A fault early warning method and system for giant hydropower units

Abstract

The present invention discloses a method and system for early warning of faults of giant hydropower units, and relates to the field of early warning of faults. The method comprises: reading a target unit and determining distributed monitoring requirements; determining an adaptive monitoring module based on the distributed monitoring requirements; determining a three-dimensional structural characteristic and a unit operation and control mechanism based on the specification configuration information, and determining a digital virtual machine unit; retrieving multi-source unit information, building a fault knowledge graph, and building a digital early warning model based on the digital virtual machine unit and the fault knowledge graph; performing sensor monitoring and data feedback of unit partitions, and generating a fault location list; and performing terminal visual early warning on the fault location list. This method solves the technical problems of high cost caused by the need to query a large amount of data when obtaining hydropower units in the prior art, and low safety and reliability caused by slow early warning, and achieves the technical effect of improving the safety and reliability of the unit and reducing operation and maintenance costs.

Description

Technical Field

[0001] The present application relates to the technical field of fault warning, and in particular to the field of fault warning for hydropower units, and specifically to a method and system for fault warning for giant hydropower units. Background Art

[0002] With the continuous growth of global energy demand and the development of clean energy, hydropower, as an important part of renewable energy, has received extensive attention and application. As an important equipment in the hydropower field, giant hydropower units are increasing in scale and complexity, which is crucial to ensure the stable and safe operation of the power system. However, during the operation of giant hydropower units, various faults are prone to occur due to long-term, high-load operation, and complex and changeable operating environment. These faults not only affect the normal operation of the unit, but also may threaten the stability of the power system. Therefore, it is of great significance to realize real-time monitoring and early warning of giant hydropower unit faults to improve the safety and reliability of the unit. Existing early warning systems are often based on simple threshold judgment or statistical methods, lacking in-depth mining and analysis of unit operation data, so the early warning accuracy is limited and prone to false alarms or missed alarms. Traditional fault early warning methods mainly rely on manual inspections and regular maintenance, which is not only inefficient, but also difficult to achieve real-time and comprehensive monitoring of the unit's operating status.

[0003] In summary, the prior art has the problems of high cost due to the large amount of data required to query when obtaining hydropower units, and low safety and reliability due to slow early warning. Summary of the invention

[0004] Based on this, it is necessary to provide a giant hydropower unit fault warning method and system that can improve the safety and reliability of the unit and reduce the operation and maintenance costs in response to the above technical problems.

[0005] In a first aspect, a giant hydropower unit fault warning method is provided, the method comprising: reading specification configuration information of a target unit and determining multi-level risk points, determining distributed monitoring requirements, the multi-level risk points being mined based on historical operation records; based on the distributed monitoring requirements, making multi-sensor configuration decisions and sensor management decisions, determining an adaptive monitoring module, the adaptive monitoring module being communicatively connected to a sensor array; determining three-dimensional structural characteristics and unit operation and control mechanisms based on the specification configuration information, performing twin-based non-equilibrium modeling, and determining a digital virtual machine group, wherein the degree of refinement of the model partition is positively correlated with the monitoring requirements; retrieving multi-source unit information, performing information fusion analysis, and building a fault knowledge graph, the fault knowledge graph comprising a fault association graph and a fault evolution graph, and building a digital warning model based on the digital virtual machine group and the fault knowledge graph; performing sensor monitoring and data feedback of unit partitions based on the adaptive monitoring module, performing real-time fault detection and fault prediction in a predetermined time zone in the digital warning model, and generating a fault location list; and performing terminal visualization warning on the fault location list.

[0006] In the second aspect, a giant hydropower unit fault warning system is provided, the system comprising: a distributed monitoring demand determination module, the distributed monitoring demand determination module is used to read the specification configuration information of the target unit and determine the multi-level risk points, determine the distributed monitoring demand, the multi-level risk points are mined based on historical operation records; an adaptive monitoring module, the adaptive monitoring module is used to make multi-sensor configuration decisions and sensor management decisions based on the distributed monitoring needs, determine the adaptive monitoring module, the adaptive monitoring module is communicatively connected with the sensor array; a digital virtual machine group determination module, the digital virtual machine group determination module is used to determine the three-dimensional structural characteristics and unit operation and control mechanism based on the specification configuration information, perform twin non-equilibrium modeling, and determine the digital virtual machine group, wherein the refinement degree of the model partition is related to the The monitoring demand is positively correlated; a fault knowledge graph building module, the fault knowledge graph building module is used to retrieve multi-source unit information, perform information fusion analysis, and build a fault knowledge graph, the fault knowledge graph includes a fault association graph and a fault evolution graph, a digital early warning model building module, the digital early warning model building module is used to build a digital early warning model based on the digital virtual machine unit and the fault knowledge graph; a fault location single column generation module, the fault location single column generation module is used to perform sensor monitoring and data feedback of unit partitions based on the adaptive monitoring module, perform real-time fault detection and fault prediction in a predetermined time zone in the digital early warning model, and generate a fault location single column; a terminal visualization early warning module, the terminal visualization early warning module is used to perform terminal visualization early warning on the fault location single column.

[0007] According to a third aspect, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps according to the first aspect when executing the computer program.

[0008] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps described in the first aspect are implemented.

