Multi-dimensional full-process hydroelectric data processing method

Abstract

The invention discloses a multi-dimensional full-process hydroelectric data processing method, which comprises the steps of data acquisition, real-time processing and fusion of various types of acquired monitoring data, collaborative sharing of multi-source heterogeneous data, and construction of an event relationship and an operation model. Establishing a distributed unit or a block space grid to carry out cloud high-performance parallel operation on the event relation and the operation model; based on a high-performance parallel computing result, constructing a data space structure oriented to the hydroelectric business; based on the characteristic that the operating environment of the drainage basin power station equipment is complex and changeable, the operating state information of the equipment is observed and sampled through the active sensing technology, and data in the time or space dimension is obtained; monitoring data, environment data and the like are subjected to fusion analysis through a data fusion technology, limitation and uncertainty of single data analysis are avoided, and more comprehensive perception and recognition of equipment and environment are formed.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower data processing, and in particular to a multi-dimensional full-process hydropower data processing method. Background Art

[0002] The data in the hydropower industry has the morphological characteristics of multimodality, strong correlation, and high heterogeneity. There are problems such as unclear data management classification, non-standard data description, and inaccurate data matching. At the same time, there are also application characteristics such as high density and low value, and there is a problem of low data value density. In the existing technology, there is no technical solution that can integrate active perception and data fusion technology to achieve the full life cycle management of hydropower's massive multi-source heterogeneous, multi-dimensional and multi-morphic data, automatic aggregation, fusion storage, business preprocessing, and high-value scenario extraction. Therefore, it is necessary to design a multi-dimensional and full-process hydropower data processing method to solve the above problems. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a multi-dimensional full-process hydropower data processing method. The method aims to solve the problem that data processing and data fusion management are relatively difficult due to the multi-modal, strong correlation and high heterogeneity characteristics of the hydropower industry in the prior art. By constructing a hydropower full-process metadata system, a real-time, accurate, scientific and reliable business service model can be created to form a multi-dimensional full-process hydropower data space.

[0004] In order to achieve the above technical effects, the technical solution adopted by the present invention is: A multi-dimensional full-process hydropower data processing method comprises the following steps: S1, Data Collection: Data on unit status, weather and water conditions, dam status and ecological environment are collected through a variety of data collection equipment, including multiple types of sensors; S2, real-time processing and fusion of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, including data preprocessing, feature extraction and feature fusion; S3, build event relationship and operation model: Through the extracted data features, the relationship between the equipment and the equipment, as well as the relationship between the equipment and the meteorological and water conditions, the dam status and the ecological environment is established, the subordinate monitoring data is associated, and the calculation model to be run when the event is triggered is constructed; S4, high performance parallel computing: Establish distributed units or block space grids to perform high-performance parallel computing of event relationships and computing models in the cloud; S5, builds a data space structure oriented towards the hydropower business based on high-performance parallel computing results.

[0005] Preferably, the collected data include temperature, vibration, current and voltage monitoring data of the generator, transformer and turbine of the unit status, dam body displacement, seepage and deformation monitoring data of the dam, upstream and downstream water level, runoff, air temperature and humidity and rainfall monitoring data of meteorological and water conditions, and water body, ecological flow and air quality monitoring data of the ecological environment.

[0006] Furthermore, in addition to multiple types of sensors, data collection also includes API interfaces for connecting to multiple systems of the hydropower plant, including API interfaces for connecting to the databases of the monitoring system, cascade dispatching data, online monitoring system, dam monitoring system and production management system, and directly obtaining monitoring data from the above systems to complete data collection; the monitoring system, cascade dispatching data, online monitoring system, dam monitoring system and production management system are existing systems of the hydropower plant.

[0007] Preferably, data preprocessing includes: According to the characteristics of hydropower data, data cleaning, data conversion and data reduction are carried out on the data; noise and abnormal values ​​in the data are removed through data cleaning, and data interruption, dead value, jump, over-range and invalid value of the hydropower equipment status data are cleaned; data conversion is used to convert the data into a unified time format; data reduction is used to reduce the amount of data to make data processing more efficient. When processing the status data of hydropower equipment, data other than operating conditions can be filtered out.

