Data Acquisition Methods and Systems Based on Hydropower Stations

Abstract

This invention relates to the field of data acquisition technology, and particularly to a data acquisition method and system based on hydropower stations. The method includes acquiring a regional map of the water area to be surveyed; segmenting the regional map based on the registered locations of hydropower stations to determine the acquisition area for each hydropower station; acquiring the operational data of each hydropower station in real time; periodically determining the characteristic data nodes of each hydropower station based on the operational data; comparing the characteristic data nodes of different hydropower stations to determine the environmental anomaly level of each hydropower station; and determining the amount of data acquisition resources based on the environmental anomaly level. The amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency. This invention enables dynamic allocation of acquisition resources on demand, avoiding the problems of resource redundancy at low-risk power stations and insufficient acquisition at high-risk power stations under a fixed acquisition mode.

Description

Technical Field

[0001] This invention relates to the field of data acquisition technology, and in particular to a data acquisition method and system based on hydropower stations. Background Technology

[0002] Data collection on the aquatic environment and unit operation of hydropower stations is a fundamental step in hydropower operation and maintenance, watershed ecological monitoring, and risk early warning. Currently, multiple hydropower stations in the watershed are conducting data collection simultaneously, which requires both monitoring of unit operation status and perception of aquatic environmental risks. In the existing scenario, in the case of a cluster of hydropower stations, data collection resources are mostly allocated in a fixed manner, which cannot adapt to the different operational anomalies and environmental fluctuations of each station, and is prone to problems such as waste of collection resources or omission of key data.

[0003] To address this issue, existing technologies often employ a uniform and fixed collection frequency, scope, and type of data for comprehensive data collection, without differentiating between the power station's operating status and the degree of anomaly in the surrounding environment when uniformly allocating collection resources. While this approach achieves full coverage of data across the entire region, it overlooks the changing patterns of operational data from different hydropower stations and the temporal differences in environmental fluctuations. Therefore, how to quantify the degree of environmental anomaly at hydropower stations in order to dynamically and adaptively allocate data collection resources for each hydropower station is the technical problem that this invention aims to solve. Summary of the Invention

[0004] The purpose of this invention is to provide a data acquisition method and system based on hydropower stations to solve the problem of "how to achieve adaptive acquisition of key data in multiple scenarios of power distribution networks" mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A data acquisition method based on hydropower stations, the method comprising:

[0007] Obtain a regional map of the water area to be sampled, and divide the regional map based on the registered locations of hydropower stations to determine the sampling areas for each hydropower station;

[0008] Real-time acquisition of operational data from each hydropower station, and periodic determination of characteristic data nodes for each hydropower station based on the operational data; the generation process of the characteristic data nodes is as follows: generating prediction data based on the operational data of the hydropower station, and determining the characteristic data nodes based on the prediction data, wherein the characteristic data nodes are time series composed of relative time points;

[0009] By comparing the characteristic data nodes of different hydropower stations, the environmental anomaly degree of each hydropower station can be determined.

[0010] The amount of data acquisition resources is determined based on the environmental anomaly level; the amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency.

[0011] As a further aspect of the present invention, the step of acquiring the operational data of each hydropower station in real time and periodically determining the characteristic data nodes of each hydropower station based on the operational data includes:

[0012] Real-time acquisition of hydropower station operation data yields an operation data sequence, and feature extraction is performed on the operation data sequence to obtain operation status features;

[0013] Based on the characteristics of the operating status, the system matches and generates data change trajectories in a preset database, and determines the confidence probability of each data change trajectory.

[0014] Based on the reliable probability comparison of the change trajectory of each data point at regular intervals, the divergence nodes of data change are identified and used as feature data nodes.

[0015] As a further aspect of the present invention, the steps of acquiring real-time hydropower station operation data, obtaining an operation data sequence, and extracting features from the operation data sequence to obtain operation status features include:

[0016] Acquire monitoring data for water turbines, generators, transformers, and auxiliary equipment;

[0017] The monitoring data is sorted into time series according to the collection time to generate time series data;

[0018] Extract the amplitude and direction of change of each monitoring parameter between continuous acquisition times to generate parameter change characteristics;

[0019] Calculate the data change distance between monitoring data corresponding to adjacent acquisition times, and generate a change distance sequence;

[0020] The parameter change characteristics are fused with the change distance sequence to generate the operating status characteristics.

