A power plant data management method and system based on multi-source data

Abstract

The present invention discloses a power plant data management method and system based on multi-source data, which relates to the technical field of power plant operation and maintenance management, including: arranging multi-source sensors to collect power plant data, transmitting power plant data through edge computing nodes, and performing data preprocessing; using the preprocessed data to evaluate the current operating status of the power plant and identify changes in power plant data; performing equipment status detection according to the identification results and generating a maintenance plan. The power plant data management method based on multi-source data provided by the present invention realizes comprehensive and real-time monitoring of the operating status of the power plant by integrating multi-source data. The real-time and accuracy of data processing are improved, and data transmission delays are reduced. Abnormal changes are identified in a timely manner, and equipment status detection and fault diagnosis are provided. Abnormal diagnosis reports are generated, maintenance resource allocation is optimized, maintenance efficiency and effects are improved, and maintenance costs are reduced. Predictive maintenance is achieved, potential failures are prevented, and the safety and reliability of power plant operation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power plant operation and maintenance management, and in particular to a power plant data management method and system based on multi-source data. Background Art

[0002] At present, power plants are important infrastructure for energy supply, and their stability and operating efficiency directly affect the quality and cost of energy supply. However, the equipment and systems of power plants are complex and diverse, involving a large amount of temperature, pressure, flow, power output and other data, which are usually scattered in different equipment and systems, making it difficult to monitor and manage them uniformly and effectively.

[0003] Traditional power plant data management methods mainly rely on a single data source and manual inspection methods, which can only monitor some key equipment and cannot fully monitor the entire power plant. Data processing is mostly concentrated on central servers, with data transmission delays and slow processing speeds, making it difficult to respond to emergencies in a timely manner.

[0004] Due to the lack of effective data analysis methods, equipment anomalies and failures are often not discovered in time, which can easily lead to safety hazards and economic losses. Maintenance management relies more on experience and regular inspections, and lacks accurate assessment and prediction of equipment status, resulting in unreasonable allocation of maintenance resources and increased maintenance costs.

[0005] With the development of the Internet of Things, edge computing and big data technologies, power plant data management has new technical means and solutions. Therefore, a power plant data management method based on multi-source data is urgently needed to improve the operation monitoring and maintenance management level of power plants. Summary of the invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by the present invention is: the existing power plant data management method has the problems of incomplete data collection, low data processing efficiency, untimely abnormality detection and lack of scientific maintenance management, as well as how to achieve real-time monitoring of the power plant operating status and optimize maintenance management.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a power plant data management method based on multi-source data, comprising:

[0009] Deploy multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing;

[0010] Use preprocessed data to evaluate the current operating status of the power plant and identify changes in power plant data;

[0011] Perform equipment status detection based on the identification results and generate a maintenance plan.

[0012] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, wherein: the power plant data includes temperature data, pressure data, flow data, power output data, equipment log data and environmental data;

[0013] The data preprocessing includes cleaning the data at the edge node, removing duplicate, erroneous and missing data, and performing data standardization.

[0014] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, wherein: the identification of power plant data changes includes calculating the pre-processed data through a primary judgment model, a secondary judgment model, a tertiary judgment model and a quaternary judgment model, and outputting a prediction result;

[0015] The first-level judgment model is expressed as:

[0016] ,

[0017] in, represents the thermodynamic state parameter, Represents temperature data, Indicates pressure data;

[0018] The secondary judgment model is expressed as:

[0019] ,

[0020] in, represents the fluid dynamics state parameter, Indicates traffic data;

[0021] The three-level judgment model is expressed as,

[0022] ,

[0023] in, Indicates the power output status parameter, Indicates power output data;

[0024] The four-level judgment model is expressed as, ,

[0025] Among them, Env represents the environmental impact state parameter, Indicates ambient temperature data, Indicates reference temperature data;

[0026] like At the threshold and If At the threshold and If At the threshold and If At the threshold and If the value is between 1 and 1, it is marked as 1, otherwise it is marked as 0;

[0027] According to the marking results, the power plant data changes are classified into normal state, abnormal thermodynamic parameters, abnormal fluid dynamic parameters, abnormal power output parameters, abnormal environmental impact parameters and global power plant failure;

[0028] If all four marking results are 1, it is judged to be a normal state; if only one of the four marking results is 0, it is judged as a single abnormality based on the abnormal situation corresponding to the mark 0; if the number of marks 0 in the four marking results is greater than or equal to 2 and less than or equal to 3, it is determined to be multiple abnormalities; if all four-level marking results are 0, it is judged to be a global fault of the power plant.

