A fault handling system, method, and medium for an energy unit powered system

Abstract

The application discloses a kind of fault processing systems, methods and media for energy unit energy supply system, it is related to energy unit management technical field, solve the existing energy unit management, cannot carry out dynamic data analysis, leading to the technical problem of low comprehensive efficiency when system faces multiple energy unit management;By obtaining the operation data of each energy subsystem;According to the project operation data, the parameter change curve corresponding to the running project is generated;According to each parameter change curve, the fault analysis data corresponding to the energy subsystem is generated;Obtain predicted environmental data, input predicted environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystem, obtain the prediction analysis result corresponding to the energy subsystem;Generate prediction processing scheme based on prediction analysis result;Through the targeted generation of fault analysis data, realize the dynamic sampling of data when fault analysis;Improve the comprehensive efficiency of analyzing multiple energy units.

Description

Technical Field

[0001] This application belongs to the field of energy unit management technology, specifically a fault handling system, method and medium for energy unit power supply systems. Background Technology

[0002] An energy unit generally refers to equipment, devices, or systems capable of independently producing, converting, storing, or utilizing energy. The energy supply system of an energy unit is the key component that converts energy into a usable form and provides energy to various energy-consuming equipment or systems. Common types include electric power supply systems, heat supply systems, and gas supply systems. The efficient operation of an energy unit's power supply system is of great significance for ensuring energy supply and meeting user needs.

[0003] The general process of existing energy unit management technology is as follows: collect a fixed number of parameters corresponding to the relevant detection items of the energy unit, analyze these parameters, and then determine the operating status of the energy unit. However, when managing multiple energy units in an entire area, if the number of sampling parameters set in the early stage is large, the accuracy of the analysis can be guaranteed, but data congestion will occur when managing multiple energy units. If the number of sampling parameters set in the early stage is small, the accuracy of the analysis cannot be guaranteed. Therefore, a fault handling system for energy unit power supply systems is needed. Summary of the Invention

[0004] This application provides a fault handling system, method, and medium for energy unit power supply systems, which solves the technical problem that existing energy unit management cannot perform dynamic data analysis, resulting in low overall efficiency when facing the management of multiple energy units.

[0005] To achieve the above objectives, this application adopts the following technical solution: Firstly, a fault handling method for an energy unit power supply system is provided, including: Acquire operational data for each energy subsystem; the operational data includes project operational data corresponding to several operational items. Generate parameter change curves corresponding to the running project based on the project operation data; Based on the change curves of each parameter, fault analysis data corresponding to the energy subsystem is generated. The fault analysis data includes several parameter values ​​corresponding to each operating item. Acquire predictive environmental data, input the predictive environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystem, and obtain the predictive analysis results corresponding to the energy subsystem; the fault prediction model is trained based on the digital twin model corresponding to the energy subsystem. A predictive processing scheme is generated based on the predictive analysis results.

[0006] Based on the above technical solution, in the fault handling method for an energy unit power supply system provided in this application, the following steps are taken: First, operational data of each energy subsystem is acquired. Then, parameter variation curves corresponding to the operational projects are generated based on the operational data. Next, fault analysis data corresponding to each energy subsystem is generated based on the parameter variation curves. Predictive environmental data is acquired, and the predicted environmental data and fault analysis data are input into the fault prediction model corresponding to the energy subsystem to obtain the predictive analysis results corresponding to the energy subsystem. A predictive processing scheme is generated based on the predictive analysis results. By specifically generating fault analysis data, dynamic sampling of data during fault analysis is achieved, ensuring both the accuracy of the analysis results and the ability to adjust the amount of data analyzed according to the actual situation during fault analysis, thereby improving the overall efficiency of analyzing multiple energy units.

[0007] In conjunction with the first aspect above, in one possible implementation, generating the parameter change curve corresponding to the running project based on project running data includes: Extract several parameter values ​​from each collection period in the project's operational data; The set number of fitting periods is obtained. The specific value of the number of fitting periods is set by experts. It can be understood that the number of fitting periods is a positive integer. In this embodiment, the number of fitting periods is set to 10. The sampling periods are grouped according to the number of fitting periods to obtain several partial fitting data groups. The sampling periods within the partial fitting data groups are continuous, and there are overlapping sampling periods between different partial fitting data groups. The parameter values ​​in each sampling period within each partial fitting data group are fitted into partial parameter curves according to their corresponding acquisition order. The partial parameter curves are fitted into the parameter change curves of the running project.