[0009] The above-mentioned giant hydropower unit fault warning method and system adopts this method to solve the technical problems of high cost caused by the need to query a lot of data when obtaining hydropower units in the prior art, and low safety and reliability caused by slow warning, thereby achieving the technical effect of improving the safety and reliability of the unit and reducing operation and maintenance costs.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A schematic diagram of a flow chart of a giant hydropower unit fault early warning method in an embodiment;

[0012] Figure 2 A schematic diagram of a flow chart of a digital virtual machine group determination of a giant hydropower unit fault early warning method in one embodiment;

[0013] Figure 3 It is a structural block diagram of a giant hydropower unit fault early warning system in one embodiment;

[0014] Figure 4 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment.

[0015] Explanation of the accompanying drawings: distributed monitoring demand determination module 11, adaptive monitoring module 12, digital virtual machine group determination module 13, fault knowledge graph construction module 14, digital early warning model construction module 15, fault location single column generation module 16, terminal visualization early warning module 17. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0017] like Figure 1As shown, the present application provides a giant hydropower unit fault early warning method, the method comprising:

[0018] Read the specification configuration information of the target unit and determine the multi-level risk points and distributed monitoring requirements, wherein the multi-level risk points are mined based on historical operation records;

[0019] Giant hydroelectric units refer to water turbine generator units with huge scale and capacity used in hydroelectric power stations. They are the core equipment of hydroelectric power stations and are responsible for converting water flow energy into electrical energy. Giant hydroelectric units are the core equipment in hydropower projects, and their operating status is directly related to the operating efficiency and safety of hydropower stations. Therefore, the maintenance, fault warning and handling of giant hydroelectric units need to be highly valued, and corresponding technologies and measures should be taken to ensure their stable operation. The present application provides a giant hydroelectric unit fault warning method, which achieves effective warning of giant hydroelectric unit faults through complex steps and decision-making processes.

[0020] The target unit refers to the giant hydropower unit that needs to be studied and is selected by the staff on their own, which is random; the specification configuration information usually covers the unit's design parameters, structural characteristics, operating conditions and other detailed information, which is used to understand the basic situation and performance characteristics of the unit; based on the specification configuration information of the target unit, the multi-level risk points of the target unit are obtained, including weak links in the unit structure, abnormal fluctuations in operating parameters, etc., which are obtained through frequently occurring problems in historical fault records. Distributed monitoring refers to the real-time monitoring and data collection of the unit's operating status by arranging sensors at key parts and potential risk points of the unit. According to the level and distribution of risk points, the monitoring requirements and sensor configurations at different locations can be determined, the specification configuration information of the target unit can be read and the multi-level risk points can be determined, and the distributed monitoring requirements can be determined, which will help the operation and maintenance team to more accurately grasp the unit's operating status and potential risks, so as to take effective measures to ensure the safe and stable operation of the unit.

[0021] Based on the distributed monitoring requirements, a multi-sensor configuration decision and a sensor management decision are made to determine an adaptive monitoring module, wherein the adaptive monitoring module is communicatively connected with the sensor array;

[0022] The configuration decision of the multi-sensor is made according to the characteristics of the unit, monitoring requirements and potential risk points. Different types of sensors have different monitoring functions and scopes of application. For example, vibration sensors can monitor the vibration of the unit, and temperature sensors can monitor the temperature changes of key parts. Sensor management decision refers to the overall management and optimization of the sensor array. In order to ensure the accuracy and real-time nature of the data, it is necessary to obtain the method of effectively managing the sensor array and determine the adaptive monitoring module. The adaptive monitoring module can automatically adjust the monitoring parameters and acquisition frequency of the sensor according to the operating status and monitoring requirements of the unit, so as to realize real-time tracking and dynamic monitoring of the unit status. Through the communication connection with the sensor array, the communication connection refers to the dedicated, virtual communication line established between two applications or devices communicating in the network to transmit messages to each other. The adaptive monitoring module can obtain the sensor data in real time, process and analyze it, and provide strong support for subsequent fault warning and diagnosis.

[0023] Identify the monitoring requirements of each distributed point, determine the sensor type and field of view, and configure a single-point array, wherein the single-point array corresponds to the distributed point one by one;

[0024] Traversing the single point array to determine a single point management mode;

[0025] Coordinate the single-point array, determine the global array, integrate the single-point management mode and configure it relatively independently, and determine the sensor management mode;

[0026] The adaptive monitoring module is managed based on the global array and the sensor management mode.

[0027] In the distributed monitoring system of giant hydropower units, the monitoring needs of each distributed point are identified. The distributed points include different parts of the turbine, key components of the generator set, cooling systems and other important operating areas. Each point has different risk points and potential failure modes, so it is necessary to determine the monitoring needs in a targeted manner. Based on the monitoring needs, the sensor type and field of view need to be determined next. The selection of sensors should be determined based on the physical quantities to be monitored, such as vibration, temperature, pressure, flow, etc., as well as the required accuracy and response time. The field of view refers to the monitoring area that the sensor can cover and the signal range that can be captured. In the configuration of a single-point array, a single-point array refers to a set of sensors configured for each distributed point, and the distributed points correspond one to one. The single-point array should be customized according to the specific monitoring needs of the point to ensure that the required data can be obtained comprehensively and accurately. Traverse the single-point array and determine the single-point management mode. The single-point management mode refers to the management strategy for sensors in each single-point array, including regulations on data collection, processing, storage and transmission, that is, for sensors configured in different partitions, the same single-point array starts and stops at the same frequency, and each point is managed relatively independently to improve the pertinence and demand compliance of the collection; the global array refers to the collection of all single-point arrays in the entire distributed monitoring system, which provides comprehensive monitoring of the operating status of the entire unit. The sensor management mode mainly refers to the management activities of sensors according to application requirements, including setting the sampling rate of sensors, data collection mode and other management work. Based on the global array and sensor management mode, the adaptive monitoring module is managed. The adaptive monitoring module is a module that automatically adjusts the monitoring strategy according to real-time data and unit status to achieve efficient and intelligent monitoring. Through the management of the adaptive monitoring module, it can be ensured that the entire monitoring system can be flexibly adjusted and optimized according to the actual situation of the unit to improve the accuracy and efficiency of monitoring.

