is performed on the historical data of each device site to extract its basic statistical characteristics, such as mean, variance, maximum, and minimum values. Time series analysis is used to analyze the long-term operating trends and change patterns of the equipment, such as gradual aging trends and load change trends. Perform periodic analysis of equipment operating data to identify cyclical patterns (such as daytime and nighttime operating patterns and seasonal fluctuations). Use outlier detection algorithms (such as box plot-based outlier detection and the Z-score algorithm) to identify abnormal fluctuations or deviations from the normal operating range. Through this analysis, extract core characteristics of the equipment data, such as the distribution of equipment operating states, operating cycles, and load variations. These characteristics are aggregated into preliminary indicators of equipment health, including efficiency, health status, and potential failure risks. Correlations between equipment are identified based on their types, uses, and relationships. For example, the operating states of wind turbines and generators are highly correlated. Different weights are assigned to each type of data based on equipment characteristics (such as importance, accuracy, and stability). The health of critical equipment warrants a higher weight, while the impact of auxiliary equipment can be appropriately weighted. Data fusion can be performed using methods such as weighted averaging, principal component analysis (PCA), or factor analysis (FA). Spatiotemporally aligned data from each equipment site is weighted and fused according to preset weights to generate comprehensive, correlated data that reflects the mutual impact of each device and its overall operating status. Store the weighted fused data locally or in the cloud, providing a comprehensive perspective for subsequent health assessments. Upload the generated fused data related to power plant equipment sites to the cloud platform via the edge computing gateway. Configure data transmission channels to ensure stable and secure data upload from the edge computing gateway to the cloud platform. Use protocols such as HTTPS and MQTT for data transmission. After receiving the fused data related to the equipment, the cloud platform performs further data integration.The cloud platform should support efficient real-time data processing and historical data storage, and be able to handle large amounts of data streams from diverse equipment sites. The platform design should include data storage, a computing engine, a visualization interface, and alarm and report generation modules. The platform should be capable of performing health assessment analysis, fault prediction, and early warning systems. Integrating cloud computing resources, the platform can perform large-scale computations and machine learning model training to improve the accuracy of equipment health assessments. Implementing equipment health assessment models on the cloud platform will visualize information such as equipment health and operating efficiency. Providing real-time monitoring, fault prediction, and maintenance recommendations will support the power plant operations team in making more accurate decisions.

[0093] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0094] Step S21: performing power plant equipment operation monitoring on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; performing multi-dimensional graph conversion on the power plant equipment operation monitoring data to generate a multi-dimensional graph of power plant equipment operation;

[0095] Step S22: performing comparative analysis of equipment multi-operating condition related trends on the power plant equipment operation monitoring data according to the multi-dimensional graph of power plant equipment operation, and generating equipment multi-operating condition trend comparative analysis data;

[0096] Step S23: performing trend fluctuation calculation on the equipment multi-operating condition trend comparison analysis data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operating modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation mode and equipment unsteady-state operation mode;

[0097] Step S24: Based on the non-steady-state operation mode of the equipment, the multi-dimensional map of the power plant equipment operation is monitored and early warning data is screened to obtain non-steady-state early warning data of the power plant equipment; abnormal state transition analysis is performed on the non-steady-state early warning data of the power plant equipment to generate non-steady-state abnormal state transition probability data of the power plant equipment.