[0008] In conjunction with the first aspect above, in one possible implementation, fault analysis data corresponding to the energy subsystem is generated based on the variation curves of various parameters, including: S1: Obtain the parameter change curves corresponding to several operating items of the energy subsystem, as well as the initially set number of sampling periods and period sampling interval; S2: Monitoring and analysis data are obtained by sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval; S3: Generate data characterization based on monitoring and analysis data and parameter change curves; extract the project characterization corresponding to each running project in the data characterization, and determine whether each project characterization is greater than or equal to the set characterization threshold. If yes, proceed to S5; otherwise, proceed to S4. S4: Adjust the initially set number of sampling periods and periodic sampling interval based on the data representation degree, and then proceed to S1; S5: Use the monitoring and analysis data as fault analysis data.

[0009] In conjunction with the first aspect above, in one possible implementation, the step of sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval to obtain monitoring and analysis data includes: The sampling period number and the period sampling interval are obtained. The sampling period number is the number of data analysis periods required, and the sampling interval is the time interval corresponding to the data sampling point. The sampling period is sampled using the period sampling interval to obtain several sampling points. Continuous sampling periods are selected as sampling segments based on the sampling period number. The curve segment corresponding to the sampling segment in the parameter change curve is extracted. Several sampling points are set based on the period sampling interval. Several parameter values ​​are obtained by sampling on the curve segment based on the sampling points. The various parameter values ​​are integrated into the monitoring and analysis data of the corresponding running project in chronological order. The monitoring and analysis data corresponding to each running project are acquired sequentially; the monitoring and analysis data corresponding to each running project are then integrated into the final monitoring and analysis data.

[0010] In conjunction with the first aspect above, in one possible implementation, the generation of data characterization based on monitoring and analysis data and parameter change curves includes: Extract several parameter values ​​corresponding to each operational item from the monitoring and analysis data, fit the parameter values ​​into a parameter value change curve, and obtain the corresponding curve segment from the parameter change curve; Input the parameter value change curve and curve segment into the set characterization generation function to obtain the corresponding project characterization; obtain the project characterization corresponding to each running project, and integrate the project characterization into data characterization; One form of the characterization generation function includes: ; in, As the project's representativeness; Number The corresponding reference weights; For curved segments; The parameter value variation curve; The designated verification point number is... , The set number of characterization verification points; Let x be the time of the characterization verification point, and , The starting point time corresponding to the parameter value change curve. The end point time corresponding to the parameter value change curve; The defined interval for characterization considerations.

[0011] In conjunction with the first aspect above, in one possible implementation, adjusting the initially set number of sampling periods and the periodic sampling interval based on data representation includes: Extract the project characterization of several running items from the data characterization; The project characterization of each running project is input into a set sampling period adjustment function to obtain the adjusted sampling period number; one way to express the sampling period adjustment function includes: ; ; in, This is the adjusted number of sampling periods. The number of sampling periods before adjustment; This is the rounding function; The maximum number of adjustable cycles is set to 100 in this embodiment; For the number The project characterization corresponding to the running project. The total number of running projects; The characterization threshold; The project characterization of each running project is input into a set sampling interval adjustment function to obtain the adjusted sampling interval; one way to express the sampling interval adjustment function includes: ; in, The adjusted sampling interval, The sampling interval before adjustment; For the number The project characterization degree corresponding to the running project; This is the characterization threshold; the specific value of the characterization threshold is set by experts based on experience.

[0012] In conjunction with the first aspect above, one possible implementation of the fault prediction model includes the following methods for constructing it: The process involves acquiring equipment parameters of an energy subsystem and constructing a digital twin model of the energy subsystem based on these parameters. These equipment parameters include line-related parameters, environment-related parameters, and equipment-related parameters. Equipment models within the energy subsystem are constructed based on the equipment-related parameters, and connection relationships between these equipment models are established based on the line-related parameters. Finally, a complete digital twin model of the energy subsystem is constructed based on the usage scenario. The input data for the digital twin model consists of environmental data and fault analysis data for the corresponding energy subsystem, while the output data consists of monitoring data. A fault identification model is constructed based on a neural network model. The input data of the fault identification model is monitoring data, and the output data is the state of the energy subsystem. The output data from the digital twin model is input into the input of the fault identification model; the energy subsystem state output by the fault identification model is used as the predictive analysis result to construct a fault prediction model.