[0028] Determine the three-dimensional structural characteristics and unit operation and control mechanism based on the specification configuration information, conduct twin non-equilibrium modeling, and determine the digital virtual machine group, wherein the refinement of the model partition is positively correlated with the monitoring requirements;

[0029] The three-dimensional structural characteristics of the unit are extracted according to the specification configuration information, including the geometric shape, size, relative position and connection method of each component of the unit. Through the three-dimensional modeling technology, a three-dimensional structural model of the unit is constructed, which can accurately reflect the actual physical structure of the unit; combined with the operation and control mechanism of the unit, the operation and control mechanism refers to the interaction and coordination between various systems and components during the operation of the unit. This includes the control strategy, sensor layout, signal transmission, etc. of the unit. By integrating the operation and control mechanism into the three-dimensional structural model, the actual operation process of the unit can be simulated, and data related to the unit status can be obtained. The twin non-equilibrium modeling method is used to determine the digital virtual machine group. Twin non-equilibrium modeling is a modeling method based on the mutual mapping and mutual influence of the actual physical system and the virtual system. By constructing a virtual machine group corresponding to the actual unit and establishing a data interaction and feedback mechanism between the two, real-time monitoring and prediction of the actual unit operation status can be achieved. In the modeling process, the refinement of the model partition is positively correlated with the monitoring needs. According to different monitoring needs, the unit is partitioned and modeled at different levels. For key components or areas that need to be focused on, more refined modeling is carried out to obtain more detailed status information. For other minor components or areas, relatively rough modeling can be performed to save computing resources and time. Through the above method, the modeling and monitoring requirements are highly consistent, model redundancy is avoided, and computing power is reasonably allocated. Based on the twin non-equilibrium modeling of the three-dimensional structural characteristics of the specification configuration information and the unit operation and control mechanism, a digital virtual machine group with high precision and high interactivity is obtained. This virtual machine group can not only be used for real-time monitoring and early warning, but also for fault simulation, performance optimization, etc., to provide strong support for the operation and maintenance management of the unit.

[0030] like Figure 2 As shown, based on the three-dimensional structural characteristics and the unit operation and control mechanism, the initial model configuration is determined;

[0031] Based on the monitoring requirements and the unit structural characteristics, the modeling refinement level is allocated and the partition modeling principle is determined. The unit structural characteristics at least include risk level, structural complexity and structural criticality;

[0032] Based on the partition modeling principle and the initial model configuration, twin modeling is performed to determine the digital virtual machine group.

[0033] Based on the three-dimensional structural characteristics and the unit operation and control mechanism, the initial model configuration is determined. The initial model configuration is the basic framework for building a digital virtual unit, which includes the basic structure, functional modules and operation and control logic of the unit. Through a detailed analysis of the three-dimensional structural characteristics, the geometric shape, size and position relationship of each component of the unit are clarified, and then the corresponding virtual model is constructed in the modeling software. Combined with the unit operation and control mechanism, the operation logic and control strategy of the virtual model are set to ensure that it can simulate the operating status of the actual unit. Combined with the monitoring requirements and the unit structural characteristics, the modeling refinement level is allocated. Monitoring requirements are the key factor guiding the allocation of modeling refinement levels. It determines which parts or systems require more refined modeling to obtain more accurate monitoring data. Unit structural characteristics, such as risk level, structural complexity and structural criticality, are also important bases for allocating modeling refinement levels. High risk level refers to the parts that often fail in the past, and higher modeling accuracy should be given to timely detect potential failures; parts with high structural complexity refer to the complex connections and architectures of the structure, which require more detailed modeling to reflect their actual operating conditions; and parts with high structural criticality are parts that are crucial to the overall performance and safety of the unit, so they also need to be finely modeled. The principle of partition modeling is to divide the unit into different modeling areas or partitions according to the structural characteristics and monitoring requirements of the unit. Each partition adopts different modeling accuracy and detail levels according to its importance, complexity and monitoring requirements. Based on the partition modeling principle and the initial model configuration, we can perform twin modeling to determine the digital virtual machine group. Twin modeling is a method of constructing a virtual machine group using actual unit data and operation information, which can achieve high consistency and real-time interactivity between the virtual machine group and the actual unit. Through twin modeling, the operating status, failure mode, performance change and other information of the actual unit are mapped to the virtual machine group in real time, so as to realize real-time monitoring, prediction and optimization of the actual unit. The initial model configuration is determined based on the three-dimensional structural characteristics and the unit operation and control mechanism. The modeling refinement level is allocated and the partition modeling principles are determined in combination with the monitoring requirements and the unit structural characteristics. Finally, a digital virtual unit is obtained through twin modeling, which can realize digital monitoring and management of the unit.