[0098] In an embodiment of the present invention, real-time operational data (such as temperature, pressure, current, and voltage) from each device is collected on a power plant equipment operation data integration platform to monitor equipment operating status. Sensors and edge computing gateways are deployed to achieve high-frequency data collection and transmission. Monitored equipment types include generators, transformers, boilers, cooling systems, and other core equipment, covering the operational status of various core power plant equipment. Equipment operating data is converted from single time series data into multidimensional graphs, showcasing relationships between multiple variables, such as equipment operating status, load, and environmental factors. Multidimensional visualization techniques (such as radar charts, heat maps, and 3D graphs) are used to display various equipment health indicators. Multidimensional graphs facilitate analysis of equipment performance under different operating conditions, providing intuitive visualization for subsequent analysis. Comparative analysis of the multidimensional operational data of each power plant device under multiple operating conditions identifies the impact of different operating conditions on equipment performance. These operating conditions include load variations, ambient temperature, and operating time. Correlation analysis and trend analysis are used to assess equipment behavior patterns under different operating conditions, revealing performance degradation or anomalies under specific conditions. Compare device performance under different operating conditions to identify optimal operating conditions and performance bottlenecks. Aggregate and generate a set of comparative trend data for devices under multiple operating conditions. This data can be used for subsequent fault prediction and health assessment. Based on this multi-condition trend analysis data, calculate the fluctuation amplitude and frequency of device operating data to determine the fluctuation characteristics of the device under different operating conditions. Fluctuation analysis techniques (such as Fourier transform and fluctuation amplitude analysis) are used to extract fluctuation information about the device's operating conditions and assess its operational stability. Based on the device's fluctuation data and certain thresholds, a distinction is made between steady-state and unstable operating modes. Steady-state operating mode: The device operates stably without significant fluctuations or anomalies, typically indicating normal operation. Unsteady-state operating mode: The device exhibits significant fluctuations or anomalies under certain circumstances, typically as precursors to faults or temporary anomalies caused by external disturbances. Combining the device's operating patterns with trend fluctuation data, a classification of steady and unstable states is generated, providing a basis for subsequent health assessment and fault warning. Based on the device's unstable operating mode obtained in the previous step, the device's multi-dimensional profile is filtered, focusing on data that exhibits abnormal or fluctuating characteristics. The screening process involves checking thresholds and identifying fluctuations across various device indicators. Data that does not meet steady-state conditions is flagged for further analysis. State transition models (such as Markov chains and state machines) are applied to the identified non-steady-state warning data to analyze the device's transition from normal to abnormal states. This analysis examines the different operating states the device experiences during non-steady-state conditions, such as the transition path from a minor anomaly to a severe failure.Based on abnormal state transition analysis, the transition probability of the equipment under different operating states is calculated to generate abnormal state transition probability data of the equipment. These transition probabilities can help predict the probability of equipment failure and provide an important reference for equipment maintenance and scheduling.

[0099] Preferably, performing multi-dimensional graph conversion on power plant equipment operation monitoring data includes:

[0100] Perform power plant equipment operation signal conversion on power plant equipment operation monitoring data to generate power plant equipment operation signals; perform waveform conversion on power plant equipment operation signals to generate power plant equipment operation signal waveforms; perform time domain analysis on power plant equipment operation signal waveforms to generate power plant equipment time domain signal waveforms;

[0101] Perform order ratio analysis on the time domain signal waveform of power plant equipment to generate an order ratio diagram of power plant equipment;

[0102] Perform signal waveform extraction on the time domain signal waveform of power plant equipment to obtain the time domain signal waveform curve of power plant equipment; perform FFT transformation on the time domain signal waveform curve of power plant equipment to generate the frequency domain waveform curve of power plant equipment; perform power spectrum analysis on the frequency domain waveform curve of power plant equipment to generate the power spectrum diagram of power plant equipment;

[0103] Extract equipment influencing parameters from the power spectrum of power plant equipment to obtain power influencing parameters of power plant equipment; perform waterfall chart analysis on the power spectrum of power plant equipment based on the power influencing parameters of power plant equipment to generate a waterfall chart of power plant equipment;

[0104] The time domain signal waveform diagram of power plant equipment, the order ratio diagram of power plant equipment, the power spectrum diagram of power plant equipment and the waterfall diagram of power plant equipment are integrated to generate a multi-dimensional diagram of power plant equipment operation.