[0013] In conjunction with the first aspect above, in one possible implementation, the step of generating a predictive processing scheme based on the predictive analysis results includes: Extract the energy subsystem status from the predictive analysis results. When the energy subsystem status is abnormal, retrieve the corresponding processing solution from the processing solution query table based on the abnormal status, and use the processing solution as the predictive processing solution. The processing solution query table includes several query items and several processing solutions corresponding to the query items. The query items are several types of abnormal statuses.

[0014] Secondly, this application provides a fault handling system for an energy unit power supply system, comprising: a data acquisition module, a data analysis module, and a database; wherein, The data acquisition module includes an operational data acquisition subunit and an environmental data acquisition subunit; The operational data acquisition subunit is used to collect operational data from each energy subsystem. The environmental data acquisition subunit is used to collect predicted environmental data from various energy subsystems; the predicted environmental data can be obtained from weather forecasts. The data analysis module includes a front-end subunit, a model invocation subunit, and a back-end subunit; The front-end subunit is used to generate parameter change curves corresponding to the operational project based on the project operation data; and to generate fault analysis data corresponding to the energy subsystem based on each parameter change curve, wherein the fault analysis data includes several parameter values ​​corresponding to each operational project. The model invocation subunit is used to invoke the corresponding fault prediction model according to the energy subsystem corresponding to the fault analysis data. The predicted environment data and fault analysis data are input into the fault prediction model corresponding to the energy subsystem to obtain the prediction analysis result corresponding to the energy subsystem. The fault prediction model is trained based on the digital twin model corresponding to the energy subsystem. The post-subunit is used to generate a predictive processing scheme based on the predictive analysis results; The database is used to store various data required and generated by this system.

[0015] Thirdly, this application provides a computer-readable storage medium storing instructions that, when executed on a fault handling system, cause the fault handling system to perform the methods described in the first aspect and any possible implementation thereof.

[0016] This application provides a fault handling method for an energy unit power supply system. The method involves: acquiring operational data from various energy subsystems; generating parameter variation curves corresponding to the operational projects based on the operational data; generating fault analysis data for each energy subsystem based on the parameter variation curves; acquiring predictive environmental data; inputting the predictive environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystem to obtain predictive analysis results for the energy subsystem; generating a predictive processing scheme based on the predictive analysis results; and achieving dynamic sampling of data during fault analysis by specifically generating fault analysis data, thus ensuring the accuracy of the analysis results while adjusting the amount of data analyzed according to actual conditions, thereby improving the overall efficiency of analyzing multiple energy units.

[0017] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram illustrating the steps of the fault handling method for the energy unit power supply system in this application. Figure 2 This is a schematic diagram of the module connection of the fault handling system of the energy unit power supply system in this application. Detailed Implementation

[0020] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Please see Figure 1 The first aspect of this application provides a fault handling method for an energy unit power supply system, comprising: Acquire operational data for each energy subsystem; the operational data includes operational data for several operational items; the energy subsystems include various energy subsystems such as cooling, heating, and electricity, such as energy systems that control heating and power supply; operational items are the items monitored during the operation of various equipment in the energy subsystems; such as motor speed, active power, reactive power, real-time load rate, rated load rate, maximum load rate, unit start-up and shutdown status, AGC / AVC regulation response speed, main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, condenser vacuum, feedwater temperature, feedwater pressure, etc. The parameter change curves corresponding to the running project are generated based on the project operation data. The parameter change curves are continuous parameter change curves fitted based on the discrete parameter values ​​collected from the running project. Curve fitting methods such as interpolation can be used for curve fitting. Fault analysis data corresponding to the energy subsystem is generated based on the parameter change curves. The fault analysis data includes several parameter values ​​corresponding to each operating item. The fault analysis data is a number of parameter values ​​extracted from the parameter change curves that can represent the operating status of the equipment represented by the parameter change curves. Acquire predictive environmental data, which is the environmental conditions of the area corresponding to the energy subsystem obtained through weather forecasts and other means, including environmental temperature and humidity; input the predictive environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystem to obtain the predictive analysis results corresponding to the energy subsystem; the fault prediction model is trained based on the digital twin model corresponding to the energy subsystem. A predictive processing scheme is generated based on the predictive analysis results.