[0021] As a further aspect of the present invention, the step of matching and generating data change trajectories in a preset database based on operational status characteristics, and determining the confidence probability corresponding to each data change trajectory, includes:

[0022] Based on the characteristics of the operating status, retrieve the reference status from the historical operating data that matches the current status;

[0023] For any reference state, extract the historical state corresponding to the subsequent time step of the reference state based on a preset time step;

[0024] Sort historical states in chronological order to construct data change trajectories;

[0025] Based on the characteristics of the running status, each data change trajectory is traversed, the matching degree is determined in real time, and the reliability probability of the data change trajectory is determined according to the matching degree at different times.

[0026] As a further aspect of the present invention, the step of identifying data change divergence nodes as feature data nodes based on a reliable probability comparison of each data change trajectory during timing includes:

[0027] At preset time intervals, extract the state nodes corresponding to each data change trajectory and construct the data state vector corresponding to the state node.

[0028] The data change trajectory with the highest confidence probability is selected as the baseline trajectory;

[0029] Calculate the state difference values ​​between the data state vectors at each position of different data change trajectories and the baseline trajectory to obtain a sequence of difference values ​​for each data change trajectory relative to the baseline trajectory;

[0030] Based on a preset trajectory order, all difference value sequences are statistically analyzed to construct a difference matrix; one dimension of the difference matrix represents the trajectory position, which corresponds to time information, and the other dimension represents the trajectory type.

[0031] Based on the location, an array is extracted from the difference matrix, and the elements in the array are accumulated to obtain the overall difference value at that location.

[0032] Select the location where the overall difference value reaches a preset threshold, and use the time point of the query location relative to the current time as the feature data node.

[0033] As a further aspect of the present invention, the step of comparing the characteristic data nodes of different hydropower stations to determine the environmental anomaly degree of each hydropower station includes:

[0034] By statistically analyzing the characteristic data nodes of different hydropower stations in chronological order, a time point sequence is obtained.

[0035] For any hydropower station, compare the time sequence of this hydropower station with the time sequence of other hydropower stations, and determine the time difference of each number in sequence to obtain the time difference sequence; wherein, to determine the time difference of each number, the starting point alignment process is executed first.

[0036] Calculate the mean sequence of the time difference series, and determine the total time difference based on the mean sequence;

[0037] The total time difference of all hydropower stations is calculated to determine the degree of environmental anomaly.

[0038] The present invention also provides a data acquisition system based on a hydropower station, the system comprising:

[0039] The area segmentation module is used to obtain an area map of the water area to be sampled, and to segment the area map based on the registered locations of hydropower stations to determine the sampling area for each hydropower station.

[0040] The feature node generation module is used to acquire the operation data of each hydropower station in real time and determine the feature data nodes of each hydropower station periodically based on the operation data. The generation process of the feature data nodes is as follows: generate prediction data based on the operation data of the hydropower station, and determine the feature data nodes based on the prediction data. The feature data nodes are time series composed of relative time points.

[0041] The environmental assessment module is used to compare the characteristic data nodes of different hydropower stations to determine the degree of environmental anomaly of each hydropower station.

[0042] The data acquisition and adjustment module is used to determine the amount of data acquisition resources based on the environmental anomaly level; the amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency.

[0043] As a further embodiment of the present invention, the feature node generation module includes:

[0044] The feature extraction unit is used to acquire hydropower station operation data in real time, obtain operation data sequences, extract features from the operation data sequences, and obtain operation status features.

[0045] The trajectory matching unit is used to match and generate data change trajectories in a preset database based on the characteristics of the running status, and to determine the confidence probability of each data change trajectory.

[0046] The trajectory comparison unit is used to periodically compare the trajectory of each data change based on the confidence probability, identify the divergence nodes of data change, and use them as feature data nodes.

[0047] As a further embodiment of the present invention, the feature extraction unit includes:

[0048] The data acquisition subunit is used to acquire monitoring data corresponding to the turbine, generator, transformer and auxiliary equipment;

[0049] The sorting subunit is used to sort the monitoring data according to the collection time to generate time series data.

[0050] The change analysis subunit is used to extract the magnitude and direction of change of each monitoring parameter between continuous acquisition times, and generate parameter change characteristics;

[0051] The distance calculation subunit is used to calculate the data change distance between monitoring data corresponding to adjacent acquisition times and generate a change distance sequence.