[0034] Retrieve multi-source unit information, perform information fusion analysis, and build a fault knowledge graph, which includes a fault association graph and a fault evolution graph.

[0035] Multi-source unit information is obtained through multiple channels such as design documents, operation logs, maintenance records, sensor data, expert opinions, and external databases. After obtaining multi-source unit information, information fusion analysis is required to extract valuable information for fault warning and diagnosis. The fault knowledge graph is a module used to represent knowledge in the field of unit faults. It includes two parts: fault association graph and fault evolution graph. According to the structure and operation principle of the unit, the main framework and key nodes of the knowledge graph are determined, and the correlation between different faults, such as causal relationship, co-occurrence relationship, etc., are analyzed. These relationships are represented in the knowledge graph to obtain the fault association graph. According to historical data and expert opinions, the evolution process of the fault from occurrence to development is described and displayed in the knowledge graph to obtain the fault evolution graph. The fault knowledge graph is constructed by combining the fault association graph and the fault evolution graph; the fault knowledge graph can be applied to multiple aspects such as fault warning, diagnosis and maintenance decision-making of the unit. By retrieving multi-source unit information, performing information fusion analysis and constructing a fault knowledge graph, a comprehensive and accurate unit fault warning and diagnosis system can be constructed, providing strong support for the safe operation and efficient maintenance of the unit.

[0036] Identify the multi-source unit information, filter the fault information and cluster the information based on the fault type to determine multiple information clusters;

[0037] Determine a first information cluster based on the multiple information clusters, perform feature fusion of homologous multi-dimensional information, and determine a first fused information cluster;

[0038] The acquisition of the Nth fusion information cluster is completed, intra-cluster correlation analysis and inter-cluster correlation analysis are performed, and the fault knowledge graph is constructed, wherein the correlation dimensions include synchronization and mutual influence.

[0039] Identifying the multi-source unit information refers to extracting information related to the unit's operating status and faults from various sources, such as sensor data, log files, maintenance records, expert opinions, etc. The identified information is screened, focusing on fault information. During the screening process, data that is irrelevant to the fault or has low information content is eliminated according to preset rules or algorithms, and information that is directly related to the fault and representative is retained. Information clustering is performed based on the fault type. Similar fault information is merged into the same information cluster according to the nature, location, impact range and other characteristics of the fault. A large amount of fault information can be divided into multiple information clusters with clear themes, which is convenient for subsequent analysis and processing. Determine a first information cluster based on the multiple information clusters, the first information cluster is one of the multiple information clusters selected by the staff, recorded as the first information cluster, and perform feature fusion of homologous multi-dimensional information. Homologous multi-dimensional information refers to information from the same data source but with different characteristics or dimensions. These information are integrated into a more representative first fusion information cluster. For example, the multiple information clusters include mechanical failure and electrical failure. From the multiple information clusters, any one of turbine failure, generator failure, bearing system failure, speed control system failure, etc. is extracted and recorded as the first information cluster. For example, turbine failure includes blade damage, deformation, and fatigue fracture. According to the turbine failure, the characteristics of the same multi-dimensional information are fused to determine the first fusion information cluster. By analogy, the acquisition of the Nth fusion information cluster is completed. Each fusion information cluster represents the multi-dimensional and comprehensive information of a certain type of fault, which provides a solid foundation for subsequent correlation analysis. Based on the above analysis results, a fault knowledge graph is constructed. This graph not only contains basic information such as the type, characteristics, and location of the fault, but also reveals the correlation and evolution trend between faults. Through the fault knowledge graph, the fault condition of the unit can be more intuitively understood, providing strong support for fault warning, diagnosis and maintenance.

[0040] Building a digital early warning model based on the digital virtual machine group and the fault knowledge graph;

[0041] Deeply integrate the digital virtual machine group with the fault knowledge graph. The digital virtual machine group provides a virtual experimental environment by simulating the actual operating status and performance of the unit. The fault knowledge graph contains rich fault information and associations, providing an in-depth understanding of the fault occurrence mechanism and evolution law. Extract the key operating parameters in the digital virtual machine group and the fault features in the fault knowledge graph, and use the deep learning algorithm to build a digital early warning model based on the extracted key parameters and fault features. Learn the complex relationship between the unit's operating status and faults from a large amount of data, and adjust the model parameters to optimize the early warning performance. Integrate the digital early warning model into the digital virtual machine group to achieve real-time monitoring and early warning of the unit's operating status. When the unit's operating status is abnormal or there is a potential risk of failure, the early warning model can issue an alarm in time and provide the corresponding fault type and possible cause, so that the operation and maintenance personnel can quickly take measures to intervene and repair.

[0042] Based on the adaptive monitoring module, sensor monitoring and data feedback of the unit division are performed, real-time fault detection and fault prediction of the predetermined time zone are performed in the digital early warning model, and a fault location list is generated;