[0022] Based on the above technical solution, in the fault handling method for an energy unit power supply system provided in this application, the following steps are taken: First, operational data of each energy subsystem is acquired. Then, parameter variation curves corresponding to the operational projects are generated based on the operational data. Next, fault analysis data corresponding to each energy subsystem is generated based on the parameter variation curves. Then, predictive environmental data is acquired, and the predictive environmental data and fault analysis data are input into the fault prediction model corresponding to the energy subsystem to obtain the predictive analysis results corresponding to the energy subsystem. Finally, a predictive handling scheme is generated based on the predictive analysis results. By specifically generating fault analysis data, indirect dynamic sampling of data during fault analysis is achieved, ensuring that during fault analysis, both the accuracy and timeliness of the data are maintained. This approach ensures the accuracy of the analysis results while allowing for adjustments to the amount of data analyzed based on actual conditions, thus improving the overall efficiency of analyzing multiple energy units. Simultaneously, indirect dynamic sampling, or two-stage sampling, involves sampling from the physical entities of the energy units using a fixed sampling method, processing the sampling results, and then performing a second dynamic sampling on the processed results. It's understood that the sampling interval during the first sampling should be as small as possible to ensure the accuracy of the subsequent parameter curves, facilitating the verification of data analysis results. The second sampling enables dynamic sampling, ensuring both the accuracy of the analysis results and the ability to adjust the amount of data analyzed based on actual conditions during fault analysis, thereby improving the overall efficiency of analyzing multiple energy units.

[0023] In one possible implementation, generating a parameter change curve corresponding to the running project based on the project operation data includes: extracting several parameter values ​​within each collection period of the project operation data; The set number of fitting periods is obtained, the specific value of which is set by experts. It is understood that the number of fitting periods is a positive integer; in this embodiment, the number of fitting periods is set to 10. The sampling periods are grouped according to the number of fitting periods to obtain several partial fitting data groups. The sampling periods within each partial fitting data group are continuous, and there are overlapping sampling periods between different partial fitting data groups. The parameter values ​​within each sampling period of each partial fitting data group are fitted into partial parameter curves according to their corresponding acquisition order. The partial parameter curves are fitted into the parameter change curves of the running project. Specifically, each partial parameter change curve is numbered sequentially according to time order, which can be based on the order of acquisition time corresponding to the starting point of the partial parameter change curve. Two partial parameter curves with overlapping sampling periods are marked as follows: and ; through the following formula: ; in, The curve is the result of merging the partial parameter curve numbered i and the partial parameter curve numbered i+1. These are the weighting coefficients corresponding to the partial parameter curve numbered i; These are the weighting coefficients corresponding to the partial parameter curve numbered i+1; The point in time at which the overlapping portion of the partial parameter curve numbered i and the partial parameter curve numbered i+1 begins; This refers to the end point of the overlapping portion of the partial parameter curve numbered i and the partial parameter curve numbered i+1; the weighting coefficient can be set based on expert experience, and can be... , The weighting coefficients can also be calculated using the following functions; one way to express the weighting coefficients includes: The sampled curves are merged sequentially according to their numbering order to obtain the final parameter numbering curve, which includes: S1: Obtain the number corresponding to each part of the sampling curve; S2: Merge the partial sampled curves numbered 1 and 2 to obtain a new merged partial sampled curve numbered 1; the weighting coefficients during fusion are determined based on the lengths of the two partial sampled curves. Specifically, ; ; in, The time when the partial sampling curve corresponding to number i ends. This represents the start time of the partial sampling curve corresponding to number i; the longer the length of the partial sampling curve, the larger its corresponding weighting coefficient should be set when performing curve fusion. S3: Renumber the sampled curves of each part; proceed to S2 until all the partial curve segments are merged into one curve; To make it easier to understand, the fusion process can be as follows: after the partial parameter curves numbered 1 and 2 are fused, they are then fused with the partial parameter curve numbered 3, and so on, until the last partial parameter curve is reached.

[0024] The original monitoring data is obtained by collecting data corresponding to each operating item of the corresponding energy subsystem at fixed time intervals. Since it is not possible to collect data without any intervals, it is difficult to achieve completely continuous data collection. If it is necessary to obtain data at non-sampling points later, the data at non-sampling points can only be inferred by fitting the collected data. In order to ensure the accuracy of the inferred data, this embodiment uses the above-mentioned segmented fitting method and then fusion method. Different data are mutually corrected to further increase the fit between the fitted curve and the continuous changes of the actual operating parameters.

[0025] It is understood that, in another embodiment, after obtaining the parameter change curve, the parameter change curve for a certain period of time in the future can be predicted based on the trend of the parameter change curve to obtain the final parameter change curve.