[0029] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, the equipment status detection includes extracting the judgment model output corresponding to a single abnormality or multiple abnormalities, using a random forest model to detect the equipment status, and using a random forest model to detect and diagnose the equipment status;

[0030] Perform data cleaning, standardization and normalization on the extracted parameter values;

[0031] Input the parameter values ​​extracted in real time into the trained random forest model to predict the output results;

[0032] If a single anomaly is detected, the extracted abnormal parameter data is cleaned, standardized and normalized; multiple regression analysis is performed, historical data and real-time data of abnormal parameters and single abnormality-related parameters are collected, and a multiple regression model is established. The abnormal parameters are used as dependent variables and the single abnormality-related parameters are used as independent variables. Regression analysis is performed to determine the degree of influence of the related parameters on the abnormal parameters, and the cause and location of the abnormality are identified; time series analysis is performed, historical time series data of abnormal parameters are collected, a time series model is established, and the change trend of abnormal parameters in historical data is analyzed to predict future changes in abnormal parameters; combining the results of multiple regression analysis and time series analysis, a single abnormal device status diagnosis report is generated, and the single abnormal device status diagnosis report includes a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range;

[0033] If multiple anomalies are detected, the extracted data of multiple abnormal parameters are cleaned, standardized and normalized; implicit relationship analysis is performed, historical data and real-time data of multiple abnormal parameters and multiple abnormal related parameters are collected, and a multiple regression model is established. Multiple abnormal parameters are used as dependent variables and multiple abnormal related parameters are used as independent variables to perform regression analysis, determine the relationship between the related parameters, and identify possible core anomalies; using the regression coefficients and residuals obtained from the multiple regression analysis, a clustering algorithm is used to perform cluster analysis on multiple abnormal parameters, identify abnormal patterns, and determine whether there are implicit associations; if the analysis results show that there is a core abnormal parameter that causes other parameter abnormalities, it is treated as a single anomaly and transferred to the single anomaly processing process; If there are indeed multiple independent anomalies, perform causal analysis on each abnormal parameter to determine the causal relationship between each abnormal parameter and multiple abnormality-related parameters, identify the cause of the anomaly, and determine the specific location and impact range of each anomaly in combination with the causal analysis results; perform time series analysis on each independent abnormal parameter, collect its historical time series data, establish a time series model, analyze the historical change trend of each parameter, and predict future changes; finally, combine the results of causal analysis and time series analysis to generate a multiple abnormal device status diagnostic report, which includes multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impact.

[0034] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, wherein: the multivariate regression model is expressed as,

[0035] ,

[0036] in, Indicates abnormal parameters, , Indicates Multiple anomaly related parameters, , Indicates The intercept of the outlier parameter, Indicates The abnormal parameters and The regression coefficients of multiple abnormal correlation parameters, Indicates The regression error term of the abnormal parameters;

[0037] The cluster analysis is expressed as,

[0038] ,

[0039] in, Indicates The cluster to which the abnormal parameters belong, , represents the indicator function, when , the value is 1, otherwise it is 0. Indicates The abnormal parameters and The square of the Euclidean distance between cluster centers, Indicates Cluster centers, ;

[0040] Calculate the mean of cluster centers and cluster center standard deviation ,like , it is judged that the core abnormal parameters cause other parameter abnormalities, the abnormal parameters in the cluster are determined as core abnormal parameters, and the core parameters are classified as single abnormalities; if , it is judged that there are multiple independent anomalies; if or , calculate the correlation strength between each abnormal parameter and multiple abnormal related parameters, expressed as,

[0041] ,

[0042] in, Indicates The abnormal parameters and The correlation strength of multiple abnormality-related parameters;

[0043] Treat each abnormal parameter as a network node and the correlation strength As the weight of the edge, we construct the association network, perform network structure analysis, calculate degree centrality and betweenness centrality, and express it as,

[0044] ,

[0045] ,

[0046] in, represents degree centrality, represents the betweenness centrality, Representation Node To Node The number of shortest paths, Indicates that through the node The number of shortest paths;

[0047] like or , it is judged that the core abnormal parameters cause other parameter abnormalities, the abnormal parameters in the cluster are determined as core abnormal parameters, and the core parameters are classified as single abnormalities; if or , it is judged that there are multiple independent anomalies, among which, represents the mean of all node degree centrality, represents the standard deviation of all node degree centrality, represents the mean of the betweenness centrality of all nodes, Represents the standard deviation of the betweenness centrality of all nodes.