[0043] Based on the adaptive monitoring module, the sensor monitoring and data feedback of the unit partition are carried out. If the collection and analysis are carried out continuously and uninterruptedly, the calculation amount is too large and there is invalid work. When the operating status changes, the monitoring is carried out again. The adaptive monitoring module will automatically adjust and optimize the layout and parameter settings of the sensor monitoring points according to the structural characteristics and operating requirements of the unit, including the reasonable configuration of the sensor type, position, sampling frequency, etc., to ensure that the operating data of each key part of the unit can be obtained comprehensively and accurately. On the basis of the unit partition, the sensor monitoring module will collect the operating data of each area in real time, and transmit these data to the digital early warning model through the data feedback system. In the digital early warning model, the data received in real time is processed and analyzed. Through the changing trends and characteristics of the real-time monitoring data, the model can timely detect the abnormal situation of the unit and perform fault detection. When a fault is detected or a potential fault is predicted, the digital early warning model will generate a fault location list. The fault location list will list the type, location, severity and possible causes of the fault in detail, providing clear fault location and guidance for operation and maintenance personnel. Through sensor monitoring and data feedback based on adaptive monitoring modules, as well as real-time fault detection and fault prediction in a predetermined time zone based on digital early warning models, real-time monitoring of the unit status and fault warning can be achieved, thus improving the safety and reliability of the unit, reducing operation and maintenance costs, and providing strong support for the intelligent management and optimized operation of the unit.

[0044] Based on the adaptive monitoring module, the same-frequency sensor monitoring data of the predetermined time zone is integrated and distributed point identification is performed, the partition modeling principle matching is traversed and data preprocessing is performed to determine the pre-transmitted data;

[0045] The pre-transmitted data is transmitted to the digital early warning model to perform fault decision analysis on the target unit.

[0046] The adaptive monitoring module will integrate the same-frequency sensor monitoring data according to the predetermined time zone. The same-frequency sensor monitoring data refers to the data collected by sensors at different locations at the same frequency or sampling rate. After these data are marked by distributed points, they can clearly reflect the operating status of each key part in the same time period. Traverse the partition modeling principle to match and preprocess the integrated data. The partition modeling principle is to divide the unit into different monitoring areas or partitions based on factors such as the structural characteristics, functional areas and historical fault data of the unit. Each partition has its specific monitoring focus and data processing method. Data preprocessing includes steps such as data cleaning, data transformation and feature extraction, aiming to improve the quality and consistency of the data and provide more valuable information for subsequent analysis. The pre-return data is determined. The pre-return data is a data set that is considered to be of great significance to the fault decision analysis after screening and preprocessing. The pre-return data will be transmitted to the digital early warning model for further fault decision analysis. The digital early warning model is able to use the pre-return data to identify the fault mode, evaluate the fault severity and predict the fault development trend. Through the digital early warning model, an accurate judgment on the fault status of the target unit can be obtained, providing decision support for the operation and maintenance team. Based on the data integration and preprocessing of the adaptive monitoring module, combined with the fault decision analysis of the digital early warning model, real-time monitoring and fault warning of the operating status of giant hydropower units can be achieved, improving the operating safety and reliability of the units.

[0047] Based on the pre-transmitted data, locally synchronize the digital virtual machine group to determine to update the virtual machine group;

[0048] Based on the updated virtual machine group, simulated operation is performed to determine simulated operation data for a predetermined time zone, wherein the time flow rate can be adjusted during the simulated operation stage;

[0049] Combined with the fault knowledge graph embedded in the fault decision layer, a fault decision is performed on the simulated operation data and the pre-transmitted data to generate the fault location single column.

[0050] The digital virtual machine group is locally synchronized using pre-returned data. The pre-returned data contains key information about the actual operating status of the unit. By analyzing these data, we can determine the parts of the digital virtual machine group that need to be updated. The process of local synchronization involves adjusting the relevant parameters and configurations of the virtual machine group to make it more accurately reflect the actual status of the unit. Through local synchronization, the updated virtual machine group is obtained, which provides a basis for subsequent simulated operation. Based on the updated virtual machine group, simulated operation is performed. Simulated operation simulates the operating status of the unit under specific conditions. By adjusting the time flow rate, the operation of the unit over a long span can be quickly simulated. In the simulated operation stage, simulated operation data of the predetermined time zone is obtained. These data reflect the operating status and performance of the unit under simulated conditions. Combined with the fault knowledge graph embedded in the fault decision layer, fault decisions are made on the simulated operation data and pre-returned data. The fault knowledge graph contains rich fault information and associations, which can identify abnormal patterns and potential faults in the data. By comparing and analyzing the simulated operation data with the pre-followed data, and combining it with the information in the fault knowledge graph, the failure risk of the unit can be accurately evaluated and predicted. Based on the results of the fault decision, a fault location column is generated. The fault location column lists in detail the type, location, severity, and possible causes of the fault, providing clear fault location and repair guidance for operation and maintenance personnel. Through the fault location column, operation and maintenance personnel can quickly locate the location of the fault and take appropriate measures to repair it, thereby ensuring the safe and stable operation of the unit. By performing local synchronization and simulated operation of the digital virtual machine unit based on the pre-followed data, and combining the fault knowledge graph in the fault decision layer to make a fault decision, accurate prediction and location of the unit fault can be achieved, providing strong support for the intelligent management and optimized operation of the unit.

[0051] For the first distributed point, determining a predicted characteristic value of the upper-level monitoring data, wherein the predicted characteristic value is any at least two items of the upper-level monitoring data;

[0052] Perform dynamic variable monitoring of the predicted characteristic value based on the intelligent sensor, and generate a sensor acquisition instruction if the variable threshold is met;

[0053] Based on the sensor collection instruction, the adaptive monitoring module controls the monitoring sensors of the first distributed points to collect and transmit data.