[0052] The feature fusion subunit is used to fuse parameter change features with change distance sequences to generate operating status features.

[0053] As a further aspect of the present invention, the trajectory matching unit includes:

[0054] The reference matching subunit is used to retrieve a reference state from historical running data that matches the current state based on the running state characteristics.

[0055] The historical query subunit is used to extract the historical state corresponding to the subsequent time of any reference state based on a preset time step.

[0056] The trajectory construction sub-unit is used to sort historical states according to time order and construct data change trajectories;

[0057] The traversal sub-unit is used to traverse each data change trajectory based on the running state characteristics, determine the matching degree in real time, and determine the reliability probability of the data change trajectory based on the matching degree at different times.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] This invention divides the watershed into collection areas, extracts operational status characteristics from multiple types of monitoring data from generating units, matches historical data change trajectories, and identifies time-series divergent data nodes to accurately quantify the environmental anomaly degree of each hydropower station. Based on the anomaly degree, it dynamically adapts the collection scope, type, frequency, and other collection resources, enabling on-demand dynamic allocation of collection resources. This avoids the problems of resource redundancy at low-risk power stations and insufficient collection at high-risk power stations under fixed collection modes, improving the overall targeting and resource utilization of watershed data collection. Simultaneously, it ensures complete collection of critical operational and environmental data under abnormal conditions, adapting to the actual operation and maintenance needs of collaborative collection from multiple hydropower stations. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0061] Figure 1 A flowchart illustrating a data acquisition method based on a hydropower station provided in an embodiment of the present invention.

[0062] Figure 2 This is a block diagram illustrating the structural composition of a data acquisition system based on a hydropower station, as provided in an embodiment of the present invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0064] Figure 1 The implementation flow of the data acquisition method based on a hydropower station provided by an embodiment of the present invention is illustrated below, and the method includes:

[0065] Step S100: Obtain a regional map of the water area to be sampled, and divide the regional map based on the registered locations of hydropower stations to determine the sampling areas for each hydropower station.

[0066] The water area to be collected refers to the area where data needs to be collected. Figure 1 Generally, the data is existing and can be read directly. Once the hydropower station is determined, its location is also determined. The location of the hydropower station is queried in the regional map, and then the water area is divided based on the location to obtain the collection range of each hydropower station. The architecture of the technical solution of this invention is that a hydropower station collects data on a region. It can be equipped with drones or drones to collect data on a region. It is more like a data transfer station. At the hydropower station, there is sufficient energy to equip it with higher performance data storage and processing modules, and it can also interact with the cloud.

[0067] Step S200: Acquire the operation data of each hydropower station in real time, and periodically determine the characteristic data nodes of each hydropower station based on the operation data;

[0068] When acquiring data within the collection area of ​​a hydropower station, the operational data of each device within the station can also be obtained. Almost all devices have built-in information acquisition modules, and the acquired data is called the hydropower station's operational data. In fact, if we separate the external acquisition process from the hydropower station's own operational process and consider the data generated internally as a whole, this whole is the operational data mentioned above. Analyzing the operational data can determine the parameters describing the hydropower station's operational process, which are the feature data nodes mentioned above. The generation process of the feature data nodes is as follows: generating predictive data based on the hydropower station's operational data, and determining the feature data nodes based on the predictive data. The feature data nodes are time series composed of relative time points. This is actually a highly simplified parameter, indicating which moments are critical for the hydropower station. The data of the entire hydropower station is originally very complex, but after processing, it can be simplified into a very concise array.

[0069] Step S300: Compare the characteristic data nodes of different hydropower stations to determine the environmental anomaly degree of each hydropower station;

[0070] Based on the characteristic data nodes of the hydropower station, the current environment can be assessed by comparing the characteristic data nodes of different hydropower stations. One point needs to be noted in this process: even if different hydropower stations have different functional positioning and different equipment (in fact, they are different), after the processing of step S200, the data structure of the obtained characteristic data nodes is the same, which makes the comparison process proceed smoothly.

[0071] Step S400: Determine the amount of data acquisition resources based on the environmental anomaly level;

[0072] Finally, under the condition that the environmental anomaly degree is known, the amount of data acquisition resources is determined according to the environmental anomaly degree. The amount of data acquisition resources is generally proportional to the environmental anomaly degree. The amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency. For example, the larger the amount of data acquisition, the wider the acquisition range, the more types of data acquired, and the higher the acquisition frequency. This process is an external acquisition process relative to the hydropower station and is not the focus of the technical solution of this invention. When the computing power resources are sufficient (the hydropower station is the control center), there are many such acquisition schemes in the prior art, which will not be elaborated here.