[0048] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, wherein: the generation of the maintenance plan includes extracting a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range from a single abnormal equipment status diagnosis report; using a particle swarm optimization algorithm to analyze the extracted single abnormality type, single abnormality location, single abnormality cause, future single abnormality changes and single abnormality impact range, optimize the configuration of maintenance resources, and generate a preliminary single maintenance plan; the preliminary single maintenance plan includes the maintenance time of the equipment, the required spare parts and human resource configuration; based on the preliminary single maintenance plan and the extracted single abnormality parameters, use the ARIMA model to perform time series analysis to predict the change trend of the future single abnormality parameters; based on the prediction results, evaluate the impact of different maintenance plans on the future operating status of the equipment, and select the best maintenance plan; combine historical maintenance data and real-time monitoring data, adjust the maintenance plan, formulate maintenance measures, time arrangements, required resources and technical support.

[0049] As a preferred solution of the power plant data management method based on multi-source data described in the present invention, wherein: the generation of the maintenance plan also includes extracting multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts from the multiple abnormality equipment status diagnosis report; using a genetic algorithm to analyze the extracted multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts, optimize the maintenance resource allocation of multiple equipment and components, and generate a preliminary multiple maintenance plan, the preliminary multiple maintenance plan includes the maintenance time of the equipment, the required spare parts and human resource allocation; based on the preliminary multiple maintenance plan and the extracted multiple abnormality parameters, use the LSTM model to perform time series analysis to predict the future change trend of the multiple abnormality parameters; based on the prediction results, evaluate the effects of different comprehensive maintenance plans and select the best maintenance plan; combine historical maintenance data and real-time monitoring data to determine the comprehensive maintenance plan, formulate specific maintenance measures, time arrangements, required resources and technical support.

[0050] Another object of the present invention is to provide a power plant data management system based on multi-source data, which can solve the problems of incomplete data collection, low data processing efficiency and untimely anomaly detection in existing power plant data management by constructing a power plant data management system based on multi-source data.

[0051] To solve the above technical problems, the present invention provides the following technical solutions: a power plant data management system based on multi-source data, comprising: an output acquisition module, a state identification module and a power plant maintenance module; the output acquisition module is used to arrange multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing; the state identification module is used to use the preprocessed data to evaluate the current operating status of the power plant and identify changes in power plant data; the power plant maintenance module is used to perform equipment status detection according to the identification results and generate a maintenance plan.

[0052] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the power plant data management method based on multi-source data when executing the computer program.

[0053] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the power plant data management method based on multi-source data as described above.

[0054] Beneficial effects of the present invention: The power plant data management method based on multi-source data provided by the present invention realizes comprehensive and real-time monitoring of the power plant's operating status by integrating multi-source data. Compared with traditional methods, the real-time and accuracy of data processing are improved, and data transmission delay is reduced. The operating status of the power plant is evaluated, abnormal changes can be identified in time, and equipment status detection and fault diagnosis can be provided. Abnormal diagnosis reports are generated, maintenance resource allocation is optimized, maintenance efficiency and effects are improved, and maintenance costs are reduced. Predictive maintenance is achieved, potential failures are prevented, and the safety and reliability of power plant operation are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0056] Figure 1 An overall flow chart of a power plant data management method based on multi-source data provided by an embodiment of the present invention.

[0057] Figure 2An overall structural diagram of a power plant data management system based on multi-source data provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments 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 ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0060] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a power plant data management method based on multi-source data, comprising:

[0061] Deploy multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing;

[0062] Use preprocessed data to evaluate the current operating status of the power plant and identify changes in power plant data;

[0063] Perform equipment status detection based on the identification results and generate a maintenance plan.

[0064] The power plant data includes temperature data, pressure data, flow data, power output data, equipment log data and environmental data;

[0065] The data preprocessing includes cleaning the data at the edge node, removing duplicate, erroneous and missing data, and performing data standardization.