[0026] In one possible implementation, fault analysis data for the energy subsystem is generated based on the variation curves of various parameters, including: S1: Obtain the parameter change curves corresponding to several operating items of the energy subsystem, as well as the initially set number of sampling periods and period sampling interval; S2: Monitoring and analysis data are obtained by sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval; S3: Generate data characterization based on monitoring and analysis data and parameter change curves; extract the project characterization corresponding to each running project in the data characterization, and determine whether each project characterization is greater than or equal to the set characterization threshold. If yes, proceed to S5; otherwise, proceed to S4. S4: Adjust the initially set number of sampling periods and periodic sampling interval based on the data representation degree, and then proceed to S1; S5: Use the monitoring and analysis data as fault analysis data.

[0027] In one possible implementation, monitoring and analysis data is obtained by sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval. This includes: obtaining the number of sampling periods and the periodic sampling interval, where the number of sampling periods is the required number of data analysis periods, and the sampling interval is the time interval corresponding to the data sampling points; sampling the sampling periods using the periodic sampling interval to obtain several sampling points; selecting consecutive sampling periods as sampling segments based on the number of sampling periods. It is worth noting that in this embodiment, the selected consecutive sampling periods are the most recent sampling period from the current time as the starting point, and the number of consecutive sampling periods selected in the direction preceding the current time are the corresponding number of sampling periods as the sampling segments; extracting the curve segment corresponding to the sampling segment in the parameter change curve; setting several sampling points based on the periodic sampling interval; specifically, taking the starting time point of each sampling period corresponding to the sampling segment as the first sampling point, and setting one sampling point every one periodic sampling interval; sequentially setting sampling points within each sampling period; sampling on the curve segment based on the sampling points to obtain several parameter values; and integrating the parameter values ​​in chronological order to form the monitoring and analysis data of the corresponding running project. The monitoring and analysis data corresponding to each running project are acquired sequentially; the monitoring and analysis data corresponding to each running project are then integrated into the final monitoring and analysis data.

[0028] In one possible implementation, data characterization is generated based on monitoring and analysis data and parameter change curves, including: extracting several parameter values ​​corresponding to each operational item in the monitoring and analysis data, fitting the parameter values ​​into a parameter value change curve; and obtaining the corresponding curve segment in the parameter change curve. Input the parameter value change curve and curve segment into the set characterization generation function to obtain the corresponding project characterization; obtain the project characterization corresponding to each running project, and integrate the project characterization into data characterization; One form of the characterization generation function includes: ; in, As the project's representativeness; Number The corresponding reference weights; For curved segments; The parameter value variation curve; The designated verification point number is... , The set number of characterization verification points; Let x be the time of the characterization verification point, and , The starting point time corresponding to the parameter value change curve. The end point time corresponding to the parameter value change curve; The specific value for the characterization interval is set by experts based on experience. For example, if the time length corresponding to the parameter value change curve is 1 hour, then the characterization interval can be set to a duration of less than half an hour. The range of values ​​is meaningful, and the characterization consideration interval in this embodiment is significant. Set to 1 minute; understandably, the larger the value of the characterization verification point count, the more characterization verification points there are, and the higher the accuracy of the calculated project characterization. The characterization verification points are obtained by setting them at even intervals within the range of characterization verification points. Each point serves as a characterization verification point; that is... One method for obtaining the reference weights includes: ; in, Number The average curvature of the corresponding curve segment can be obtained using the following formula: ; in, For curve segments The corresponding first derivative; For curve segments The corresponding second derivative; In this embodiment, the number is calculated using the above formula. The average curvature of a corresponding curve segment can represent the degree of change of the curve within that segment. The greater the degree of change within a curve segment, the more details it contains, and the higher its importance. If the curve segment cannot be accurately simulated when the digital twin model simulates its operating parameters, the simulation process will suffer from poor detail control, leading to deviations in the final simulation results. Therefore, the reference weight for the representation of this curve segment is set to be relatively large.

[0029] In this embodiment, the above-described method calculates several parameter values ​​to represent the characterization degree of the curve segment. The smaller the deviation between the curve fitted using these parameter values ​​and the corresponding curve segment, the stronger the ability of these parameter values ​​to represent the curve segment. Subsequently, when simulating the operation of the energy subsystem using these parameter values, the operating state of each component in the energy subsystem is highly consistent with the actual operating state. At this time, the corresponding project characterization degree is set to be larger. At the same time, this embodiment also further increases the accuracy of the project characterization degree by considering the deviation of different parts of the curve segment in segments and performing weighted summation on the deviation.