[0008] Preferably, feature extraction includes: Features are extracted from each data source. Features include original features of the original data or new features calculated by algorithms. Features include dimensions such as logical devices, instantiated devices, organizations, and time.

[0009] Preferably, feature fusion includes: The features of different data sources are fused to form a new feature representation, and the new features are used to complete the aggregation and fusion of multivariate heterogeneous data; the feature fusion methods include feature selection, feature weighting and feature transformation.

[0010] Preferably, constructing the event relationship and operation model includes: S301, define events and relationships: Define various events and time types of hydropower station units, including startup, shutdown and failure, and define the relationship between events, including sequential relationship, parallel relationship and causal relationship; S302, establish an event relationship diagram: According to the defined events and relationships, an event relationship diagram is established; the parent event is formed according to the superior-subordinate relationship of the event, and the predecessor event is formed according to the order in which the events occur; S303, establishing a computational model: According to the event relationship diagram, a corresponding operation model is established; the operation model includes a mathematical model and a computer model, which is used to calculate and predict the operating status of the hydropower station unit; S304, Model Verification and Adjustment: The model is verified and adjusted through actual operating data; the unit operating data over a period of time is collected and used as a test set to verify the accuracy of the model; if the model's prediction results do not match the actual operating status, the operation model is adjusted by adjusting the model parameters and improving the model structure.

[0011] Preferably, establishing a distributed unit or a block space grid to perform high-performance parallel computing of event relationships and computing models on the cloud includes: S401, task breakdown: Decompose the event into multiple subtasks, which are executed simultaneously on multiple processors in parallel. In the hydropower station unit event relationship model, the subsystem scheduling instruction task is decomposed into the smallest unit according to the steps of "unit-parameter type-indicator". S402, parallel strategy definition: According to the characteristics of the event, select one or both of data parallelism and task parallelism; in the event relationship model of the hydropower station unit, use the space decomposition method to perform parallel computing, the units with superior-subordinate relationships are serially calculated, and the units without superior-subordinate relationships are parallel or serially calculated; S403, task allocation and scheduling: Subtasks are assigned to multiple processors or computing nodes through task scheduling algorithms. In the event relationship model of hydropower station units, the hydropower system is decomposed into multiple subsystems according to the characteristics and spatial distribution pattern of the hydropower station, and scheduling and coordination optimization are carried out area by area. The cyclic scheduling algorithm is used as the task scheduling algorithm to assign all subtasks to all processors in a certain order. After each processor completes a subtask, it takes out a new subtask from the remaining subtasks for processing. S404, Synchronization and Communication: Data exchange and task synchronization are performed between multiple processors or computing nodes, so that all processors or computing nodes can correctly complete the assigned tasks; in the event relationship model of the hydropower station unit, the neighborhood information is updated for each spatial subdomain before each calculation; the message passing interface MPI is used for synchronization and communication, and each processor establishes a message sending and receiving channel with other processors to perform data exchange and task synchronization between processors; S405, parallel computing optimization: Perform algorithm optimization, data structure optimization, parallel algorithm optimization and hardware optimization on the parallel computing process; in the event relationship model of the hydropower station unit, parallel computing optimization is performed by applying the large system decomposition and coordination method and multi-core cluster parallel computing; S406, performance evaluation: The performance of parallel computing is evaluated through actual operation data to determine the effect of parallel computing. In the event relationship model of the hydropower station unit, the number of floating-point operations FLOPS is used to evaluate the performance of parallel computing. The performance of parallel computing is obtained by calculating the number of floating-point operations that can be performed per second.

[0012] Preferably, in step S4, a data space structure oriented towards the hydropower business is constructed based on the high-performance parallel computing results, including an object-centric model, a business subject model and an event relationship model.