[0105] In an embodiment of the present invention, key signal information is extracted from raw monitoring data (such as temperature, pressure, vibration, and current) collected from power plant equipment. Preliminary processing of the monitoring data, including data cleaning and noise removal, is performed to ensure the accuracy of the collected signals. The operating signals of the power plant equipment are then generated, typically comprising a multi-parameter signal sequence of the equipment, such as equipment temperature and current fluctuations. The operating signals of the power plant equipment are converted into waveform graphs, displaying their changes over time. Waveform graphs can intuitively display the dynamic behavior of the equipment during operation, such as startup, shutdown, and load changes. Time-domain waveform graphs can be used to preliminarily identify abnormal fluctuations in the equipment. Time-domain analysis is performed on the waveform graphs of the power plant equipment operating signals, focusing on the patterns of signal changes along the time axis. Time-domain analysis involves filtering and smoothing the signals to remove noise and highlight the actual fluctuations of the equipment. The resulting time-domain signal waveform graphs can help detect periodic changes and sudden anomalies in equipment operation. Order ratio analysis is performed on the time-domain signal waveform graphs. In particular, in vibration signal analysis, order ratio analysis can reveal the operating conditions of the equipment at specific operating frequencies. An order ratio plot displays the energy distribution corresponding to each order of a device (e.g., 1X, 2X, and 3X frequency components), helping to identify potential mechanical issues such as imbalance and misalignment. Signal waveforms for specific time periods can be extracted from a time-domain signal waveform. Signal waveform extraction can select key time windows or peak fluctuation regions for more detailed analysis of device operation. A fast Fourier transform (FFT) is performed on the time-domain signal waveforms of power plant equipment to convert them to the frequency domain. The FFT transform converts the time-domain signal into frequency components, helping to analyze the device's behavior at different frequencies and identify frequency characteristics such as vibration and noise. The resulting frequency-domain waveform displays the intensity distribution of the device across different frequency ranges. Power spectrum analysis is performed on the frequency-domain waveforms of power plant equipment to understand the signal's power distribution. Power spectrum analysis reveals the energy distribution of the device at different frequencies and identifies abnormal frequency bands, such as peaks at specific frequencies that are associated with device failures. Key parameters affecting device health and performance, such as vibration amplitude and harmonic components, can be extracted from the power spectrum. These parameters reveal the device's performance at specific frequencies and help analyze the device's health. Based on the power impact parameters, a waterfall chart analysis is performed on the power spectrum of the equipment. The waterfall chart can show how the frequency characteristics of the equipment change over time. The generated waterfall chart of the power plant equipment helps to reveal the changes in the frequency characteristics of the equipment during long-term operation, such as frequency drift during equipment aging. The time domain signal waveform chart, order ratio chart, power spectrum chart and waterfall chart generated above are integrated to form a multi-dimensional equipment operation map. After the multi-dimensional map is integrated, it can provide a full range of perspectives for equipment health monitoring, including time changes, frequency analysis, signal strength and multi-level display of equipment status. The final generated multi-dimensional map of power plant equipment operation provides support for subsequent equipment health assessment, fault prediction and maintenance decisions.

[0106] Preferably, dividing the power plant equipment operation monitoring data into equipment operation modes according to the equipment operating condition trend fluctuation data includes:

[0107] Segmenting the equipment operating condition trend fluctuation data into adjacent fluctuation curve segments based on a preset time range to obtain a first equipment operating condition trend fluctuation curve segment and a second equipment operating condition trend fluctuation curve segment; and calculating trend fluctuation peak values ​​for the first equipment operating condition trend fluctuation curve segment and the second equipment operating condition trend fluctuation curve segment to obtain a first fluctuation curve peak value and a second fluctuation curve peak value.

[0108] Calculating the frequency domain signal ratio of the first fluctuation curve peak value and the second fluctuation curve peak value to obtain first fluctuation curve frequency domain signal ratio data and second fluctuation curve frequency domain signal ratio data, wherein the frequency domain signal includes a low-frequency signal and a high-frequency signal; performing a frequency characteristic analysis based on the first fluctuation curve frequency domain signal ratio data and the first fluctuation curve peak value to generate first frequency fluctuation characteristic data; performing a frequency characteristic analysis based on the second fluctuation curve frequency domain signal ratio data and the second fluctuation curve peak value to generate second frequency fluctuation characteristic data;

[0109] The first frequency fluctuation characteristic data and the second frequency fluctuation characteristic data are subjected to similarity calculation to obtain frequency fluctuation characteristic similarity data; the frequency fluctuation characteristic similarity data is compared with a preset standard similarity threshold value; when the frequency fluctuation characteristic similarity data is greater than or equal to the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a steady-state operation mode; when the frequency fluctuation characteristic similarity data is less than the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a non-steady-state operation mode.

