Basin collaborative prediction optimization system and method of water conservancy big data

Abstract

The application discloses a watershed collaborative forecasting optimization system and method for water conservancy big data, and belongs to the technical field of water conservancy forecasting. The method sets up a watershed monitoring station and deploys an Internet of Things sensor, divides a monitoring area and synchronizes a time stamp, binds a station identifier to generate a hydrological characteristic flow, constructs a full-watershed data matrix, calculates a state difference update matrix element, and finally generates a downstream station collaborative forecasting matrix to determine a flood forecasting range. The system comprises a monitoring deployment and time synchronization unit, an identifier binding and characteristic flow generation unit, a data capture and matrix construction unit, a difference calculation and matrix update unit, and a collaborative forecasting and alarm generation unit. The application can enhance the correlation of collaborative forecasting and the consistency of actual flood propagation paths, improve the accuracy and pertinence of the boundary effect of the alarm range, and shorten the response time from data collection to alarm generation while ensuring the spatiotemporal consistency of data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water conservancy forecasting, in particular to a basin collaborative forecasting optimization system and method for water conservancy big data. BACKGROUND

[0002] In the field of water conservancy forecasting, accurate and timely flood forecasting is crucial for disaster prevention and mitigation. However, traditional water conservancy forecasting techniques have significant limitations in data collection, processing, and collaborative analysis:

[0003] Data collection and synchronization problems: the existing monitoring site deployment lacks systematicness, and the time stamp of the sensor is not synchronized, leading to spatio-temporal dislocation of hydrological data, making it difficult to reflect the real evolution process of the basin hydrological characteristics. For example, different site sensors may record different stage data of the same flood event due to time deviation, causing analysis results to be distorted.

[0004] Insufficient basin collaborative analysis: traditional methods are mostly based on independent analysis of single site or local area data, lacking dynamic modeling and collaborative correlation of full basin hydrological data. For example, only the water level change of individual sites is used to predict flood propagation, without considering the hydrological conduction relationship between upstream and downstream areas, resulting in inaccurate prediction range.

[0005] Data processing method lags behind: existing technologies for processing multi-source hydrological data (such as secondary flood rating influence parameters and water level values) are limited to simple statistical level, lacking depth mining of data spatio-temporal dimensions. For example, a full basin data matrix is not constructed to capture the double state difference of monitoring period and area replacement, making it difficult to quantify the dominant factors of basin hydrological changes.

[0006] Low efficiency of alarm response: traditional forecasting models rely on manual alarm range setting, lack of dynamic optimization mechanism, resulting in fuzzy alarm boundary, much redundant information, and long response time. For example, it is difficult to quickly identify the key path of flood propagation, delaying emergency decision-making.

[0007] To solve the above problems, the present application proposes a basin collaborative forecasting optimization system and method for water conservancy big data, which realizes efficient integration and dynamic feedback of basin hydrological data through synchronization of Internet of Things sensor time stamp, construction of full basin data matrix, state difference processing, and collaborative forecasting analysis, which is conducive to improving the accuracy and timeliness of flood forecasting, and recommending auxiliary solutions for water conservancy disaster warning. SUMMARY

[0008] The present application aims to provide a basin collaborative forecasting optimization system and method for water conservancy big data to solve the problems raised in the background.

[0009] To solve the above technical problems, the present application provides the following technical solutions:

[0010] The water conservancy big data basin collaborative prediction optimization system comprises a monitoring deployment and time synchronization unit, an identification binding and feature flow generation unit, a data capture and matrix construction unit, a difference calculation and matrix update unit and a collaborative prediction and alarm generation unit.

[0011] The monitoring deployment and time synchronization unit is configured to set up basin monitoring sites and deploy Internet of Things sensors, divide monitoring areas within the coverage of the basin monitoring sites, so that each Internet of Things sensor collects hydrological data in the same monitoring area, and trigger an initialization instruction through an upstream starting basin monitoring site to realize the synchronous calibration of the time stamp of each Internet of Things sensor.

[0012] The identification binding and feature flow generation unit is configured to bind a unique identification to each basin monitoring site and write site information, and configure an initialization synchronous time stamp. When a flood event occurs in a monitoring area, the identification binding and feature flow generation unit triggers the Internet of Things sensors in the monitoring area to collect data and generates a hydrological feature flow with a synchronous time stamp.