[0030] In one possible implementation, the initially set number of sampling periods and the periodic sampling interval are adjusted based on the data representation degree, including: extracting the project representation degree of several running items from the data representation degree; The project characterization of each running project is input into a set sampling period adjustment function to obtain the adjusted sampling period number; one way to express the sampling period adjustment function includes: ; ; in, This is the adjusted number of sampling periods. The number of sampling periods before adjustment; This is the rounding function; The maximum number of adjustable cycles is set to 100 in this embodiment; For the number The project characterization corresponding to the running project. The total number of running projects; The characteristic threshold is used to adjust the number of sampling periods based on the characteristic of the project, thereby expanding the simulated data. When the characteristic of the project is lower, it indicates that the current monitoring and analysis data is difficult to represent the actual operating state of the equipment. When the monitoring and analysis data is used for simulation, the simulation results deviate significantly from the actual operating results. Therefore, by increasing the amount of data, it is ensured that the input data can make the simulated operating state highly consistent with the actual operating state of the equipment. Therefore, when the characteristic of the monitoring and analysis data is low, this embodiment expands the data simultaneously through two paths: increasing the number of sampling periods and reducing the sampling interval. The project characterization of each running project is input into a set sampling interval adjustment function to obtain the adjusted sampling interval; one way to express the sampling interval adjustment function includes: ; in, The adjusted sampling interval, The sampling interval before adjustment; For the number The project characterization degree corresponding to the running project; This is the characterization threshold; the specific value of the characterization threshold is set by experts based on experience. This embodiment adjusts the sampling interval using the above formula. When the deviation between the project characterization degree and the characterization degree threshold corresponding to the running project is larger, it indicates that the currently collected monitoring and analysis data has a weaker ability to represent the curve fitting the actual running data. That is, when using the monitoring and analysis data for simulation, the simulation result deviates significantly from reality. At this time, it is necessary to increase the amount of sampled data, i.e., reduce the sampling interval to ensure sufficient data. The more data points sampled, the higher the characterization degree of the data used for simulation, resulting in stronger control over the details in the subsequent simulation process.

[0031] In one possible implementation, the fault prediction model is constructed in the following ways: acquiring equipment parameters of the energy subsystem, and constructing a digital twin model of the energy subsystem based on the equipment parameters; the equipment parameters include line-related parameters, environment-related parameters, and equipment-related parameters of the energy subsystem; constructing equipment models in the energy subsystem based on the equipment-related parameters, and constructing connection relationships between the various equipment models based on the line-related parameters; and then constructing a complete digital twin model of the energy subsystem based on the usage scenario; the input data of the digital twin model is the environmental data and fault analysis data of the corresponding energy subsystem, and the output data is the monitoring data. A fault identification model is constructed based on a neural network model. The input data of the fault identification model is monitoring data, and the output data is the state of the energy subsystem. The output data from the digital twin model is input into the input of the fault identification model; the energy subsystem state output by the fault identification model is used as the predictive analysis result to construct a fault prediction model.

[0032] It is understood that the digital twin model in this embodiment is used to simulate the operation of the energy subsystem under the input operation-related data and environmental-related data. Then, by acquiring the digital twin model to simulate the operation, the monitoring data of each component in the energy subsystem is collected and output. The monitoring data includes relevant data such as the temperature, current, voltage and vibration of the equipment. The output monitoring data is input into the fault identification model to obtain the prediction results; One training method for the fault identification model includes acquiring several actual monitoring data collected from the energy subsystem, and the corresponding energy subsystem status; the energy subsystem status is the operating status of the energy subsystem obtained by experts through analysis of the monitoring data, including normal status and several types of abnormal status, such as abnormal voltage, abnormal vibration, and abnormal temperature of a certain device in the energy subsystem; the monitoring data and its corresponding energy subsystem status are integrated into several training data and verification data. The artificial intelligence model is trained using training data and tested using validation data, ultimately resulting in a fault prediction model that takes monitoring data as input and outputs the state of the energy subsystem; the artificial intelligence model includes a neural network model.

[0033] If the parameters corresponding to each operating item remain almost unchanged, inputting a small amount of monitoring data into the fault prediction model is sufficient to grasp the operating characteristics of the energy subsystem and obtain accurate simulation results. However, if the parameters of each operating item change drastically, a larger amount of monitoring data needs to be input into the fault prediction model to grasp the operating characteristics of the energy subsystem and obtain accurate simulation results. Conventional fault prediction often inputs a fixed amount of data into the fault prediction model for prediction, which leads to wasted data processing or insufficient prediction accuracy. Therefore, this embodiment sets the amount of data input into the fault prediction model according to the current parameter changes of the energy subsystem, which can adaptively grasp the dynamic changes in the operating status of the energy subsystem and achieve adaptive data input. This ensures that there is no waste of computation while also guaranteeing the accuracy of the prediction results.