[0073] Regarding step S100, the step of acquiring the operational data of each hydropower station in real time and periodically determining the characteristic data nodes of each hydropower station based on the operational data includes:

[0074] Real-time acquisition of hydropower station operation data yields an operation data sequence, and feature extraction is performed on the operation data sequence to obtain operation status features;

[0075] Based on the characteristics of the operating status, the system matches and generates data change trajectories in a preset database, and determines the confidence probability of each data change trajectory.

[0076] Based on the reliable probability comparison of the change trajectory of each data point at regular intervals, the divergence nodes of data change are identified and used as feature data nodes.

[0077] In one example of the technical solution of this invention, the generation process of feature data nodes is described. Real-time acquisition of hydropower station operation data yields an operation data sequence. Hydropower station operation data refers to the operation data of all equipment, which is essentially multi-source data. Data sequences for each type of data are constructed according to the data type, and then the statistical data sequences are collected according to a preset order. The resulting comprehensive data body is the hydropower station operation data. Feature extraction is performed on each operation data sequence to obtain operation status features. It can be imagined that when the operation data sequence is not unique, the obtained operation status features are also not unique and need to be statistically analyzed. The operation status features are actually a feature set.

[0078] Based on the obtained operational status characteristics, data prediction is performed to determine the data sequence at future moments. One feasible data prediction method is to generate data change trajectories by matching the operational status characteristics in a preset database and determining the confidence probability corresponding to each data change trajectory. The confidence probability is the probability of changing according to the corresponding trajectory. The trajectory refers to what the subsequent data will be like. Of course, other prediction processes such as function-based or neural network-based prediction processes can also be applied. Under the condition that the operational status characteristics are known, the prediction process is extremely conventional.

[0079] The data change trajectories obtained from data prediction are generally not unique. On the one hand, the database-based matching process mentioned above can itself match multiple data change trajectories. In addition, multiple prediction schemes can be used to obtain multiple data change trajectories. The data change trajectories are compared periodically based on the confidence probability to identify divergence nodes in data change, which are then used as feature data nodes. The meaning of this process is that for the multiple data change trajectories obtained, the data change trajectories are compared to determine which positions (times, after how much time) are key positions. That is, different data change trajectories have different directions at these positions, and these positions are the feature data nodes.

[0080] As a preferred embodiment of the technical solution of the present invention, the steps of acquiring real-time hydropower station operation data, obtaining an operation data sequence, and extracting features from the operation data sequence to obtain operation status features include:

[0081] Acquire monitoring data for water turbines, generators, transformers, and auxiliary equipment;

[0082] The monitoring data is sorted into time series according to the collection time to generate time series data;

[0083] Extract the amplitude and direction of change of each monitoring parameter between continuous acquisition times to generate parameter change characteristics;

[0084] Calculate the data change distance between monitoring data corresponding to adjacent acquisition times, and generate a change distance sequence;

[0085] The parameter change characteristics are fused with the change distance sequence to generate the operating status characteristics.

[0086] In one example of the technical solution of this invention, the process of generating data sequences and extracting features is described. Monitoring data corresponding to water turbines, generators, transformers and auxiliary equipment are acquired. The monitoring data are sorted into time series according to the acquisition time to generate time series data. The amplitude and direction of change of each monitoring parameter between consecutive acquisition times are extracted to generate parameter change features. The data change distance between monitoring data corresponding to adjacent acquisition times is calculated to generate a change distance sequence. Finally, the parameter change features and the change distance sequence are statistically analyzed to obtain the operating status features.

[0087] The above process takes water turbines, generators, transformers and auxiliary equipment as examples. Of course, there are many other devices in a hydropower station, all of which can be used for data acquisition and analysis. The analysis results of each device are then statistically combined to obtain the feature set, which is the operating status feature.

[0088] As a preferred embodiment of the technical solution of the present invention, the step of matching and generating data change trajectories in a preset database based on the characteristics of the operating state, and determining the confidence probability corresponding to each data change trajectory includes:

[0089] Based on the characteristics of the operating status, retrieve the reference status from the historical operating data that matches the current status;

[0090] For any reference state, extract the historical state corresponding to the subsequent time step of the reference state based on a preset time step;

[0091] Sort historical states in chronological order to construct data change trajectories;

[0092] Based on the characteristics of the running status, each data change trajectory is traversed, the matching degree is determined in real time, and the reliability probability of the data change trajectory is determined according to the matching degree at different times.