[0110] In an embodiment of the present invention, fluctuation data within adjacent time periods is extracted from equipment operating condition trend fluctuation data. The fluctuation data is segmented according to preset time ranges to generate multiple independent fluctuation curve segments: a first equipment operating condition trend fluctuation curve segment and a second equipment operating condition trend fluctuation curve segment. Each curve segment represents the operating trend fluctuation of the equipment within a specific time period. Peak analysis is performed on the first equipment operating condition trend fluctuation curve segment, and the maximum value is extracted to generate a first fluctuation curve peak value. The same peak analysis is performed on the second equipment operating condition trend fluctuation curve segment, and the maximum value is extracted to generate a second fluctuation curve peak value. Frequency domain conversion is performed on the first and second fluctuation curve peak values ​​to decompose them into low-frequency and high-frequency signals. The signal proportions of each frequency band are calculated to generate the following data: first fluctuation curve frequency domain signal proportion data and second fluctuation curve frequency domain signal proportion data. Frequency characteristics analysis is performed on the first fluctuation curve frequency domain signal proportion data and the first fluctuation curve peak value to identify the main frequency components and their variation characteristics, generating first frequency fluctuation characteristic data. The same frequency characteristics analysis is performed on the second fluctuation curve frequency domain signal proportion data and the second fluctuation curve peak value to generate second frequency fluctuation characteristic data. The first frequency fluctuation characteristic data and the second frequency fluctuation characteristic data are compared and their similarity is calculated using a similarity algorithm to generate frequency fluctuation characteristic similarity data, which is used to measure the degree of similarity between the two trend fluctuation curves in terms of frequency characteristics. The frequency fluctuation characteristic similarity data is compared with a preset standard similarity threshold. If the frequency fluctuation characteristic similarity data is greater than or equal to the standard similarity threshold, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as being in a steady-state operating mode. If the frequency fluctuation characteristic similarity data is less than the standard similarity threshold, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as being in a non-steady-state operating mode.

[0111] Preferably, performing abnormal state transition analysis on the non-steady-state warning data of power plant equipment includes:

[0112] Extract warning parameters from the non-steady-state warning data of power plant equipment to obtain non-steady-state warning parameters of power plant equipment; confirm warning timestamps of non-steady-state warning parameters of power plant equipment to obtain non-steady-state warning timestamps of power plant equipment;

[0113] Dynamically screen the warning parameters of the power plant equipment's non-steady-state warning parameters with adjacent timestamps using the non-steady-state warning timestamps to obtain warning correlation parameters; perform abnormal state correlation analysis on the non-steady-state warning parameters of the power plant equipment using the warning correlation parameters to generate non-steady-state abnormal state correlation data for the power plant equipment;

[0114] A Markov chain is constructed on the correlation data of the non-steady abnormal state of power plant equipment to generate a power plant equipment state transition graph; and edge distance calculation is performed on the power plant equipment state transition graph to generate the non-steady abnormal state transition probability data of power plant equipment.

[0115] In this embodiment of the present invention, by adding timestamp information to each warning parameter and recording the data generation time, time series data is generated, generating power plant equipment non-steady-state warning timestamps to identify the temporal characteristics of each abnormal state. Using the power plant equipment non-steady-state warning timestamps, warning parameters within adjacent time periods are selected to form dynamic associations. A screening algorithm is applied to filter invalid or noisy parameters to ensure data accuracy, generating warning association parameters to reflect the correlation between temporally adjacent states. Based on the warning association parameters, statistical analysis is performed on non-steady-state warning parameters within different time periods to identify correlation patterns. Causal or co-occurrence relationships between abnormal states are established to generate power plant equipment non-steady-state abnormal state association data to describe the relationship characteristics between abnormal states. The power plant equipment non-steady-state abnormal state association data is modeled as a finite set of states. Each abnormal state is defined as a node, and transition relationships between states are established. Using Markov chain theory, a power plant equipment state transition graph is generated to reflect the transition paths of abnormal states. Distance calculation is performed on the edges of the state transition graph (i.e., transitions between abnormal states), taking into account factors such as transition probabilities between states, time intervals, and parameter change amplitudes. The weighted formula is used for comprehensive calculation to generate the probability data of non-steady-state abnormal state transition of power plant equipment, describing the possibility and intensity of transition between various abnormal states.

[0116] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0117] Step S31: Acquire rated life data of power plant equipment;

[0118] Step S32: performing a power plant equipment operating load analysis on the multi-dimensional graph of power plant equipment operation based on the steady-state operating mode of the equipment to generate steady-state operating power plant equipment operating load data; performing an equipment loss analysis on the rated life data of the power plant equipment based on the steady-state operating power plant equipment operating load data to generate steady-state operating mode loss data of the power plant equipment;

[0119] Step S33: dividing the data sets of the power plant equipment's non-steady-state abnormal state transition probability data and the power plant equipment's steady-state operation mode loss data into a model training set and a model test set;

[0120] Step S34: Perform model training on the model training set using a feedforward neural network algorithm to generate a power plant equipment health assessment pre-model; perform model optimization iteration on the power plant equipment health assessment pre-model using the model test set to generate a power plant equipment health assessment model.