[0034] In one possible implementation, generating a predictive processing scheme based on the predictive analysis results includes: extracting the energy subsystem state from the predictive analysis results; when the energy subsystem state is an abnormal state, querying a processing scheme lookup table based on the abnormal state to obtain the corresponding processing scheme, and using the processing scheme as the predictive processing scheme; the processing scheme lookup table includes several query items and several processing schemes corresponding to the query items, and the query items are several types of abnormal states.

[0035] Please see Figure 2 Secondly, this application provides a fault handling system for an energy unit power supply system, comprising: a data acquisition module, a data analysis module, and a database; wherein, The data acquisition module includes an operational data acquisition subunit and an environmental data acquisition subunit; The operational data acquisition subunit is used to collect operational data from each energy subsystem. The environmental data acquisition subunit is used to collect predicted environmental data from various energy subsystems; the predicted environmental data can be obtained from weather forecasts. The data analysis module includes a front-end subunit, a model invocation subunit, and a back-end subunit; The front-end subunit is used to generate parameter change curves corresponding to the operational project based on the project operation data; and to generate fault analysis data corresponding to the energy subsystem based on each parameter change curve, wherein the fault analysis data includes several parameter values ​​corresponding to each operational project. The model invocation subunit is used to invoke the corresponding fault prediction model according to the energy subsystem corresponding to the fault analysis data. The predicted environment data and fault analysis data are input into the fault prediction model corresponding to the energy subsystem to obtain the prediction analysis result corresponding to the energy subsystem. The fault prediction model is trained based on the digital twin model corresponding to the energy subsystem. The post-subunit is used to generate a predictive processing scheme based on the predictive analysis results; The database is used to store various data required and generated by this system.

[0036] Thirdly, this application provides a computer-readable storage medium storing instructions that, when executed on a fault handling system, cause the fault handling system to perform the methods described in the first aspect and any possible implementation thereof.

[0037] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0038] How this application works: By acquiring operational data from each energy subsystem; generating parameter variation curves corresponding to the operational projects based on the project operational data; generating fault analysis data corresponding to the energy subsystems based on the parameter variation curves; acquiring predictive environmental data; inputting the predictive environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystems to obtain predictive analysis results for the energy subsystems; generating predictive processing schemes based on the predictive analysis results; and achieving dynamic sampling of data during fault analysis by selectively generating fault analysis data, thereby ensuring the accuracy of the analysis results while adjusting the amount of data analyzed according to the actual situation, and improving the overall efficiency of analyzing multiple energy units.

[0039] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A fault handling method for an energy unit power supply system, characterized in that, include: Acquire operational data for each energy subsystem; the operational data includes project operational data corresponding to several operational items. Generate parameter change curves corresponding to the running project based on the project operation data; Based on the change curves of each parameter, fault analysis data corresponding to the energy subsystem is generated. The fault analysis data includes several parameter values ​​corresponding to each operating item. Acquire predictive environmental data, input the predictive environmental data and fault analysis data into the fault prediction model corresponding to the energy subsystem, and obtain the predictive analysis results corresponding to the energy subsystem; The fault prediction model is trained based on the digital twin model corresponding to the energy subsystem; Generate a predictive processing scheme based on the predictive analysis results; The fault analysis data for the energy subsystem is generated based on the variation curves of various parameters, including: S1: Obtain the parameter change curves corresponding to several operating items of the energy subsystem, as well as the initially set number of sampling periods and period sampling interval; S2: Monitoring and analysis data are obtained by sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval; S3: Generate data characterization based on monitoring and analysis data and parameter change curves; extract the project characterization corresponding to each running project in the data characterization, and determine whether each project characterization is greater than or equal to the set characterization threshold. If yes, proceed to S5; otherwise, proceed to S4. S4: Adjust the initially set number of sampling periods and periodic sampling interval based on the data representation degree, and then proceed to S1; S5: Use the monitoring and analysis data as fault analysis data; The generation of data characterization based on monitoring and analysis data and parameter change curves includes: Extract several parameter values ​​corresponding to each operational item from the monitoring and analysis data, fit the parameter values ​​into a parameter value change curve, and obtain the corresponding curve segment from the parameter change curve; Input the parameter value change curve and curve segment into the set characterization generation function to obtain the corresponding project characterization; obtain the project characterization corresponding to each running project, and integrate the project characterization into data characterization; One form of the characterization generation function includes: ； in, As the project's representativeness; Number The corresponding reference weights; For curved segments; The parameter value variation curve; The designated verification point number is... , The set number of characterization verification points; Let x be the time of the characterization verification point, and , The starting point time corresponding to the parameter value change curve. The end point time corresponding to the parameter value change curve; The defined interval for characterization considerations.