[0013] Preferably, a multi-dimensional full-process hydropower data processing system is provided, wherein the system executes the multi-dimensional full-process hydropower data processing method described above; the system comprises a perception layer, a platform layer and an application layer; The perception layer is used to realize data collection, including data collection on unit status, meteorological and water conditions, dam status and ecological environment; The platform layer is used to implement data governance, which includes real-time processing and integration of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, construction of event relationship and operation models, and establishment of distributed units or block space grids to perform high-performance parallel computing of event relationship and operation models in the cloud; The application layer is used for data applications, including intelligent operation and maintenance, intelligent maintenance, smart scheduling and intelligent decision-making.

[0014] The beneficial effects of the present invention are as follows: 1. Aiming at the complex random disturbance and dynamic time-varying characteristics of power stations in river basins, the present invention collects dynamic data of equipment through real-time monitoring, conducts high-quality and comprehensive fusion analysis on multi-source time-series basic data such as measurement information and external environment information, dynamically perceives the operating conditions of the units, and establishes a panoramic situation awareness architecture of real-time monitoring, intelligent analysis, optimization control, and remote management.

[0015] 2. The present invention establishes hydropower event relationships and operation models based on hydropower business, and monitors and analyzes the state parameters of the unit in real time through the operation model; at the same time, it provides high-performance parallel scheduling calculations to improve the efficiency of model operations and enhance the real-time performance of data analysis.

[0016] 3. The present invention creates a multi-dimensional, cross-process business data mart through data processing tools oriented to business scenarios and business processes, constructs a business-centric trusted and available data middle layer, and forms a data space structure with hydropower business characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the architecture of the multi-dimensional full-process hydropower data processing system in Example 2 of the present invention. DETAILED DESCRIPTION

[0018] Embodiment 1: like Figure 1 As shown, a multi-dimensional full-process hydropower data processing method includes the following steps: S1, Data Collection: Data on unit status, weather and water conditions, dam status and ecological environment are collected through a variety of data collection equipment, including multiple types of sensors; S2, real-time processing and fusion of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, including data preprocessing, feature extraction and feature fusion; S3, build event relationship and operation model: Through the extracted data features, the relationship between the equipment and the equipment, as well as the relationship between the equipment and the meteorological and water conditions, the dam status and the ecological environment is established, the subordinate monitoring data is associated, and the calculation model to be run when the event is triggered is constructed; S4, high performance parallel computing: Establish distributed units or block space grids to perform high-performance parallel computing of event relationships and computing models in the cloud; S5, builds a data space structure oriented towards the hydropower business based on high-performance parallel computing results.

[0019] Preferably, the collected data include temperature, vibration, current and voltage monitoring data of the generator, transformer and turbine of the unit status, dam body displacement, seepage and deformation monitoring data of the dam, upstream and downstream water level, runoff, air temperature and humidity and rainfall monitoring data of meteorological and water conditions, and water body, ecological flow and air quality monitoring data of the ecological environment.

[0020] Furthermore, in addition to multiple types of sensors, data collection also includes API interfaces for connecting to multiple systems of the hydropower plant, including API interfaces for connecting to the databases of the monitoring system, cascade dispatching data, online monitoring system, dam monitoring system and production management system, and directly obtaining monitoring data from the above systems to complete data collection; the monitoring system, cascade dispatching data, online monitoring system, dam monitoring system and production management system are existing systems of the hydropower plant.

[0021] Preferably, data preprocessing includes: According to the characteristics of hydropower data, data cleaning, data conversion and data reduction are carried out on the data; noise and abnormal values ​​in the data are removed through data cleaning, and data interruption, dead value, jump, over-range and invalid value of the hydropower equipment status data are cleaned; data conversion is used to convert the data into a unified time format; data reduction is used to reduce the amount of data to make data processing more efficient. When processing the status data of hydropower equipment, data other than operating conditions can be filtered out.

[0022] Preferably, feature extraction includes: Features are extracted from each data source. Features include original features of the original data or new features calculated by algorithms. Features include dimensions such as logical devices, instantiated devices, organizations, and time.