[0066] The identifying of power plant data changes includes calculating the pre-processed data through a primary judgment model, a secondary judgment model, a tertiary judgment model, and a quaternary judgment model, and outputting a prediction result;

[0067] The first-level judgment model is expressed as:

[0068] ,

[0069] in, represents the thermodynamic state parameter, Represents temperature data, Indicates pressure data;

[0070] The secondary judgment model is expressed as:

[0071] ,

[0072] in, represents the fluid dynamics state parameter, Indicates traffic data;

[0073] The three-level judgment model is expressed as,

[0074] ,

[0075] in, Indicates the power output status parameter, Indicates power output data;

[0076] The four-level judgment model is expressed as, ,

[0077] Among them, Env represents the environmental impact state parameter, Indicates ambient temperature data, Indicates reference temperature data;

[0078] In order to maintain the data characteristics while enhancing the stability and robustness of the model, the square root transformation coefficient was found to be 0.5 through analysis of historical data. In order to find a balance point so that the power output data can reflect the actual situation without excessive fluctuations and ensure the stability of the model, the power transformation coefficient was found to be 0.3 through analysis of historical data.

[0079] like At the threshold and If At the threshold and If At the threshold and If At the threshold and If the value is between 1 and 1, it is marked as 1, otherwise it is marked as 0;

[0080] According to the marking results, the power plant data changes are classified into normal state, abnormal thermodynamic parameters, abnormal fluid dynamic parameters, abnormal power output parameters, abnormal environmental impact parameters and global power plant failure;

[0081] If it is classified as normal, all parameters are continuously monitored to ensure that all parameters are always within the normal range, and all monitoring data are recorded in the database in real time for subsequent analysis and reference.

[0082] If it is classified as a global fault of the power plant, a fault alarm will be issued immediately to notify the relevant maintenance personnel and management personnel, and the shutdown protection program will be started according to the type and severity of the fault to avoid further damage to the equipment and the expansion of the accident. At the same time, the safety and stability of the non-faulty parts will be ensured by switching and isolating the faulty equipment.

[0083] If all four marking results are 1, it is judged to be a normal state; if only one of the four marking results is 0, it is judged as a single abnormality based on the abnormal situation corresponding to the mark 0; if the number of marks 0 in the four marking results is greater than or equal to 2 and less than or equal to 3, it is determined to be multiple abnormalities; if all four-level marking results are 0, it is judged to be a global fault of the power plant.

[0084] The device status detection includes extracting the judgment model output corresponding to a single anomaly or multiple anomalies, using a random forest model to detect the device status, and using a random forest model to detect and diagnose the device status; performing data cleaning, standardization and normalization on the extracted parameter values; using historical data under normal operating conditions to train the random forest model, setting the number of trees to 100 and the depth to 10. The random forest learns the characteristic patterns under normal conditions by constructing multiple decision trees and integrating their results. Input the parameter values ​​extracted in real time into the trained random forest model to predict the output results;

[0085] If a single anomaly is detected, the extracted abnormal parameter data is cleaned, standardized and normalized; multiple regression analysis is performed, historical data and real-time data of abnormal parameters and single abnormality-related parameters are collected, and a multiple regression model is established. The abnormal parameters are used as dependent variables and the single abnormality-related parameters are used as independent variables. Regression analysis is performed to determine the degree of influence of the related parameters on the abnormal parameters, and the cause and location of the abnormality are identified; time series analysis is performed, historical time series data of abnormal parameters are collected, a time series model is established, and the change trend of abnormal parameters in historical data is analyzed to predict future changes in abnormal parameters; combining the results of multiple regression analysis and time series analysis, a single abnormal device status diagnosis report is generated, and the single abnormal device status diagnosis report includes a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range;

[0086] If multiple anomalies are detected, the extracted data of multiple abnormal parameters are cleaned, standardized and normalized; implicit relationship analysis is performed, historical data and real-time data of multiple abnormal parameters and multiple abnormal related parameters are collected, and a multiple regression model is established. Multiple abnormal parameters are used as dependent variables and multiple abnormal related parameters are used as independent variables to perform regression analysis, determine the relationship between the related parameters, and identify possible core anomalies; using the regression coefficients and residuals obtained from the multiple regression analysis, a clustering algorithm is used to perform cluster analysis on multiple abnormal parameters, identify abnormal patterns, and determine whether there are implicit associations; if the analysis results show that there is a core abnormal parameter that causes other parameter abnormalities, it is treated as a single anomaly and transferred to the single anomaly processing process; If there are indeed multiple independent anomalies, perform causal analysis on each abnormal parameter to determine the causal relationship between each abnormal parameter and multiple abnormality-related parameters, identify the cause of the anomaly, and determine the specific location and impact range of each anomaly in combination with the causal analysis results; perform time series analysis on each independent abnormal parameter, collect its historical time series data, establish a time series model, analyze the historical change trend of each parameter, and predict future changes; finally, combine the results of causal analysis and time series analysis to generate a multiple abnormal device status diagnostic report, which includes multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impact.