[0054] For the first distributed point, determine the prejudgment characteristic value of the upper monitoring data, the prejudgment characteristic value is any at least two data in the upper monitoring data, and determine the prejudgment characteristic value in the upper monitoring data of the first distributed point. These prejudgment characteristic values ​​are usually key data items in the upper monitoring data that can reflect the state or operation trend of the unit, which can be any at least two data. By selecting appropriate characteristic values, the operating state of the unit can be preliminarily judged and predicted. Use intelligent sensors to perform dynamic variable monitoring on the prejudgment characteristic values. Intelligent sensors have the ability of real-time monitoring and data processing, and can obtain dynamic change information of prejudgment characteristic values ​​in real time. When the change amount of these characteristic values ​​meets the preset variable threshold, it means that the operating state of the unit may have changed significantly or there is a potential risk of failure. At this time, the intelligent sensor will generate a sensor acquisition instruction. After receiving the sensor acquisition instruction, the adaptive monitoring module will control the monitoring sensor class of the first distributed point to collect data according to the instruction. The monitoring sensor class here may include multiple different types of sensors for obtaining various operating parameters and status information of the unit at this point. The adaptive monitoring module will select appropriate sensors for data acquisition according to actual needs and ensure the accuracy and integrity of the data. After completing data collection, the adaptive monitoring module will transmit the collected data back to the central processing system or the digital early warning model for further analysis and processing. Through the above method, accurate monitoring and fault early warning of the first distributed point can be achieved, thereby improving the safety and reliability of the unit.

[0055] A terminal visual warning is performed on the fault location column.

[0056] Terminal visual warning refers to the display of warning information on terminal devices in a graphical and intuitive way so that operation and maintenance personnel can quickly understand and respond. This warning method converts complex fault data and status information into easy-to-understand visual elements, such as color, shape, size, etc., to help operation and maintenance personnel quickly locate problems, determine the severity of faults, and take appropriate treatment measures. Integrate the data in the fault location column and convert it into a format suitable for visual display. According to actual needs, select a suitable visualization tool and design an intuitive and easy-to-understand interface in the visualization tool. This can include using colors, icons, labels, etc. to distinguish different fault types and severity. Implement terminal visual warning for fault location columns to improve operation and maintenance efficiency and fault handling speed.

[0057] like Figure 3 As shown, the embodiment of the present application includes a giant hydropower unit fault warning system, the system comprising:

[0058] A distributed monitoring demand determination module 11, which is used to read the specification configuration information of the target unit and determine the multi-level risk points, and determine the distributed monitoring demand, wherein the multi-level risk points are mined based on historical operation records;

[0059] An adaptive monitoring module 12, the adaptive monitoring module 12 is used to make a multi-sensor configuration decision and a sensor management decision based on the distributed monitoring requirements, and determine an adaptive monitoring module, the adaptive monitoring module is communicatively connected with the sensor array;

[0060] A digital virtual machine group determination module 13, the digital virtual machine group determination module 13 is used to determine the three-dimensional structural characteristics and unit operation and control mechanism based on the specification configuration information, perform twin non-equilibrium modeling, and determine the digital virtual machine group, wherein the refinement degree of the model partition is positively correlated with the monitoring demand;

[0061] A fault knowledge graph building module 14 is used to retrieve multi-source unit information, perform information fusion analysis, and build a fault knowledge graph. The fault knowledge graph includes a fault association graph and a fault evolution graph.

[0062] A digital early warning model building module 15, the digital early warning model building module 15 is used to build a digital early warning model based on the digital virtual machine group and the fault knowledge graph;

[0063] A fault location single column generation module 16, the fault location single column generation module 16 is used to perform sensor monitoring and data feedback of the unit partition based on the adaptive monitoring module, perform real-time fault detection and fault prediction of the predetermined time zone in the digital early warning model, and generate a fault location single column;

[0064] The terminal visualization warning module 17 is used to perform terminal visualization warning on the fault location column.

[0065] Furthermore, the embodiment of the present application also includes:

[0066] A single-point array configuration module, which is used to identify the monitoring requirements of each distributed point, determine the sensor type and field of view, and configure a single-point array, wherein the single-point array corresponds to the distributed point one by one;

[0067] A single point management mode determination module, the single point management mode determination module is used to traverse the single point array and determine the single point management mode;

[0068] A sensor management mode determination module, the sensor management mode determination module is used to coordinate the single-point array, determine the global array, integrate the single-point management mode and relatively independently configure it, and determine the sensor management mode;

[0069] An adaptive monitoring module is used to manage the adaptive monitoring module based on the global array and the sensor management mode.

[0070] Furthermore, the embodiment of the present application also includes:

[0071] An initial model configuration determination module, the initial model configuration determination module is used to determine the initial model configuration based on the three-dimensional structural characteristics and the unit operation and control mechanism;

[0072] A partition modeling principle determination module, which is used to allocate a modeling refinement level and determine a partition modeling principle in combination with monitoring requirements and unit structural characteristics, wherein the unit structural characteristics at least include risk level, structural complexity and structural criticality;

[0073] A digital virtual machine group determination module is used to perform twin modeling to determine the digital virtual machine group based on the partition modeling principle and the initial model configuration.

[0074] Furthermore, the embodiment of the present application also includes:

[0075] A multi-source unit information identification module, the multi-source unit information identification module is used to identify the multi-source unit information, screen fault information and perform information clustering based on fault types to determine multiple information clusters;

[0076] A feature fusion module, the feature fusion module is used to determine a first information cluster based on the multiple information clusters, perform feature fusion of homologous multi-dimensional information, and determine a first fused information cluster;

[0077] The association analysis module is used to complete the acquisition of the Nth fusion information cluster, perform intra-cluster association analysis and inter-cluster association analysis, and build the fault knowledge graph, wherein the association dimensions include synchronization and mutual influence.