[0023] Preferably, feature fusion includes: The features of different data sources are fused to form a new feature representation, and the new features are used to complete the aggregation and fusion of multivariate heterogeneous data; the feature fusion methods include feature selection, feature weighting and feature transformation.

[0024] Preferably, constructing the event relationship and operation model includes: S301, define events and relationships: Define various events and time types of hydropower station units, including startup, shutdown and failure, and define the relationship between events, including sequential relationship, parallel relationship and causal relationship; S302, establish an event relationship diagram: According to the defined events and relationships, an event relationship diagram is established; the parent event is formed according to the superior-subordinate relationship of the event, and the predecessor event is formed according to the order in which the events occur; Taking the startup event as an example, during the startup process, the event can be divided into startup operation, cooling equipment operation, auxiliary equipment operation, starting the hydraulic system to remove the lock spindle, adding high-pressure oil, starting the speed governor, closing the GCB switch knife switch to the synchronization device, adding the excitation system, and grid connection operation. The sub-events are in a sequential relationship; the cooling equipment operation can be divided into lower-level events such as cooler operation and electric valve operation, and the cooling equipment operation is the parent event; the event relationship diagram is a graphical representation method that can clearly show the relationship between events; S303, establishing a computational model: According to the event relationship diagram, a corresponding operation model is established; the operation model includes a mathematical model and a computer model, which is used to calculate and predict the operating status of the hydropower station unit; Taking the startup event as an example, during the startup event, when the governor startup sub-event is completed and the speed is greater than 5%, the unit operation state model calculation is triggered, the unit operation state time interval is recorded, and the key unit status data is interrupted and audited in real time; when the grid-connected operation sub-time is completed, and the unit speed is greater than 95%, and the active power is greater than 5MW, the unit grid-connected interval model and steady-state interval model calculation are triggered. The grid-connected interval model performs real-time audits on the dead values, jumps, over-ranges, and invalid values ​​of the key unit status data, and the steady-state interval model calculates the relevant indicators of the key status data during the operation of this unit in real time, such as: maximum value, minimum value, median, and other related indicators; S304, Model Verification and Adjustment: The model is verified and adjusted through actual operating data; the unit operating data over a period of time is collected and used as a test set to verify the accuracy of the model; if the model's prediction results do not match the actual operating status, the operation model is adjusted by adjusting the model parameters and improving the model structure.

[0025] Preferably, establishing a distributed unit or a block space grid to perform high-performance parallel computing of event relationships and computing models on the cloud includes: S401, task breakdown: Decompose the event into multiple subtasks, which are executed simultaneously on multiple processors in parallel. In the hydropower station unit event relationship model, the steps of "unit-parameter type-indicator" are decomposed into the smallest unit to execute the subsystem scheduling instruction task. For example, for the unit startup event, "1F-temperature-upper guide temperature maximum value" can be used as the smallest scheduling instruction task. S402, parallel strategy definition: According to the characteristics of the event, select one or both of data parallelism and task parallelism; in the event relationship model of the hydropower station unit, use the spatial decomposition method to perform parallel calculations, and the units with a superior-subordinate relationship are calculated serially, while the units without a superior-subordinate relationship choose parallel calculations or serial calculations; for example, during the unit startup process, the steady-state interval should be calculated first, and then the key parameter indicators of the unit should be calculated, while parameters such as temperature, vibration, current, and voltage do not have a superior-subordinate relationship and can be calculated in parallel and synchronously; S403, task allocation and scheduling: Subtasks are assigned to multiple processors or computing nodes through task scheduling algorithms. In the event relationship model of hydropower station units, the hydropower system is decomposed into multiple subsystems according to the characteristics and spatial distribution pattern of the hydropower station, and scheduling and coordination optimization are carried out area by area. The cyclic scheduling algorithm is used as the task scheduling algorithm to assign all subtasks to all processors in a certain order. After each processor completes a subtask, it takes out a new subtask from the remaining subtasks for processing. S404, Synchronization and Communication: Data exchange and task synchronization are performed between multiple processors or computing nodes, so that all processors or computing nodes can correctly complete the assigned tasks; in the event relationship model of the hydropower station unit, the neighborhood information is updated for each spatial subdomain before each calculation; the message passing interface MPI is used for synchronization and communication, and each processor establishes a message sending and receiving channel with other processors to perform data exchange and task synchronization between processors; S405, parallel computing optimization: Perform algorithm optimization, data structure optimization, parallel algorithm optimization and hardware optimization on the parallel computing process; in the event relationship model of the hydropower station unit, parallel computing optimization is performed by applying the large system decomposition and coordination method and multi-core cluster parallel computing; S406, performance evaluation: The performance of parallel computing is evaluated through actual operation data to determine the effect of parallel computing. In the event relationship model of the hydropower station unit, the number of floating-point operations FLOPS is used to evaluate the performance of parallel computing. The performance of parallel computing is obtained by calculating the number of floating-point operations that can be performed per second.