[0087] Single abnormality related parameters include environmental factor data, equipment operating parameters, historical maintenance records and operation records.

[0088] Multiple anomaly-related parameters include system comprehensive parameters, interaction impact data, external environment data, and system-level operation records.

[0089] The multivariate regression model is expressed as,

[0090] ,

[0091] in, Indicates abnormal parameters, , Indicates Multiple anomaly related parameters, , Indicates The intercept of the outlier parameter, Indicates The abnormal parameters and The regression coefficients of multiple abnormal correlation parameters, Indicates The regression error term of the abnormal parameters;

[0092] The cluster analysis is expressed as,

[0093] ,

[0094] in, Indicates The cluster to which the abnormal parameters belong, , represents the indicator function, when , the value is 1, otherwise it is 0. Indicates The abnormal parameters and The square of the Euclidean distance between cluster centers, Indicates Cluster centers, ;

[0095] Calculate regression coefficients Estimated by the least squares method, expressed as,

[0096] ,

[0097] in, is the design matrix, is the response vector.

[0098] Calculate residuals , the residual is the difference between the observed value and the predicted value, expressed as,

[0099] ,

[0100] Calculate cluster centers , the initial cluster center is randomly selected and iteratively updated, expressed as,

[0101] ,

[0102] in, It is The sample set contained in the clusters, It is The number of samples in a cluster.

[0103] Calculate the Euclidean distance, expressed as,

[0104] ,

[0105] Calculate the mean of cluster centers and cluster center standard deviation ,like , it is judged that the core abnormal parameters cause other parameter abnormalities, the abnormal parameters in the cluster are determined as core abnormal parameters, and the core parameters are classified as single abnormalities; if , it is judged that there are multiple independent anomalies; if or , calculate the correlation strength between each abnormal parameter and multiple abnormal related parameters, expressed as,

[0106] ,

[0107] in, Indicates The abnormal parameters and The correlation strength of multiple abnormality-related parameters;

[0108] Treat each abnormal parameter as a network node and the correlation strength As the weight of the edge, we construct the association network, perform network structure analysis, calculate degree centrality and betweenness centrality, and express it as,

[0109] ,

[0110] ,

[0111] in, represents degree centrality, represents the betweenness centrality, Representation Node To Node The number of shortest paths, Indicates that through the node The number of shortest paths;

[0112] like or , it is judged that the core abnormal parameters cause other parameter abnormalities, the abnormal parameters in the cluster are determined as core abnormal parameters, and the core parameters are classified as single abnormalities; if or , it is judged that there are multiple independent anomalies, among which, represents the mean of all node degree centrality, represents the standard deviation of all node degree centrality, represents the mean of the betweenness centrality of all nodes, Represents the standard deviation of the betweenness centrality of all nodes.

[0113] The generation of the maintenance plan includes extracting a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range from a single abnormality equipment status diagnosis report; using a particle swarm optimization algorithm to analyze the extracted single abnormality type, single abnormality location, single abnormality cause, future single abnormality changes and single abnormality impact range, optimize the configuration of maintenance resources, and generate a preliminary single maintenance plan; the preliminary single maintenance plan includes the maintenance time of the equipment, the required spare parts and human resource configuration; based on the preliminary single maintenance plan and the extracted single abnormality parameters, using the ARIMA model to perform time series analysis to predict the change trend of the future single abnormality parameters; based on the prediction results, evaluating the impact of different maintenance plans on the future operating status of the equipment, and selecting the best maintenance plan; combining historical maintenance data and real-time monitoring data, adjusting the maintenance plan, formulating maintenance measures, time arrangements, required resources and technical support.