[0093] The above provides a deterministic data change trajectory generation scheme, which adopts a matching scheme to obtain multiple data change trajectories. Based on the characteristics of the operating state, a reference state matching the current state is retrieved from historical operating data. This means that after each acquisition of operating data, feature extraction is performed, and the extracted data is stored in a database as a historical database, where the data is referred to as historical operating data. Retrieving matching data is a simple data comparison process based on the same data structure. The data involved in the technical solution of this invention are all comprehensive basic data structures, and the data comparison process can be simplified to the most basic vector comparison and then synthesis process, which will not be elaborated here.

[0094] Then, for any reference state, the historical state corresponding to the subsequent time of the reference state is extracted based on a preset time step. This process must ensure that the time step is the same so that the dimension of the obtained data change trajectory is the same, and various comparison processes can be carried out smoothly in subsequent processing. The historical states are sorted according to the time order to construct the data change trajectory.

[0095] Finally, the obtained data change trajectory is traversed again using the running state features to determine whether it conforms to the running state features overall. If it conforms to the features, the data change trajectory is likely to appear and has a high probability of being reliable. The specific operation process is as follows: traverse each data change trajectory based on the running state features, determine the matching degree in real time, and determine the reliability probability of the data change trajectory based on the matching degree at different times. For example, the mean of the matching degree can be calculated, and the reliability probability can be determined based on the direct proportion of the mean.

[0096] As a preferred embodiment of the technical solution of the present invention, the step of identifying data change divergence nodes as feature data nodes by comparing the data change trajectories based on the confidence probability of the timing includes:

[0097] At preset time intervals, extract the state nodes corresponding to each data change trajectory and construct the data state vector corresponding to the state node.

[0098] The data change trajectory with the highest confidence probability is selected as the baseline trajectory;

[0099] Calculate the state difference values ​​between the data state vectors at each position of different data change trajectories and the baseline trajectory to obtain a sequence of difference values ​​for each data change trajectory relative to the baseline trajectory;

[0100] Based on a preset trajectory order, all difference value sequences are statistically analyzed to construct a difference matrix; one dimension of the difference matrix represents the trajectory position, which corresponds to time information, and the other dimension represents the trajectory type.

[0101] Based on the location, an array is extracted from the difference matrix, and the elements in the array are accumulated to obtain the overall difference value at that location.

[0102] Select the location where the overall difference value reaches a preset threshold, and use the time point of the query location relative to the current time as the feature data node.

[0103] In one example of the technical solution of this invention, the comparison process of data trajectories is described to identify divergences in data changes as feature data nodes. Specifically, a feature data node generation process is performed every preset time period to determine the amount of resources to be collected in the next period. State nodes corresponding to each data change trajectory are extracted, and a data state vector corresponding to the state node is constructed. The state node is each element in the data change trajectory. Each state node corresponds to multiple devices. Therefore, it is a vector, which is the data state vector mentioned above. It represents the internal state of the hydropower station at a future moment.

[0104] Then, the data change trajectory with the highest confidence probability is selected, which is the most likely state to occur in the future, and is used as the baseline trajectory. The state difference value between the data state vectors at each position of different data change trajectories and the baseline trajectory is calculated. On the basis that the dimensions of the data change trajectories are the same, this process is a simple comparison process. The state difference value is actually the vector distance of the data state vectors, and thus a sequence of difference values ​​of each data change trajectory relative to the baseline trajectory is obtained.

[0105] Then, according to the preset trajectory order, all difference value sequences are statistically analyzed to construct a difference matrix. The difference value sequence is an array, and statistical analysis of it yields a matrix called the difference matrix. One dimension of the difference matrix represents the trajectory position, which corresponds to time information, and the other dimension represents the trajectory type. Based on the position, an array is extracted from the difference matrix, and the elements in the array are accumulated to obtain the overall difference value at that position. This process is also easy to understand. When the number of columns represents the position (time), each row represents the difference value between a certain trajectory and the baseline trajectory at different times. At this time, each column of data is the array that needs to be extracted in the above content, representing the difference value between different trajectories and the baseline trajectory at each position. These are accumulated to obtain the overall difference value. The position where the overall difference value reaches a preset threshold (large enough) is selected, and the time point of the position relative to the current time is queried as the feature data node. It can be imagined that the feature data node is not unique. The obtained feature data node is actually an array composed of time labels, and the time labels use absolute time.