[0026] Preferably, in step S4, a data space structure oriented towards the hydropower business is constructed based on the high-performance parallel computing results, including an object-centric model, a business subject model and an event relationship model.

[0027] Furthermore, the object-centric model takes the hydropower equipment object or business object as the center, integrates multi-source, multi-type and multi-dimensional data according to the data fusion characteristics, and comprehensively displays the data associated with the center; taking the equipment object as an example, the instantiated hydropower equipment can be associated with all the monitoring data and all business data of the equipment through the equipment KKS code, which includes static information such as basic equipment information and setting sheets, dynamic data such as monitoring data, equipment operation, equipment maintenance, rules and algorithm data of advanced applications such as status assessment and maintenance decision-making, as well as knowledge bases such as fault case libraries and sample libraries, to form a data panorama centered on the instantiated equipment.

[0028] Furthermore, the business theme model is centered on the hydropower business object, and conducts exploratory multi-dimensional analysis on the wide table of business data related to the business; during the analysis process, the data is presented using charts, graphs, and dashboards to more intuitively display the information reflected by the business; taking the defect business as an example, the wide table of defect data is combined with the actual hydropower business to conduct exploratory analysis from the dimensions of profession, system, completion status, time, etc., to quickly build an ideal analysis chart and form a defect business theme model; at the same time, the analysis dashboard has a sharing function, which can quickly serve practical applications, reports / statements or solutions, etc.

[0029] Furthermore, with hydropower events as the center, business activities are disassembled to form minimum event units and their subordinate data. The minimum event units are associated through event relationships to form a completed hydropower event process, and the subordinate data in each link are displayed at the same time. Through the construction of event models, the subordinate data can be put into the operation model when the event is triggered, thereby improving the real-time and operation efficiency of data analysis.

[0030] Preferably, a multi-dimensional full-process hydropower data processing system is provided, wherein the system executes the multi-dimensional full-process hydropower data processing method described above; the system comprises a perception layer, a platform layer and an application layer; The perception layer is used to realize data collection, including data collection on unit status, meteorological and water conditions, dam status and ecological environment; The platform layer is used to implement data governance, which includes real-time processing and integration of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, construction of event relationship and operation models, and establishment of distributed units or block space grids to perform high-performance parallel computing of event relationship and operation models in the cloud; The application layer is used for data applications, including intelligent operation and maintenance, intelligent maintenance, smart scheduling and intelligent decision-making.

[0031] Embodiment 2: like Figure 2 As shown, this embodiment provides a specific architecture of a multi-dimensional full-process hydropower data processing system, including a perception layer, a platform layer and an application layer; The perception layer is used to realize data collection, including data collection on unit status, meteorological and water conditions, dam status and ecological environment; the perception layer connects data with the oil and gas monitoring system, equipment status online monitoring system, cascade dispatching center, dam monitoring system, monitoring system, EPMS system and watershed support system of the hydropower plant to obtain relevant monitoring data; The platform layer is used to implement data governance, which includes real-time processing and fusion of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, construction of event relationship and operation models, and establishment of distributed units or block space grids to perform high-performance parallel computing of event relationship and operation models in the cloud; the platform layer includes four sub-modules: data asset management tools, object construction tools, data service tools, and processing and computing tools.