[0013] The data capture and matrix construction unit captures the hydrological data of all Internet of Things sensors in the flood event based on the hydrological feature flow and the monitoring area, and establishes a full-basin data matrix with the hydrological feature flow as the horizontal dimension and the monitoring area as the vertical dimension.

[0014] The difference calculation and matrix update unit calculates the state difference in the horizontal and vertical directions with the full-basin data matrix elements as the center, and updates the matrix elements according to the maximum difference.

[0015] The collaborative prediction and alarm generation unit removes the first row and the first column of the full-basin data matrix to obtain a downstream site collaborative prediction matrix, calculates the correlation degree of the collaborative prediction between sites to determine the prediction range of the flood event, selects the largest continuous range segment after removing the duplicate monitoring sites in the prediction range to form an alarm range, and sends the alarm range to a staff terminal.

[0016] Further, the identification binding and feature flow generation unit comprises a regional device group module and an initialization time stamp module.

[0017] The regional device group module is configured to count all Internet of Things sensors deployed in different monitoring areas to form a regional device group.

[0018] The initialization time stamp module is configured to configure an initialization synchronous time stamp.

[0019] Further, the data capture and matrix construction unit comprises a monitoring period generation module and a monitoring feature label generation module.

[0020] The monitoring period generation module is configured to divide the monitoring period of different monitoring areas under the hydrological feature flow.

[0021] The monitoring feature label generation module is configured to capture hydrological data of all Internet of Things sensors generated under a monitoring flood event to generate a monitoring feature label.

[0022] Further, the difference calculation and matrix updating unit comprises a state replacement module and a matrix updating module.

[0023] The state replacement module is configured to replace a lateral state and a longitudinal state, wherein the lateral state refers to a state of a monitoring area itself when a monitoring period is replaced, and the longitudinal state refers to a double state between two monitoring areas when the monitoring area is replaced downstream.

[0024] The matrix updating module is configured to perform gradient updating of a full-basin data matrix downstream.

[0025] The basin collaborative prediction optimization method of water conservancy big data comprises the following steps:

[0026] Step S1: Setting up basin monitoring sites and deploying Internet of Things sensors in the basin, dividing monitoring areas of each basin monitoring site, and counting Internet of Things sensors in the monitoring areas to collect hydrological data containing secondary flood rating influence parameters and water level values. Meanwhile, based on the upstream and downstream relationship of basin runoff, an initialization instruction is triggered through an upstream starting basin monitoring site to realize synchronization calibration of the time stamps of each Internet of Things sensor.

[0027] Step S2: Binding a unique identifier to each basin monitoring site and writing site information, and configuring an initialization synchronization time stamp. When a flood event occurs in a monitoring area, the Internet of Things sensors in the monitoring area are triggered to collect data, and a hydrological feature stream with a synchronization time stamp is generated.

[0028] Step S3: Based on the hydrological feature stream and the monitoring area, the hydrological data of all Internet of Things sensors in the flood event is captured to generate a monitoring feature label containing a monitoring period, an influence parameter cluster, and a water level change value. A full-basin data matrix is established with the hydrological feature stream as the lateral dimension and the monitoring area as the longitudinal dimension.

[0029] Step S4: Taking the elements of the full-basin data matrix as the center, the state differences in the lateral direction (the state of the area itself when the monitoring period is replaced) and the longitudinal direction (the double state between the two areas when the area is replaced downstream) are calculated, and the matrix elements are updated according to the maximum difference.

[0030] Step S5: Removing the first row and the first column of the full-basin data matrix to obtain a downstream site collaborative prediction matrix, calculating the collaborative prediction correlation between sites to determine the prediction range of the flood event, and selecting the largest continuous range segment after removing the duplicate monitoring sites in the prediction range to form an alarm range and send it to the staff port.