[0106] Regarding step S300, the step of comparing the characteristic data nodes of different hydropower stations to determine the environmental anomaly degree of each hydropower station includes:

[0107] By statistically analyzing the characteristic data nodes of different hydropower stations in chronological order, a time point sequence is obtained.

[0108] For any hydropower station, compare the time sequence of this hydropower station with the time sequence of other hydropower stations, and determine the time difference of each number in sequence to obtain the time difference sequence; wherein, to determine the time difference of each number, the starting point alignment process is executed first.

[0109] Calculate the mean sequence of the time difference series, and determine the total time difference based on the mean sequence;

[0110] The total time difference of all hydropower stations is calculated to determine the degree of environmental anomaly.

[0111] In one example of the technical solution of this invention, the application process of feature data nodes is described. Feature data nodes of different hydropower stations are statistically analyzed according to time sequence to obtain a time point sequence. For any hydropower station, the time point sequence of that hydropower station is compared with the time point sequences of other hydropower stations, and the time difference of each sequence number is determined sequentially to obtain a time difference sequence. It should be noted that the time point sequences are no longer of the same dimension at this point, and the comparison method is slightly different. The shorter time point sequence needs to be expanded first, continuously copying the last time point to make them of the same dimension. Then, the elements are compared sequentially. For any two time point sequences, the time difference of the first element is calculated first. When calculating the time difference of the second element, the difference between the time difference of the second element and the time difference of the first element needs to be calculated. This may have some computational complexity. In fact, it is about determining the actual time difference generated after removing the existing time difference. It is worth mentioning that if the two time point sequences were originally of different dimensions, the time difference calculated from the last few elements will be very large.

[0112] The above scheme is actually just one feasible calculation method. In fact, any parameter that can represent the differences in time series can be used. After calculating the time difference, a time difference sequence is obtained. At this time, there is a time difference sequence for each hydropower station. All time difference sequences are aligned (upsampled), and then the mean sequence of the time difference sequences is calculated. The values ​​in the mean sequence are superimposed to obtain the total time difference value.

[0113] Finally, the total time difference of all hydropower stations is calculated. Each total time difference is normalized based on its maximum and minimum values. Then, the environmental anomaly level is determined based on the normalized values, with the highest set at level five and the lowest at level one. This process actually eliminates a large number of macroscopic influences, such as the conditions under different weather conditions, because all hydropower stations are in the same weather. In this way, what is obtained is the anomaly of a hydropower station relative to other hydropower stations, rather than the anomaly relative to the weather. It can be combined with existing anomalies relative to the weather to obtain a more accurate judgment result.

[0114] It is worth mentioning that the calculation process of time difference is actually very complicated because the time difference itself can be positive or negative. To address this, when calculating the time difference, the absolute value is calculated directly. When calculating the next time difference, the previous moment is aligned first, and then the absolute value of the time difference is calculated. This method is clearer. Of course, the specific calculation process is a conventional technique for those skilled in the art, and will not be elaborated here.

[0115] Figure 2 The block diagram of the data acquisition system based on a hydropower station provided in the embodiments of the present invention is shown. As a preferred embodiment of the technical solution of the present invention, a data acquisition system 10 based on a hydropower station is provided, the data acquisition system 10 based on the hydropower station includes:

[0116] The area segmentation module 11 is used to obtain an area map of the water area to be sampled, segment the area map based on the registered hydropower station locations, and determine the sampling area of ​​each hydropower station.

[0117] The feature node generation module 12 is used to acquire the operation data of each hydropower station in real time and determine the feature data nodes of each hydropower station at regular intervals based on the operation data. The generation process of the feature data nodes is as follows: generating prediction data based on the operation data of the hydropower station, and determining the feature data nodes based on the prediction data. The feature data nodes are time series composed of relative time points.

[0118] The environmental assessment module 13 is used to compare the characteristic data nodes of different hydropower stations to determine the environmental anomaly degree of each hydropower station.

[0119] The data acquisition and adjustment module 14 is used to determine the amount of data acquisition resources based on the environmental anomaly level; the amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency.