[0114] The generation of the maintenance plan also includes extracting multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts from the multiple abnormality equipment status diagnosis report; using a genetic algorithm to analyze the extracted multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts, optimize the maintenance resource allocation of multiple equipment and components, and generate a preliminary multiple maintenance plan, the preliminary multiple maintenance plan includes the maintenance time of the equipment, the required spare parts and human resource allocation; based on the preliminary multiple maintenance plan and the extracted multiple abnormality parameters, use the LSTM model to perform time series analysis to predict the future change trend of the multiple abnormality parameters; based on the prediction results, evaluate the effects of different comprehensive maintenance plans and select the best maintenance plan; combine historical maintenance data and real-time monitoring data to determine the comprehensive maintenance plan, formulate specific maintenance measures, time arrangements, required resources and technical support.

[0115] Example 2, reference Figure 2 , is an embodiment of the present invention, and provides a power plant data management system based on multi-source data, comprising:

[0116] Output acquisition module, status identification module and power plant maintenance module;

[0117] The output acquisition module is used to arrange multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing;

[0118] The state recognition module is used to evaluate the current operating state of the power plant using preprocessed data and identify changes in power plant data;

[0119] The power plant maintenance module is used to detect equipment status based on the identification results and generate a maintenance plan.

[0120] Embodiment 3, an embodiment of the present invention, is different from the first two embodiments in that:

[0121] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0123] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0124] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0125] Example 4 is an embodiment of the present invention, which provides a power plant data management method based on multi-source data. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0126] This experiment compares two sets of equipment management systems in a power plant for 30 consecutive days. The test data types include temperature, pressure, flow, power output, equipment logs, and environmental data. Evaluation indicators include data collection coverage, data processing delay, anomaly detection timeliness, diagnosis accuracy, diagnosis time, daily maintenance cost, and failure rate.

[0127] The traditional method uses a single sensor and local automated data collection, and transmits and processes data through a central server. Data preprocessing includes basic cleaning and standardization. Equipment status evaluation and anomaly detection use rule-based judgment models, and fault diagnosis uses basic statistical analysis methods. In terms of maintenance management, maintenance resource allocation lacks systematicity based on regular maintenance and experience adjustments.

[0128] The method invented by us deploys multi-source sensors and performs data preprocessing at the edge computing node. A multi-level judgment model is used to evaluate the operating status of the power plant and identify abnormal changes. Equipment status detection and fault diagnosis are performed, and a scientific maintenance plan is generated based on the diagnosis results and optimization algorithm. The experimental results are shown in Table 1.

[0129] Table 1 Comparison of experimental results

[0130] ,

[0131] The method of the present invention uses multi-source sensors for comprehensive data collection, improves data coverage, and avoids the defect that single sensors and local automated data collection cannot fully monitor. The introduction of edge computing nodes reduces the delay time of data processing, allowing the system to respond to data changes in a more real-time manner.

[0132] Due to the limitations of the algorithm, the rule judgment model in the traditional method is not as good as the multi-level judgment model of the method of the present invention in terms of detection accuracy and real-time performance, which makes the evaluation of the power plant operating status more accurate and the abnormality detection more timely.

[0133] Random forest model and multivariate regression analysis provide higher diagnostic accuracy and faster diagnostic speed. The traditional method uses basic statistical analysis methods, and the diagnostic accuracy and speed are not as good as the algorithm used in the present invention. Based on the generated maintenance plan, the maintenance cost is reduced and the utilization rate of maintenance resources is improved. The traditional method is based on regular maintenance and experience adjustment, lacks systematic resource allocation, and leads to waste of resources and high maintenance costs. The method of the present invention optimizes the maintenance plan through analysis and prediction, and realizes more efficient resource allocation.