[0121] In an embodiment of the present invention, rated life data for power plant equipment is obtained from equipment operating manuals, historical maintenance records, or manufacturer-provided information. This data includes the equipment's theoretical service life, rated operating hours, and durability indicators for related operating parameters. Based on the equipment's time-domain signals, frequency-domain signals, and power parameters in a multi-dimensional graph of power plant equipment operation, combined with the equipment's steady-state operating mode, a multi-angle analysis of the equipment's operating load is performed to generate steady-state power plant equipment operating load data, reflecting actual operating conditions under steady-state operation. Using this steady-state power plant equipment operating load data, a loss analysis is performed on the rated life data of the power plant equipment to assess the impact of the operating load on its lifespan, such as the lifespan reduction rate caused by overload. Equipment loss per unit time is calculated to generate steady-state power plant equipment loss data, providing a basis for equipment health assessment. The power plant equipment's non-steady-state abnormal state transition probability data is integrated with the steady-state loss data to form a comprehensive dataset containing both steady-state and non-steady-state operating information. The comprehensive dataset is divided into a model training set and a model testing set according to a specific ratio (e.g., 80% training set, 20% test set). Ensure the diversity and representativeness of the dataset to avoid overfitting or underfitting during model training. Using a feedforward neural network (FFNN) algorithm, with the model training set as input, a multi-layer neural network structure is designed, consisting of an input layer, hidden layers, and an output layer. The input layer contains key features related to equipment health, such as operating load, loss rate, and transition probability. The hidden layer uses activation functions (such as ReLU or Sigmoid) for nonlinear transformation to capture complex feature relationships. The output layer generates equipment health scores or health status classification results, forming a preliminary power plant equipment health assessment pre-model. The performance of the power plant equipment health assessment pre-model is tested using a model test set to evaluate prediction accuracy and generalization ability. Optimization is performed through the following methods: adjusting the number of neural network layers and neurons. Using the cross-entropy loss function or mean squared error (MSE) for error feedback. Applying the Adam optimizer or SGD (stochastic gradient descent) for weight updates. Through multiple rounds of iterative optimization, a highly accurate and robust power plant equipment health assessment model is ultimately generated.

[0122] Preferably, step S4 includes the following steps:

[0123] Step S41: Importing the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate power plant equipment health assessment data;

[0124] Step S42: Visualize the power plant equipment health assessment data to generate a power plant equipment health assessment report; construct equipment repair decisions based on the power plant equipment health assessment report to perform power plant equipment health operations.

[0125] In this embodiment of the present invention, data on the probability of transitions from non-steady-state abnormal states and loss data from steady-state operation modes of power plant equipment are imported into a power plant equipment health assessment model. The imported data is standardized to ensure balanced weighting of the two types of data in the model for comprehensive analysis. Using the power plant equipment health assessment model, a comprehensive assessment is performed through the following steps: The non-steady-state abnormal state transition probability data is scored for state stability. The cumulative loss rate and health risk of the equipment are calculated in combination with the loss data from steady-state operation modes. The output layer of the health assessment model generates power plant equipment health assessment data (e.g., health scores or health status classifications). This comprehensive health assessment data reflects the current health status of the equipment and provides a basis for further action. Graphical tools (such as line charts, pie charts, and heat maps) are used to visually display the equipment health assessment data in a power plant equipment health assessment report. Based on the health status displayed in the assessment report, the following decision logic is constructed: Healthy state: No repair required, only routine maintenance. Sub-healthy state: Develop a preventive maintenance plan and schedule equipment optimization and adjustments. Faulty state: Generate a repair priority list and schedule emergency repairs. Use a maintenance management system or intelligent scheduling platform to automatically generate equipment repair plans, including repair time, required resources, and steps. Re-enter the health assessment model with completed repair data to verify the repair results and optimize the model's decision-making capabilities.

[0126] In this specification, a system for constructing a power plant equipment health assessment model is provided, which is used to execute the above-mentioned method for constructing a power plant equipment health assessment model. The system for constructing a power plant equipment health assessment model includes:

[0127] The data integration module is used to obtain the site information data of the power plant equipment; configure the site equipment data transmission channel for the power plant equipment site information data to generate the site equipment data transmission channel; use the site equipment data transmission channel to integrate the power plant equipment site information data to generate the power plant equipment operation data integration platform;

[0128] The operating condition analysis module is used to monitor the operation of power plant equipment on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; perform equipment multi-operating condition trend fluctuation calculations on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; classify the power plant equipment operation monitoring data into equipment operating modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; perform equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data;

[0129] The health prediction module is used to obtain the rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate the steady-state operation mode loss data of the power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment to generate the power plant equipment health assessment model;

[0130] The evaluation and decision-making module is used to import the non-steady-state abnormal state transition probability data of power plant equipment and the loss data of the steady-state operation mode of power plant equipment into the power plant equipment health assessment model to conduct a comprehensive health assessment and generate a power plant equipment health assessment report; based on the power plant equipment health assessment report, equipment repair decision-making is constructed to execute power plant equipment health operations.