[0032] The data asset management tool includes a source data layer and a data warehouse. The source data layer is directly connected to the perception layer, and the data warehouse stores advanced application business data. The monitoring data of the source data layer and the advanced application business data of the data warehouse are sent to the processing and calculation tool sub-module, which performs data quality audits, builds structured data sets, end-to-end processing, indicator configuration, and model configuration to process and calculate the source layer data. The processed and calculated data enters the data warehouse and is stored as the model output results and the processing flow output results. An asset catalog is constructed using the data of the source data layer and the data warehouse, and the asset catalog includes organizational structure, business entities, security levels, and asset forms.

[0033] The data asset management tool is connected with the object construction tool and the data service tool. The object construction tool builds a data space structure oriented towards the hydropower business based on the results of high-performance parallel computing, including an object-centric model, a business subject model and an event relationship model. The object construction tool is connected with the data service tool, and the data service includes a data mart and data subscription.

[0034] The data service tool is connected to the application layer, which is used for data applications, including intelligent operation and maintenance, intelligent maintenance, smart scheduling and intelligent decision-making; the new data generated by the application layer is re-input into the data asset management tool.

Claims

1. A multi-dimensional full-process hydropower data processing method, It is characterized in that The following steps are involved: S1, Data Collection: Data on unit status, weather and water conditions, dam status and ecological environment are collected through a variety of data collection equipment, including multiple types of sensors; S2, real-time processing and fusion of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, including data preprocessing, feature extraction and feature fusion; S3, build event relationship and operation model: Through the extracted data features, the relationship between the equipment and the equipment, as well as the relationship between the equipment and the meteorological and water conditions, the dam status and the ecological environment is established, the subordinate monitoring data is associated, and the calculation model to be run when the event is triggered is constructed; S4, high performance parallel computing: Establish distributed units or block space grids to perform high-performance parallel computing of event relationships and computing models in the cloud; S5, builds a data space structure oriented towards the hydropower business based on high-performance parallel computing results.

2. A multi-dimensional full-process hydropower data processing method according to claim 1, It is characterized in that In step S1, the collected data include the temperature, vibration, current and voltage monitoring data of the generator, transformer and turbine of the unit status, the dam body displacement, seepage and deformation monitoring data of the dam, the upstream and downstream water level, runoff, air temperature and humidity and rainfall monitoring data of the meteorological and water conditions, and the water body, ecological flow and air quality monitoring data of the ecological environment.

3. A multi-dimensional full-process hydropower data processing method according to claim 2, It is characterized in that In step S2, data preprocessing includes: According to the characteristics of hydropower data, data cleaning, data conversion and data reduction are carried out on the data; noise and abnormal values ​​in the data are removed through data cleaning, and data interruption, dead value, jump, over-range and invalid value of the hydropower equipment status data are cleaned; data conversion is used to convert the data into a unified time format; and data reduction is used to reduce the amount of data.

4. A multi-dimensional full-process hydropower data processing method according to claim 3, It is characterized in that In step S2, feature extraction includes: Features are extracted from each data source. Features include original features of the original data or new features calculated by algorithms. Features include dimensions such as logical devices, instantiated devices, organizations, and time.

5. A multi-dimensional full-process hydropower data processing method according to claim 4, It is characterized in that In step S2, feature fusion includes: The features of different data sources are fused to form a new feature representation, and the new features are used to complete the aggregation and fusion of multivariate heterogeneous data; the feature fusion methods include feature selection, feature weighting and feature transformation.