[0120] Furthermore, the feature node generation module 12 includes:

[0121] The feature extraction unit is used to acquire hydropower station operation data in real time, obtain operation data sequences, extract features from the operation data sequences, and obtain operation status features.

[0122] The trajectory matching unit is used to match and generate data change trajectories in a preset database based on the characteristics of the running status, and to determine the confidence probability of each data change trajectory.

[0123] The trajectory comparison unit is used to periodically compare the trajectory of each data change based on the confidence probability, identify the divergence nodes of data change, and use them as feature data nodes.

[0124] Specifically, the feature extraction unit includes:

[0125] The data acquisition subunit is used to acquire monitoring data corresponding to the turbine, generator, transformer and auxiliary equipment;

[0126] The sorting subunit is used to sort the monitoring data according to the collection time to generate time series data.

[0127] The change analysis subunit is used to extract the magnitude and direction of change of each monitoring parameter between continuous acquisition times, and generate parameter change characteristics;

[0128] The distance calculation subunit is used to calculate the data change distance between monitoring data corresponding to adjacent acquisition times and generate a change distance sequence.

[0129] The feature fusion subunit is used to fuse parameter change features with change distance sequences to generate operating status features.

[0130] Furthermore, the trajectory matching unit includes:

[0131] The reference matching subunit is used to retrieve a reference state from historical running data that matches the current state based on the running state characteristics.

[0132] The historical query subunit is used to extract the historical state corresponding to the subsequent time of any reference state based on a preset time step.

[0133] The trajectory construction sub-unit is used to sort historical states according to time order and construct data change trajectories;

[0134] The traversal sub-unit is used to traverse each data change trajectory based on the running state characteristics, determine the matching degree in real time, and determine the reliability probability of the data change trajectory based on the matching degree at different times.

[0135] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data acquisition method based on a hydropower station, characterized in that, The method includes: Obtain a regional map of the water area to be sampled, and divide the regional map based on the registered locations of hydropower stations to determine the sampling areas for each hydropower station; Real-time acquisition of operational data from each hydropower station, and periodic determination of characteristic data nodes for each hydropower station based on the operational data; the generation process of the characteristic data nodes is as follows: generating prediction data based on the operational data of the hydropower station, and determining the characteristic data nodes based on the prediction data, wherein the characteristic data nodes are time series composed of relative time points; By comparing the characteristic data nodes of different hydropower stations, the environmental anomaly degree of each hydropower station can be determined. The amount of data acquisition resources is determined based on the environmental anomaly level; the amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency.

2. The data acquisition method based on a hydropower station according to claim 1, characterized in that, The steps of acquiring real-time operational data of each hydropower station and periodically determining the characteristic data nodes of each hydropower station based on the operational data include: Real-time acquisition of hydropower station operation data yields an operation data sequence, and feature extraction is performed on the operation data sequence to obtain operation status features; Based on the characteristics of the operating status, the system matches and generates data change trajectories in a preset database, and determines the confidence probability of each data change trajectory. Based on the reliable probability comparison of the change trajectory of each data point at regular intervals, the divergence nodes of data change are identified and used as feature data nodes.

3. The data acquisition method based on a hydropower station according to claim 2, characterized in that, The steps of acquiring real-time hydropower station operation data, obtaining an operation data sequence, and extracting features from the operation data sequence to obtain operation status features include: Acquire monitoring data for water turbines, generators, transformers, and auxiliary equipment; The monitoring data is sorted into time series according to the collection time to generate time series data; Extract the amplitude and direction of change of each monitoring parameter between continuous acquisition times to generate parameter change characteristics; Calculate the data change distance between monitoring data corresponding to adjacent acquisition times, and generate a change distance sequence; The parameter change characteristics are fused with the change distance sequence to generate the operating status characteristics.

4. The data acquisition method based on a hydropower station according to claim 2, characterized in that, The steps of matching and generating data change trajectories in a preset database based on operational status characteristics, and determining the confidence probability of each data change trajectory, include: Based on the characteristics of the operating status, retrieve the reference status from the historical operating data that matches the current status; For any reference state, extract the historical state corresponding to the subsequent time step of the reference state based on a preset time step; Sort historical states in chronological order to construct data change trajectories; Based on the characteristics of the running status, each data change trajectory is traversed, the matching degree is determined in real time, and the reliability probability of the data change trajectory is determined according to the matching degree at different times.