[0131] The beneficial effects of the present invention lie in the efficient integration and real-time data transmission of multi-site equipment information through the configuration of site equipment data transmission channels and the establishment of a power plant equipment operation data integration platform, ensuring data integrity and accuracy. The use of this integrated platform simplifies the data management process, provides unified data support for real-time monitoring and subsequent analysis of equipment status, and improves overall operational efficiency. Equipment operation monitoring, combined with multi-operating condition trend fluctuation calculations, comprehensively captures dynamic changes in equipment operating status and accurately identifies steady-state and unsteady-state operating modes. The classification of operating modes effectively distinguishes between normal and abnormal equipment operating states, providing a foundation for subsequent abnormality analysis and health assessment. Abnormal state transition probability analysis, using unsteady-state operating mode data, quantifies the probability of equipment transitioning from one state to another. Markov chain analysis models provide a scientific basis for predicting equipment operation, helping managers identify potential risks in advance and take preventive measures. Based on equipment steady-state operating modes and unsteady abnormal state transition probability data, and taking into account equipment losses and abnormal conditions, a health assessment model is constructed. This assessment method can comprehensively quantify the health status of equipment. The construction of this health assessment model provides a scientific equipment status diagnostic tool that can accurately assess the current operating status and remaining life of the equipment. Health assessment reports visually display the operational and health status of equipment, providing a clear scientific basis for repair or replacement. Repair decisions, combined with health assessment data, can dynamically adjust maintenance plans, avoid blind repairs, and optimize resource allocation. Therefore, this invention improves the accuracy of power plant equipment health assessment models through data integration, refined operating condition analysis, analysis of non-steady-state abnormal state transitions, and dynamic loss assessment.

[0132] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0133] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a power plant equipment health assessment model, characterized in that: The following steps are involved: Step S1: Obtaining power plant equipment site information data; Configure the site equipment data transmission channel for the power plant equipment site information data and generate the site equipment data transmission channel; Use the site equipment data transmission channel to integrate the power plant equipment site information data and generate a power plant equipment operation data integration platform; Step S2: monitoring the power plant equipment operation on the power plant equipment operation data integration platform, thereby obtaining power plant equipment operation monitoring data; Calculate the equipment multi-operating condition trend fluctuations on the power plant equipment operation monitoring data to obtain the equipment operating condition trend fluctuation data; According to the equipment operating condition trend fluctuation data, the power plant equipment operation monitoring data is divided into equipment operation modes, and the equipment steady-state operation mode and equipment unsteady-state operation mode are generated; Based on the non-steady-state operation mode of the equipment, an abnormal state transfer analysis of the equipment is performed to generate the non-steady-state abnormal state transfer probability data of the power plant equipment, wherein the generation of the non-steady-state abnormal state transfer probability data of the power plant equipment includes: based on the non-steady-state operation mode of the equipment, monitoring and early warning data of the multi-dimensional map of the power plant equipment operation is screened to obtain the non-steady-state early warning data of the power plant equipment; abnormal state transfer analysis of the non-steady-state early warning data of the power plant equipment is performed to generate the non-steady-state abnormal state transfer probability data of the power plant equipment; wherein the abnormal state transfer analysis of the non-steady-state early warning data of the power plant equipment includes: Extract warning parameters from the non-steady-state warning data of power plant equipment to obtain non-steady-state warning parameters of power plant equipment; confirm warning timestamps of non-steady-state warning parameters of power plant equipment to obtain non-steady-state warning timestamps of power plant equipment; Dynamically screen the warning parameters of the power plant equipment's non-steady-state warning parameters with adjacent timestamps using the non-steady-state warning timestamps to obtain warning correlation parameters; perform abnormal state correlation analysis on the non-steady-state warning parameters of the power plant equipment using the warning correlation parameters to generate non-steady-state abnormal state correlation data for the power plant equipment; Construct a Markov chain for the associated data of the non-steady abnormal state of power plant equipment to generate a power plant equipment state transition graph; calculate the edge distance of the power plant equipment state transition graph to generate the non-steady abnormal state transition probability data of the power plant equipment; Step S3: Acquire rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate steady-state operation mode loss data of power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of power plant equipment and the steady-state operation mode loss data of power plant equipment to generate a power plant equipment health assessment model; Step S4: Import the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate a power plant equipment health assessment report; construct an equipment repair decision based on the power plant equipment health assessment report to perform power plant equipment health operations.