6. A multi-dimensional full-process hydropower data processing method according to claim 1, It is characterized in that In step S3, constructing the event relationship and operation model includes: S301, define events and relationships: Define various events and time types of hydropower station units, including startup, shutdown and failure, and define the relationship between events, including sequential relationship, parallel relationship and causal relationship; S302, establish an event relationship diagram: According to the defined events and relationships, an event relationship diagram is established; the parent event is formed according to the superior-subordinate relationship of the event, and the predecessor event is formed according to the order in which the events occur; S303, establishing a computational model: According to the event relationship diagram, a corresponding operation model is established; the operation model includes a mathematical model and a computer model, which is used to calculate and predict the operating status of the hydropower station unit; S304, Model Verification and Adjustment: The model is verified and adjusted through actual operating data; the unit operating data over a period of time is collected and used as a test set to verify the accuracy of the model; if the model's prediction results do not match the actual operating status, the operation model is adjusted by adjusting the model parameters and improving the model structure.

7. A multi-dimensional full-process hydropower data processing method according to claim 6, It is characterized in that In step S4, establishing a distributed unit or a block space grid to perform high-performance parallel computing on the event relationship and computing model in the cloud includes: S401, task breakdown: Decompose the event into multiple subtasks, which are executed simultaneously on multiple processors in parallel. In the hydropower station unit event relationship model, the subsystem scheduling instruction task is decomposed into the smallest unit according to the steps of "unit-parameter type-index". S402, parallel strategy definition: According to the characteristics of the event, select one or both of data parallelism and task parallelism; in the event relationship model of the hydropower station unit, use the space decomposition method to perform parallel computing, the units with superior-subordinate relationships are serially calculated, and the units without superior-subordinate relationships are parallel or serially calculated; S403, task allocation and scheduling: Subtasks are assigned to multiple processors or computing nodes through task scheduling algorithms. In the event relationship model of hydropower station units, the hydropower system is decomposed into multiple subsystems according to the characteristics and spatial distribution pattern of the hydropower station, and scheduling and coordination optimization are carried out area by area. The cyclic scheduling algorithm is used as the task scheduling algorithm to assign all subtasks to all processors in a certain order. After each processor completes a subtask, it takes out a new subtask from the remaining subtasks for processing. S404, Synchronization and Communication: Data exchange and task synchronization are performed between multiple processors or computing nodes, so that all processors or computing nodes can correctly complete the assigned tasks; in the event relationship model of the hydropower station unit, the neighborhood information is updated for each spatial subdomain before each calculation; the message passing interface MPI is used for synchronization and communication, and each processor establishes a message sending and receiving channel with other processors to perform data exchange and task synchronization between processors; S405, parallel computing optimization: Perform algorithm optimization, data structure optimization, parallel algorithm optimization and hardware optimization on the parallel computing process; in the event relationship model of the hydropower station unit, parallel computing optimization is performed by applying the large system decomposition and coordination method and multi-core cluster parallel computing; S406, performance evaluation: The performance of parallel computing is evaluated through actual operation data to determine the effect of parallel computing. In the event relationship model of the hydropower station unit, the number of floating-point operations FLOPS is used to evaluate the performance of parallel computing. The performance of parallel computing is obtained by calculating the number of floating-point operations that can be performed per second.

8. A multi-dimensional full-process hydropower data processing method according to claim 7, It is characterized in that In step S4, a data space structure oriented towards the hydropower business is constructed based on the high-performance parallel computing results, including an object-centric model, a business subject model and an event relationship model.

9. A multi-dimensional full-process hydropower data processing system, It is characterized in that The system implements the multi-dimensional full-process hydropower data processing method of any one of claims 1 to 8; the system includes a perception layer, a platform layer and an application layer; The perception layer is used to realize data collection, including data collection on unit status, meteorological and water conditions, dam status and ecological environment; the platform layer is used to realize data governance, which includes real-time processing and fusion of various types of collected monitoring data, collaborative sharing of multi-source heterogeneous data, construction of event relationship and operation models, and establishment of distributed units or block spatial grids to perform high-performance parallel computing of event relationship and operation models in the cloud; the application layer is used for data application.

Images