[0078] Furthermore, the embodiment of the present application also includes:

[0079] A prejudgment characteristic value determination module, the prejudgment characteristic value determination module is used to determine the prejudgment characteristic value of the upper monitoring data for the first distributed point, the prejudgment characteristic value being any at least two items of the upper monitoring data;

[0080] A sensing acquisition instruction generation module, the sensing acquisition instruction generation module is used to perform dynamic variable monitoring of the predicted characteristic value based on the intelligent sensor, and generate a sensing acquisition instruction if the variable threshold is met;

[0081] A data collection and feedback module, wherein the data collection and feedback module is used to control the monitoring sensor type of the first distributed point to collect and feedback data based on the sensor collection instruction.

[0082] Furthermore, the embodiment of the present application also includes:

[0083] A pre-return data determination module, which is used to integrate the same-frequency sensor monitoring data of the predetermined time zone and perform distributed point identification based on the adaptive monitoring module, traverse the partition modeling principle matching and perform data preprocessing to determine the pre-return data;

[0084] A fault decision analysis module is used to transmit the pre-transmitted data to the digital early warning model to perform a fault decision analysis on the target unit.

[0085] Furthermore, the embodiment of the present application also includes:

[0086] A local synchronization module, the local synchronization module is used to perform local synchronization on the digital virtual machine group based on the pre-transmitted data, and determine to update the virtual machine group;

[0087] A simulation operation data determination module, the simulation operation data determination module is used to perform simulation operation based on the updated virtual machine group, and determine the simulation operation data of the predetermined time zone, wherein the time flow rate can be adjusted during the simulation operation stage;

[0088] A fault location single column generation module is used to combine the fault knowledge graph embedded in the fault decision layer, make a fault decision on the simulated operation data and the pre-transmission data, and generate the fault location single column.

[0089] For a specific embodiment of a giant hydropower unit fault warning system, please refer to the embodiment of a giant hydropower unit fault warning method described above, which will not be repeated here. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0090] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store news data and data such as time attenuation factors. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement a giant hydropower unit fault warning method.

[0091] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0092] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of a giant hydropower unit fault early warning method when executing the computer program.

[0093] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of a giant hydropower unit fault early warning method are implemented.

[0094] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A giant hydropower unit fault early warning method, characterized in that: The method comprises: Read the specification configuration information of the target unit and determine the multi-level risk points and distributed monitoring requirements, wherein the multi-level risk points are mined based on historical operation records; Based on the distributed monitoring requirements, a multi-sensor configuration decision and a sensor management decision are made to determine an adaptive monitoring module, wherein the adaptive monitoring module is communicatively connected with the sensor array; Determine the three-dimensional structural characteristics and unit operation and control mechanism based on the specification configuration information, perform twin-based non-equilibrium modeling, and determine the digital virtual machine unit; Retrieve multi-source unit information, perform information fusion analysis, and build a fault knowledge graph, which includes a fault association graph and a fault evolution graph. Building a digital early warning model based on the digital virtual machine group and the fault knowledge graph; Based on the adaptive monitoring module, sensor monitoring and data feedback of the unit division are performed, real-time fault detection and fault prediction of the predetermined time zone are performed in the digital early warning model, and a fault location list is generated; Perform terminal visual warning on the fault location column; Among them, multi-sensor configuration decision and sensor management decision are made, including: Identify the monitoring requirements of each distributed point, determine the sensor type and field of view, and configure a single-point array, wherein the single-point array corresponds to the distributed point one by one; Traversing the single point array to determine a single point management mode; Coordinate the single-point array, determine the global array, integrate the single-point management mode and configure it relatively independently, and determine the sensor management mode; Managing the adaptive monitoring module based on the global array and the sensor management mode; Among them, the digital virtual machine group is determined, including: Determining an initial model configuration based on the three-dimensional structural characteristics and the unit operation and control mechanism; Based on the monitoring requirements and the unit structural characteristics, the modeling refinement level is allocated and the partition modeling principle is determined. The unit structural characteristics at least include risk level, structural complexity and structural criticality; Based on the partition modeling principle and the initial model configuration, the digital virtual machine group is determined based on twin unbalanced modeling, wherein the refinement of the model partition is positively correlated with the monitoring demand; Conduct information fusion analysis, including: Identify the multi-source unit information, filter the fault information and cluster the information based on the fault type to determine multiple information clusters; Determine a first information cluster based on the multiple information clusters, perform feature fusion of homologous multi-dimensional information, and determine a first fused information cluster; Complete the acquisition of the Nth fusion information cluster, perform intra-cluster correlation analysis and inter-cluster correlation analysis, and build the fault knowledge graph, wherein the correlation dimensions include synchronization and mutual influence; The data monitoring and management of each distributed point is relatively independent, and the method further includes: Based on the adaptive monitoring module, the same-frequency sensor monitoring data of the predetermined time zone is integrated and distributed point identification is performed, the partition modeling principle matching is traversed and data preprocessing is performed to determine the pre-transmitted data; Transmitting the pre-transmitted data to the digital early warning model to perform fault decision analysis on the target unit; The digital early warning model includes an operation simulation layer and a fault decision layer, and performs fault decision analysis on the target unit, including: Based on the pre-transmitted data, locally synchronize the digital virtual machine group to determine to update the virtual machine group; Based on the updated virtual machine group, simulated operation is performed to determine simulated operation data for a predetermined time zone, wherein the time flow rate can be adjusted during the simulated operation stage; Combined with the fault knowledge graph embedded in the fault decision layer, a fault decision is performed on the simulated operation data and the pre-transmitted data to generate the fault location single column.