2. The fault handling method for an energy unit power supply system according to claim 1, characterized in that, The process of generating parameter change curves corresponding to the running project based on project operation data includes: Extract several parameter values ​​from each collection period in the project's operational data; Obtain the set number of fitting periods, group each sampling period according to the number of fitting periods to obtain several partial fitting data groups, wherein the sampling periods within the partial fitting data groups are continuous; and there are overlapping sampling periods between different partial fitting data groups; fit the parameter values ​​within each sampling period of each partial fitting data group into a partial parameter curve according to its corresponding acquisition order; fit the partial parameter curve into the parameter change curve of the running project.

3. The fault handling method for an energy unit power supply system according to claim 1, characterized in that, The monitoring and analysis data obtained by sampling the parameter change curve based on the number of sampling periods and the periodic sampling interval includes: Obtain the number of sampling periods and the periodic sampling interval. Select consecutive sampling periods as sampling segments based on the number of sampling periods. Extract the curve segment corresponding to the sampling segment in the parameter change curve. Set several sampling points based on the periodic sampling interval. Sample the curve segment based on the sampling points to obtain several parameter values. Integrate the parameter values ​​in chronological order to form the monitoring and analysis data of the corresponding running project. The monitoring and analysis data corresponding to each running project are acquired sequentially; the monitoring and analysis data corresponding to each running project are then integrated into the final monitoring and analysis data.

4. A fault handling method for an energy unit power supply system according to claim 1, characterized in that, The adjustment of the initially set number of sampling periods and the period sampling interval based on data representation includes: Extract the project characterization of several running items from the data characterization; Input the project characterization of each running project into the set sampling period number adjustment function to obtain the adjusted sampling period number; Input the project characterization of each running project into the set sampling interval adjustment function to obtain the adjusted sampling interval.

5. A fault handling method for an energy unit power supply system according to claim 1, characterized in that, One method for constructing the fault prediction model includes: The equipment parameters of the energy subsystem are obtained, and a digital twin model of the energy subsystem is constructed based on the equipment parameters; the input data of the digital twin model are the environmental data and fault analysis data of the corresponding energy subsystem, and the output data is the monitoring data. A fault identification model is constructed based on a neural network model. The input data of the fault identification model is monitoring data, and the output data is the state of the energy subsystem. The output data from the digital twin model is input into the input of the fault identification model; the energy subsystem state output by the fault identification model is used as the predictive analysis result to construct a fault prediction model.

6. A fault handling method for an energy unit power supply system according to claim 1, characterized in that, The generation of a predictive processing scheme based on the predictive analysis results includes: Extract the energy subsystem status from the predictive analysis results. When the energy subsystem status is abnormal, retrieve the corresponding processing solution from the processing solution query table based on the abnormal status, and use the processing solution as the predictive processing solution. The processing solution query table includes several query items and several processing solutions corresponding to the query items. The query items are several types of abnormal statuses.

7. A fault handling system for an energy unit power supply system, based on the application of a fault handling method for an energy unit power supply system according to any one of claims 1 to 6, characterized in that, include: Data acquisition module and data analysis module; among which, The data acquisition module includes an operational data acquisition subunit and an environmental data acquisition subunit; The operational data acquisition subunit is used to collect operational data from each energy subsystem. The environmental data acquisition subunit is used to collect predicted environmental data from various energy subsystems. The data analysis module includes a front-end subunit, a model invocation subunit, and a back-end subunit; The front-end subunit is used to generate parameter change curves corresponding to the operational project based on the project operation data; and to generate fault analysis data corresponding to the energy subsystem based on each parameter change curve, wherein the fault analysis data includes several parameter values ​​corresponding to each operational project. The model invocation subunit is used to invoke the corresponding fault prediction model according to the energy subsystem corresponding to the fault analysis data. The predicted environment data and fault analysis data are input into the fault prediction model corresponding to the energy subsystem to obtain the prediction analysis result corresponding to the energy subsystem. The fault prediction model is trained based on the digital twin model corresponding to the energy subsystem. The post-subunit is used to generate a predictive processing scheme based on the predictive analysis results.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on the fault handling system for an energy unit power supply system, perform a fault handling method for an energy unit power supply system as described in any one of claims 1 to 6.

Images