[0031] Further, the specific implementation process of step S1 comprises:

[0032] Set the basin monitoring station, deploy the Internet of Things sensor at each basin monitoring station, divide the monitoring area for the coverage range of the basin monitoring station, and count the Internet of Things sensors deployed inside each monitoring area, the Internet of Things sensors are used to collect hydrological data in the basin and add a synchronous timestamp, the hydrological data includes the influence parameter of the secondary flood rating and the water level value in the monitoring area;

[0033] Based on the upstream and downstream relationship of the basin runoff, the time stamp of each Internet of Things sensor is calibrated, the initialization instruction of the synchronization calibration is set, and the initialization instruction is triggered by the starting basin monitoring station upstream of the basin runoff and distributed to each basin monitoring station downstream, to instruct the time stamp synchronization calibration of each Internet of Things sensor;

[0034] In the above method, by setting the monitoring station in the basin and deploying the Internet of Things sensor, the monitoring area is divided to collect hydrological data (secondary flood rating influence parameter, water level value); based on the upstream and downstream relationship of the basin runoff, the initialization instruction is triggered by the upstream starting station, and the time stamp synchronization calibration mechanism is used to ensure the time consistency of the sensor data in the whole basin; further, the data misplacement problem caused by the time difference of the sensor can be eliminated to ensure the space-time consistency of the hydrological data, and lay a foundation for subsequent whole basin data modeling.

[0035] Further, the specific implementation process of step S2 includes:

[0036] Bind a unique identifier to each basin monitoring station and write the station information, the station information includes:

[0037] Any mth basin monitoring station is recorded as , the flood event with flood number n is recorded as , any e th monitoring area is recorded as , any r th Internet of Things sensor is recorded as , the monitoring area inside the deployment is counted, and the regional device group is formed, R represents the total number of Internet of Things sensors;

[0038] The initialized synchronous timestamp is recorded as , The g th synchronization time node in the initialized synchronous timestamp is recorded as G, which represents the total number of synchronization time nodes in the initialized synchronous timestamp; when the flood event occurs in the monitoring area , trigger the Internet of Things sensors in the regional device group to collect hydrological data, and generate a hydrological feature stream with a synchronous timestamp , and q represents the number of triggering times;

[0039] In the above method, a unique identifier is bound for each monitoring site, a site information model containing the flood number, monitoring area and sensor is constructed; when the flood event is triggered, a hydrological feature flow with a time label is generated by initializing a synchronization timestamp to realize accurate triggering and structured storage of data acquisition; further, one-to-one correspondence between the flood event and the monitoring data can be ensured, which facilitates subsequent data tracing and feature analysis to improve data management efficiency.

[0040] Further, the specific implementation process of step S3 includes:

[0041] Based on the hydrological feature flow and the monitoring area, all hydrological data of the Internet of Things sensors generated during the monitoring of the flood event is captured to generate a monitoring feature label, denoted as , wherein represents the hydrological feature flow of the monitoring area in the monitoring period, represents the influence parameter cluster of the monitoring area in the monitoring period , and represents the water level change value in the monitoring area in the monitoring period , which is the difference between the water level value in the monitoring area at the starting synchronization time node in the monitoring period and the water level value in the monitoring area at the ending synchronization time node in the monitoring period. If the water level change value is positive, it indicates that the water level in the monitoring area decreases, and if the water level change value is negative, it indicates that the water level in the monitoring area increases.

[0042] A full-basin data matrix is established with the hydrological feature flow as the horizontal dimension and the monitoring area as the vertical dimension. The full-basin data matrix generated during the monitoring of the flood event is denoted as , and the matrix element corresponding to the qth row and the e th column in the full-basin data matrix is denoted as .

[0043] In the above method, based on the hydrological feature flow and the monitoring area, all basin sensor data is captured to generate a monitoring feature label containing the monitoring period, the influence parameter cluster, and the water level change value; a full-basin data matrix is constructed with the hydrological feature flow as the horizontal dimension and the monitoring area as the vertical dimension to realize spatial gridding and time sequencing of multi-source data; further, discrete hydrological data can be converted into a structured matrix, which facilitates matrix operation to mine the spatiotemporal rules of basin hydrological changes and provides a data basis for collaborative analysis.