2. The method for constructing a power plant equipment health assessment model according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Acquire power plant equipment site information data; Step S12: performing data preprocessing on the power plant equipment site information data to generate standard power plant equipment site information data, wherein the data preprocessing includes data cleaning, data denoising, data missing value filling and data standardization; Step S13: configuring a site equipment data transmission channel for the standard power plant equipment site information data to generate a site equipment data transmission channel; Step S14: Use the site equipment data transmission channel to deploy the edge perception computing gateway for the standard power plant equipment site information data to generate edge perception computing gateway deployment data; integrate the standard power plant equipment site information data based on the edge perception computing gateway deployment data to generate a power plant equipment operation data integration platform.

3. The method for constructing a power plant equipment health assessment model according to claim 2, characterized in that: The construction method of the power plant equipment health assessment model includes: Based on the edge perception computing gateway deployment data, the standard power plant equipment site information data is uploaded to the edge perception computing gateway for data spatiotemporal alignment to generate power plant equipment site spatiotemporal alignment data; Perform data characteristic analysis on the spatiotemporal alignment data of power plant equipment sites to generate data characteristics of power plant equipment sites; perform weighted fusion of associated data on the spatiotemporal alignment data of power plant equipment sites based on the data characteristics of power plant equipment sites to generate associated fusion data of power plant equipment sites; The associated integrated data of power plant equipment sites are connected to the cloud platform to generate a power plant equipment operation data integration platform.

4. The method for constructing a power plant equipment health assessment model according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing power plant equipment operation monitoring on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; performing multi-dimensional graph conversion on the power plant equipment operation monitoring data to generate a multi-dimensional graph of power plant equipment operation; Step S22: performing comparative analysis of equipment multi-operating condition related trends on the power plant equipment operation monitoring data according to the multi-dimensional graph of power plant equipment operation, and generating equipment multi-operating condition trend comparative analysis data; Step S23: performing trend fluctuation calculation on the equipment multi-operating condition trend comparison analysis data to obtain equipment operating condition trend fluctuation data; dividing the power plant equipment operation monitoring data into equipment operation modes according to the equipment operating condition trend fluctuation data to generate equipment steady-state operation mode and equipment unsteady-state operation mode.

5. The method for constructing a power plant equipment health assessment model according to claim 4, characterized in that: The multi-dimensional graph conversion of power plant equipment operation monitoring data includes: Perform power plant equipment operation signal conversion on power plant equipment operation monitoring data to generate power plant equipment operation signals; perform waveform conversion on power plant equipment operation signals to generate power plant equipment operation signal waveforms; perform time domain analysis on power plant equipment operation signal waveforms to generate power plant equipment time domain signal waveforms; Perform order ratio analysis on the time domain signal waveform of power plant equipment to generate an order ratio diagram of power plant equipment; Perform signal waveform extraction on the time domain signal waveform of power plant equipment to obtain the time domain signal waveform curve of power plant equipment; perform FFT transformation on the time domain signal waveform curve of power plant equipment to generate the frequency domain waveform curve of power plant equipment; perform power spectrum analysis on the frequency domain waveform curve of power plant equipment to generate the power spectrum diagram of power plant equipment; Extract equipment influencing parameters from the power spectrum of power plant equipment to obtain power influencing parameters of power plant equipment; perform waterfall chart analysis on the power spectrum of power plant equipment based on the power influencing parameters of power plant equipment to generate a waterfall chart of power plant equipment; The time domain signal waveform diagram of power plant equipment, the order ratio diagram of power plant equipment, the power spectrum diagram of power plant equipment and the waterfall diagram of power plant equipment are integrated to generate a multi-dimensional diagram of power plant equipment operation.