2. The method according to claim 1, characterized in that The multi-sensor includes intelligent sensors and monitoring sensors. The sensor monitoring and data feedback of the unit partition are performed based on the adaptive monitoring module. The method also includes: For the first distributed point, determining a predicted characteristic value of the upper-level monitoring data, wherein the predicted characteristic value is any at least two items of the upper-level monitoring data; Perform dynamic variable monitoring of the predicted characteristic value based on the intelligent sensor, and generate a sensor acquisition instruction if the variable threshold is met; Based on the sensor collection instruction, the adaptive monitoring module controls the monitoring sensors of the first distributed points to collect and transmit data.

3. A giant hydropower unit fault warning system, characterized in that: The system comprises: A distributed monitoring demand determination module, which is used to read the specification configuration information of the target unit and determine the multi-level risk points to determine the distributed monitoring demand, wherein the multi-level risk points are mined based on historical operation records; An adaptive monitoring module, the adaptive monitoring module is used to make multi-sensor configuration decisions and sensor management decisions based on the distributed monitoring requirements, determine an adaptive monitoring module, and the adaptive monitoring module is communicatively connected with the sensor array; A digital virtual machine group determination module, the digital virtual machine group determination module is used to determine the three-dimensional structural characteristics and unit operation and control mechanism based on the specification configuration information, perform twin non-equilibrium modeling, and determine the digital virtual machine group; A fault knowledge graph building module is used to retrieve multi-source unit information, perform information fusion analysis, and build a fault knowledge graph. The fault knowledge graph includes a fault association graph and a fault evolution graph. A digital early warning model building module, the digital early warning model building module is used to build a digital early warning model based on the digital virtual machine group and the fault knowledge graph; A fault location single column generation module, the fault location single column generation module is used to perform sensor monitoring and data feedback of the unit partition based on the adaptive monitoring module, perform real-time fault detection and fault prediction of the predetermined time zone in the digital early warning model, and generate a fault location single column; A terminal visualization warning module, which is used to perform terminal visualization warning on the fault location column; The system further comprises: A single-point array configuration module, which is used to identify the monitoring requirements of each distributed point, determine the sensor type and field of view, and configure a single-point array, wherein the single-point array corresponds to the distributed point one by one; A single point management mode determination module, the single point management mode determination module is used to traverse the single point array and determine the single point management mode; A sensor management mode determination module, the sensor management mode determination module is used to coordinate the single-point array, determine the global array, integrate the single-point management mode and relatively independently configure it, and determine the sensor management mode; An adaptive monitoring module, the adaptive monitoring module being used to manage the adaptive monitoring module based on the global array and the sensor management mode; An initial model configuration determination module, the initial model configuration determination module is used to determine the initial model configuration based on the three-dimensional structural characteristics and the unit operation and control mechanism; A partition modeling principle determination module, which is used to allocate a modeling refinement level and determine a partition modeling principle in combination with monitoring requirements and unit structural characteristics, wherein the unit structural characteristics at least include risk level, structural complexity and structural criticality; A digital virtual machine group determination module, the digital virtual machine group determination module is used to determine the digital virtual machine group based on twin unbalanced modeling based on the partition modeling principle and the initial model configuration, wherein the refinement degree of the model partition is positively correlated with the monitoring demand; A multi-source unit information identification module, the multi-source unit information identification module is used to identify the multi-source unit information, screen fault information and perform information clustering based on fault types to determine multiple information clusters; A feature fusion module, the feature fusion module is used to determine a first information cluster based on the multiple information clusters, perform feature fusion of homologous multi-dimensional information, and determine a first fused information cluster; A correlation analysis module, which is used to complete the acquisition of the Nth fusion information cluster, perform intra-cluster correlation analysis and inter-cluster correlation analysis, and build the fault knowledge graph, wherein the correlation dimensions include synchronization and mutual influence; A pre-return data determination module, which is used to integrate the same-frequency sensor monitoring data of the predetermined time zone and perform distributed point identification based on the adaptive monitoring module, traverse the partition modeling principle matching and perform data preprocessing to determine the pre-return data; A fault decision analysis module, the fault decision analysis module is used to transmit the pre-transmitted data to the digital early warning model to perform a fault decision analysis on the target unit, the digital early warning model includes an operation simulation layer and a fault decision layer; A local synchronization module, the local synchronization module is used to perform local synchronization on the digital virtual machine group based on the pre-transmitted data, and determine to update the virtual machine group; A simulation operation data determination module, the simulation operation data determination module is used to perform simulation operation based on the updated virtual machine group, and determine the simulation operation data of the predetermined time zone, wherein the time flow rate can be adjusted during the simulation operation stage; A fault location single column generation module is used to combine the fault knowledge graph embedded in the fault decision layer, make a fault decision on the simulated operation data and the pre-transmission data, and generate the fault location single column.

4. 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 method according to any one of claims 1 to 2 are implemented.

5. 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 method according to any one of claims 1 to 2 are implemented.

Images