[0134] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power plant data management method based on multi-source data, characterized in that: include: Deploy multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing; Use preprocessed data to evaluate the current operating status of the power plant and identify changes in power plant data; Perform equipment status detection based on the identification results and generate a maintenance plan; The power plant data includes temperature data, pressure data, flow data, power output data, equipment log data and environmental data; The data preprocessing includes cleaning the data at the edge node, removing duplicate, erroneous and missing data, and performing data standardization; The identifying of power plant data changes includes calculating the pre-processed data through a primary judgment model, a secondary judgment model, a tertiary judgment model, and a quaternary judgment model, and outputting a prediction result; The first-level judgment model is expressed as: H=T·log(P+1) Among them, H represents the thermodynamic state parameter, T represents the temperature data, and P represents the pressure data; The secondary judgment model is expressed as: F=H·Q 0.5 Among them, F represents the fluid dynamics state parameter, Q represents the flow data; The three-level judgment model is expressed as, E=F·(P elec +1) 0.3 Where E represents the power output state parameter, P elec Indicates power output data; The four-level judgment model is expressed as, Among them, Env represents the environmental impact state parameter, Env temp Indicates ambient temperature data, T ref Indicates reference temperature data; If H is above the threshold H min and H max If F is between the threshold F min and F max If E is between the threshold E min and E max If Env is between the threshold Env, it is marked as 1, otherwise it is marked as 0. min and Env max If the value is between 1 and 1, it is marked as 1, otherwise it is marked as 0; According to the marking results, the power plant data changes are classified into normal state, abnormal thermodynamic parameters, abnormal fluid dynamic parameters, abnormal power output parameters, abnormal environmental impact parameters and global power plant failure; If all four marking results are 1, it is judged to be a normal state; if only one of the four marking results is 0, it is judged as a single abnormality based on the abnormal situation corresponding to the mark 0; if the number of marks 0 in the four marking results is greater than or equal to 2 and less than or equal to 3, it is determined to be multiple abnormalities; if all four-level marking results are 0, it is judged to be a global fault of the power plant.

2. The power plant data management method based on multi-source data according to claim 1, characterized in that: The device status detection includes extracting the judgment model output corresponding to a single abnormality or multiple abnormalities, using a random forest model to detect the device status, and using the random forest model to detect and diagnose the device status; Input the parameter values ​​extracted in real time into the trained random forest model to predict the output results; If a single anomaly is detected, the extracted abnormal parameter data is cleaned, standardized and normalized; multiple regression analysis is performed, historical data and real-time data of abnormal parameters and single abnormality-related parameters are collected, and a multiple regression model is established. The abnormal parameters are used as dependent variables and the single abnormality-related parameters are used as independent variables. Regression analysis is performed to determine the degree of influence of the related parameters on the abnormal parameters, and the cause and location of the abnormality are identified; time series analysis is performed, historical time series data of abnormal parameters are collected, a time series model is established, and the change trend of abnormal parameters in historical data is analyzed to predict future changes in abnormal parameters; combining the results of multiple regression analysis and time series analysis, a single abnormal device status diagnosis report is generated, and the single abnormal device status diagnosis report includes a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range; If multiple anomalies are detected, the extracted data of multiple abnormal parameters are cleaned, standardized and normalized; the data after data cleaning, standardization and normalization are analyzed for implicit relationships, historical data and real-time data of multiple abnormal parameters and multiple abnormal related parameters are collected, and a multivariate regression model is established. Multiple abnormal parameters are used as dependent variables and multiple abnormal related parameters are used as independent variables to perform regression analysis, determine the relationship between the related parameters, and identify possible core anomalies; using the regression coefficients and residuals obtained from the multivariate regression analysis, a clustering algorithm is used to perform cluster analysis on multiple abnormal parameters, identify abnormal patterns, and determine whether there are implicit associations; if the analysis results show that there is a core abnormal parameter that causes other parameter abnormalities, it is treated as a single anomaly and transferred to the single anomaly processing process; If there are indeed multiple independent anomalies, perform causal analysis on each abnormal parameter to determine the causal relationship between each abnormal parameter and multiple abnormality-related parameters, identify the cause of the anomaly, and determine the specific location and impact range of each anomaly in combination with the causal analysis results; perform time series analysis on each independent abnormal parameter, collect its historical time series data, establish a time series model, analyze the historical change trend of each parameter, and predict future changes; finally, combine the results of causal analysis and time series analysis to generate a multiple abnormal device status diagnostic report, which includes multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impact.