5. The data acquisition method based on a hydropower station according to claim 4, characterized in that, The step of identifying divergence nodes in data changes based on a reliable probability comparison of data change trajectories, and using these nodes as feature data nodes, includes: At preset time intervals, extract the state nodes corresponding to each data change trajectory and construct the data state vector corresponding to the state node. The data change trajectory with the highest confidence probability is selected as the baseline trajectory; Calculate the state difference values ​​between the data state vectors at each position of different data change trajectories and the baseline trajectory to obtain a sequence of difference values ​​for each data change trajectory relative to the baseline trajectory; Based on a preset trajectory order, all difference value sequences are statistically analyzed to construct a difference matrix; one dimension of the difference matrix represents the trajectory position, which corresponds to time information, and the other dimension represents the trajectory type. Based on the location, an array is extracted from the difference matrix, and the elements in the array are accumulated to obtain the overall difference value at that location. Select the location where the overall difference value reaches a preset threshold, and use the time point of the query location relative to the current time as the feature data node.

6. The data acquisition method based on a hydropower station according to claim 1, characterized in that, The step of comparing characteristic data nodes of different hydropower stations to determine the environmental anomaly degree of each hydropower station includes: By statistically analyzing the characteristic data nodes of different hydropower stations in chronological order, a time point sequence is obtained. For any hydropower station, compare the time sequence of this hydropower station with the time sequence of other hydropower stations, and determine the time difference of each number in sequence to obtain the time difference sequence; wherein, to determine the time difference of each number, the starting point alignment process is executed first. Calculate the mean sequence of the time difference series, and determine the total time difference based on the mean sequence; The total time difference of all hydropower stations is calculated to determine the degree of environmental anomaly.

7. A data acquisition system based on a hydropower station, characterized in that, The system includes: The area segmentation module is used to obtain an area map of the water area to be sampled, and to segment the area map based on the registered locations of hydropower stations to determine the sampling area for each hydropower station. The feature node generation module is used to acquire the operation data of each hydropower station in real time and determine the feature data nodes of each hydropower station periodically based on the operation data. The generation process of the feature data nodes is as follows: generate prediction data based on the operation data of the hydropower station, and determine the feature data nodes based on the prediction data. The feature data nodes are time series composed of relative time points. The environmental assessment module is used to compare the characteristic data nodes of different hydropower stations to determine the degree of environmental anomaly of each hydropower station. The data acquisition and adjustment module is used to determine the amount of data acquisition resources based on the environmental anomaly level; the amount of data acquisition resources is used to adjust the amount of data acquisition within the acquisition area of ​​each hydropower station, including but not limited to the acquisition range, the number of acquisition types, and the acquisition frequency.

8. The data acquisition system based on a hydropower station according to claim 7, characterized in that, The feature node generation module includes: The feature extraction unit is used to acquire hydropower station operation data in real time, obtain operation data sequences, extract features from the operation data sequences, and obtain operation status features. The trajectory matching unit is used to match and generate data change trajectories in a preset database based on the characteristics of the running status, and to determine the confidence probability of each data change trajectory. The trajectory comparison unit is used to periodically compare the trajectory of each data change based on the confidence probability, identify the divergence nodes of data change, and use them as feature data nodes.

9. The data acquisition system based on a hydropower station according to claim 8, characterized in that, The feature extraction unit includes: The data acquisition subunit is used to acquire monitoring data corresponding to the turbine, generator, transformer and auxiliary equipment; The sorting subunit is used to sort the monitoring data according to the collection time to generate time series data. The change analysis subunit is used to extract the magnitude and direction of change of each monitoring parameter between continuous acquisition times, and generate parameter change characteristics; The distance calculation subunit is used to calculate the data change distance between monitoring data corresponding to adjacent acquisition times and generate a change distance sequence. The feature fusion subunit is used to fuse parameter change features with change distance sequences to generate operating status features.

10. The data acquisition system based on a hydropower station according to claim 8, characterized in that, The trajectory matching unit includes: The reference matching subunit is used to retrieve a reference state from historical running data that matches the current state based on the running state characteristics. The historical query subunit is used to extract the historical state corresponding to the subsequent time of any reference state based on a preset time step. The trajectory construction sub-unit is used to sort historical states according to time order and construct data change trajectories; The traversal sub-unit is used to traverse each data change trajectory based on the running state characteristics, determine the matching degree in real time, and determine the reliability probability of the data change trajectory based on the matching degree at different times.

Images