6. The method for constructing a power plant equipment health assessment model according to claim 4, characterized in that: The equipment operation mode classification of power plant equipment operation monitoring data based on equipment operating condition trend fluctuation data includes: Segmenting the equipment operating condition trend fluctuation data into adjacent fluctuation curve segments based on a preset time range to obtain a first equipment operating condition trend fluctuation curve segment and a second equipment operating condition trend fluctuation curve segment; and calculating trend fluctuation peak values ​​for the first equipment operating condition trend fluctuation curve segment and the second equipment operating condition trend fluctuation curve segment to obtain a first fluctuation curve peak value and a second fluctuation curve peak value. Calculating the frequency domain signal ratio of the first fluctuation curve peak value and the second fluctuation curve peak value to obtain first fluctuation curve frequency domain signal ratio data and second fluctuation curve frequency domain signal ratio data, wherein the frequency domain signal includes a low-frequency signal and a high-frequency signal; performing a frequency characteristic analysis based on the first fluctuation curve frequency domain signal ratio data and the first fluctuation curve peak value to generate first frequency fluctuation characteristic data; performing a frequency characteristic analysis based on the second fluctuation curve frequency domain signal ratio data and the second fluctuation curve peak value to generate second frequency fluctuation characteristic data; The first frequency fluctuation characteristic data and the second frequency fluctuation characteristic data are subjected to similarity calculation to obtain frequency fluctuation characteristic similarity data; the frequency fluctuation characteristic similarity data is compared with a preset standard similarity threshold value; when the frequency fluctuation characteristic similarity data is greater than or equal to the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a steady-state operation mode; when the frequency fluctuation characteristic similarity data is less than the preset standard similarity threshold value, the power plant equipment corresponding to the equipment operating condition trend fluctuation data is marked as a non-steady-state operation mode.

7. The method for constructing a power plant equipment health assessment model according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: Acquire rated life data of power plant equipment; Step S32: performing a power plant equipment operating load analysis on the multi-dimensional graph of power plant equipment operation based on the steady-state operating mode of the equipment to generate steady-state operating power plant equipment operating load data; performing an equipment loss analysis on the rated life data of the power plant equipment based on the steady-state operating power plant equipment operating load data to generate steady-state operating mode loss data of the power plant equipment; Step S33: dividing the data sets of the power plant equipment's non-steady-state abnormal state transition probability data and the power plant equipment's steady-state operation mode loss data into a model training set and a model test set; Step S34: Perform model training on the model training set using a feedforward neural network algorithm to generate a power plant equipment health assessment pre-model; perform model optimization iteration on the power plant equipment health assessment pre-model using the model test set to generate a power plant equipment health assessment model.

8. The method for constructing a power plant equipment health assessment model according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Importing the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment into the power plant equipment health assessment model to perform a comprehensive health assessment and generate power plant equipment health assessment data; Step S42: Visualize the power plant equipment health assessment data to generate a power plant equipment health assessment report; construct equipment repair decisions based on the power plant equipment health assessment report to perform power plant equipment health operations.

9. A system for constructing a power plant equipment health assessment model, characterized in that: A method for constructing a power plant equipment health assessment model according to claim 1, wherein the system for constructing the power plant equipment health assessment model comprises: The data integration module is used to obtain the site information data of the power plant equipment; configure the site equipment data transmission channel for the power plant equipment site information data to generate the site equipment data transmission channel; use the site equipment data transmission channel to integrate the power plant equipment site information data to generate the power plant equipment operation data integration platform; The operating condition analysis module is used to monitor the operation of power plant equipment on the power plant equipment operation data integration platform to obtain power plant equipment operation monitoring data; perform equipment multi-operating condition trend fluctuation calculations on the power plant equipment operation monitoring data to obtain equipment operating condition trend fluctuation data; classify the power plant equipment operation monitoring data into equipment operating modes based on the equipment operating condition trend fluctuation data to generate equipment steady-state operation modes and equipment unsteady-state operation modes; perform equipment abnormal state transition analysis based on the equipment unsteady operation mode to generate power plant equipment unsteady-state abnormal state transition probability data; The health prediction module is used to obtain the rated life data of power plant equipment; perform equipment loss analysis on the rated life data of power plant equipment based on the steady-state operation mode of the equipment to generate the steady-state operation mode loss data of the power plant equipment; construct a health assessment model based on the non-steady-state abnormal state transition probability data of the power plant equipment and the steady-state operation mode loss data of the power plant equipment to generate the power plant equipment health assessment model; The evaluation and decision-making module is used to import the non-steady-state abnormal state transition probability data of power plant equipment and the loss data of the steady-state operation mode of power plant equipment into the power plant equipment health assessment model to conduct a comprehensive health assessment and generate a power plant equipment health assessment report; based on the power plant equipment health assessment report, equipment repair decision-making is constructed to execute power plant equipment health operations.

Images