3. The power plant data management method based on multi-source data according to claim 2, characterized in that: The multivariate regression model is expressed as, Among them, Y i represents the i-th abnormal parameter, i=1,2,…,n, X j represents the jth multiple anomaly correlation parameter, j = 1, 2, ..., m, β i0 represents the intercept of the i-th abnormal parameter, β ij Represents the regression coefficient of the i-th abnormal parameter and the j-th multiple abnormal correlation parameter, ∈ i represents the regression error term of the i-th abnormal parameter; The cluster analysis is expressed as, Among them, c i Indicates the cluster to which the i-th abnormal parameter belongs, i = 1, 2, ..., n, 1 (c i =k) ​​represents the indicator function, when c i = k, the value is 1, otherwise it is 0, ||Y i -μ k || 2 represents the square of the Euclidean distance between the i-th anomaly parameter and the k-th cluster center, μ k represents the kth cluster center, k = 1, 2, ..., K; Calculate the mean of cluster centers and the standard deviation of the cluster center σ, if If it is judged that the core abnormal parameter causes other parameter abnormalities, the abnormal parameters in the cluster are determined as the core abnormal parameters, and the core parameters are classified as a single abnormality; It is judged that there are multiple independent anomalies; if Calculate the correlation strength between each abnormal parameter and multiple abnormal related parameters, expressed as, Among them, A ij represents the correlation strength between the ith abnormal parameter and the jth multiple abnormal correlation parameter, β ik represents the regression coefficient between the ith anomaly parameter and the kth multiple anomaly correlation parameter; Each abnormal parameter is regarded as a network node, and the correlation strength A ij As the weight of the edge, we construct the association network, perform network structure analysis, calculate degree centrality and betweenness centrality, and express it as, Among them, C D (i) represents degree centrality, C B (i) represents the betweenness centrality, σ st represents the number of shortest paths from node s to node t, σ st (i) represents the number of shortest paths passing through node i; like or If it is judged that the core abnormal parameter causes other parameter abnormalities, the abnormal parameters in the cluster are determined as the core abnormal parameters, and the core parameters are classified as a single abnormality; or It is judged that there are multiple independent anomalies, among which, represents the mean of all node degree centrality, represents the standard deviation of all node degree centrality, represents the mean of the betweenness centrality of all nodes, Represents the standard deviation of the betweenness centrality of all nodes.

4. The power plant data management method based on multi-source data according to claim 3, characterized in that: Generating the maintenance plan includes extracting a single abnormality type, a single abnormality location, a single abnormality cause, future single abnormality changes and a single abnormality impact range from a single abnormality equipment status diagnosis report; using a particle swarm optimization algorithm to analyze the extracted single abnormality type, single abnormality location, single abnormality cause, future single abnormality changes and single abnormality impact range, optimizing the configuration of maintenance resources, and generating a preliminary single maintenance plan; the preliminary single maintenance plan includes the maintenance time of the equipment, the required spare parts and the human resource configuration; Based on the preliminary single maintenance plan and the extracted single abnormal parameters, the ARIMA model is used to perform time series analysis to predict the changing trend of the single abnormal parameters in the future; Based on the prediction results, evaluate the impact of different maintenance plans on the future operating status of the equipment and select the best maintenance plan; combine historical maintenance data and real-time monitoring data to adjust the maintenance plan, formulate maintenance measures, schedules, required resources and technical support.

5. The power plant data management method based on multi-source data according to claim 4, characterized in that: The generation of the maintenance plan also includes extracting multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts from the multiple abnormality equipment status diagnosis report; using a genetic algorithm to analyze the extracted multiple abnormality types, multiple abnormality locations, multiple abnormality causes, correlation analysis results, future multiple abnormality changes and overall impacts, optimize the maintenance resource allocation of multiple equipment and components, and generate a preliminary multiple maintenance plan, the preliminary multiple maintenance plan includes the maintenance time of the equipment, the required spare parts and human resource allocation; based on the preliminary multiple maintenance plan and the extracted multiple abnormality parameters, use the LSTM model to perform time series analysis to predict the future change trend of the multiple abnormality parameters; based on the prediction results, evaluate the effects of different comprehensive maintenance plans and select the best maintenance plan; combine historical maintenance data and real-time monitoring data to determine the comprehensive maintenance plan, formulate specific maintenance measures, time arrangements, required resources and technical support.

6. A system using the power plant data management method based on multi-source data as described in any one of claims 1 to 5, characterized in that: include: Output acquisition module, status identification module and power plant maintenance module; The output acquisition module is used to arrange multi-source sensors to collect power plant data, transmit the power plant data through edge computing nodes, and perform data preprocessing; The state identification module is used to evaluate the current operating state of the power plant using the preprocessed data and identify changes in the power plant data; The power plant maintenance module is used to perform equipment status detection according to the identification results and generate a maintenance plan.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the power plant data management method based on multi-source data are implemented as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power plant data management method based on multi-source data are implemented as described in any one of claims 1 to 5.

Images