[0044] Furthermore, the specific implementation process of step S4 includes:

[0045] Based on the data matrix of the entire basin, select the matrix elements is the center of the matrix, calculate the matrix elements The horizontal state difference and the vertical state difference of the monitoring period Monitoring area during replacement The longitudinal state difference refers to the monitoring area When the monitoring area is replaced downstream, the monitoring area The dual status difference between the downstream monitoring area and the monitoring period refers to the two dimensions of monitoring area:

[0046] ;

[0047] Where, Indicates the horizontal state difference, represents the longitudinal state difference, Indicates monitoring period Change to monitoring period Time monitoring area The difference of its own state, and the difference value of its own state is the influencing parameter cluster of the secondary flood rate and The number of identical influencing parameters contained in the intersection set;

[0048] Select the maximum value between the horizontal state difference and the vertical state difference ,like , then for the matrix elements Update and let the matrix elements Is 0, if , then for the matrix elements Update and let the matrix elements is 1;

[0049] In the above method, horizontal state differences (changes in the state of the region itself when the monitoring period changes) and vertical state differences (double state changes when the monitoring area changes downstream) are defined, and the matrix values ​​(0 or 1) are updated by calculating the differences in matrix elements to characterize the dominant factors of the basin state changes (period changes or regional evolution); furthermore, through dynamic matrix updates, the key nodes and propagation directions of the basin's hydrological changes are highlighted, secondary information is filtered out, and the data processing efficiency and the sensitivity of the forecast model are improved.

[0050] Furthermore, the specific implementation process of step S5 includes:

[0051] Basin-wide data matrix After each matrix element is updated, the full-basin data matrix is removed The matrix element in the first row and the first column is obtained, and a downstream station collaborative prediction matrix is obtained, denoted as Based on the downstream station collaborative prediction matrix, the collaborative prediction correlation between the monitoring stations is calculated, and the prediction range of the flood event is optimized:

[0052] ;

[0053] In the formula, represents the collaborative prediction correlation between the monitoring stations under the flood event and the monitoring stations , the downstream station collaborative prediction matrix represents the number of values 1 contained after the Boolean logical AND operation between the downstream station collaborative prediction matrix and the downstream station collaborative prediction matrix , the downstream station collaborative prediction matrix represents the number of values 1 contained after the Boolean logical OR operation between the downstream station collaborative prediction matrix and the downstream station collaborative prediction matrix ;

[0054] A preset collaborative prediction threshold is set, and if the collaborative prediction correlation is greater than or equal to the collaborative prediction threshold, the monitoring stations and the monitoring stations are recorded in the prediction range of the flood event In the prediction range , the monitoring stations are subjected to a deduplication process, and after the deduplication process, the monitoring stations in the largest continuous range segment are selected to form an alarm range and sent to a staff terminal.

[0055] In the above method, the generation of the downstream station collaborative prediction matrix is based only on the downstream objects of the monitoring area, and when studying the differences between the downstream objects, the time dimension is also considered. The first row and the first column of the full-basin matrix are removed to obtain the downstream station collaborative prediction matrix. By calculating the collaborative prediction correlation between the stations (based on Boolean logical AND and OR operations), the prediction range is determined in combination with a preset threshold, and after deduplication, the largest continuous range segment is selected to generate an alarm. Furthermore, the hydrological correlation between the stations can be quantified, the prediction range can be dynamically optimized, redundant alarms can be avoided, and the accuracy and pertinence of the boundary effect of the alarm can be improved.

[0056] ​​Compared with the prior art, the beneficial effects achieved by the application are that: in the basin collaborative prediction optimization system and method of water conservancy big data provided by the application, the basin monitoring station is set and the Internet of Things sensor is deployed, the monitoring area is divided and the time stamp is synchronized, the station identifier is bound to generate the hydrological feature flow, the full-basin data matrix is constructed, the state difference update matrix element is calculated, and finally the downstream station collaborative prediction matrix is generated to determine the flood prediction range. The system includes a monitoring deployment and time synchronization unit, an identifier binding and feature flow generation unit, a data capture and matrix construction unit, a difference calculation and matrix update unit, and a collaborative prediction and alarm generation unit. While ensuring the spatiotemporal consistency of data, the application can enhance the relevance of collaborative prediction and the consistency of the actual flood propagation path, improve the accuracy and pertinence of the boundary effect of the alarm range, and shorten the response time from data acquisition to alarm generation. BRIEF DESCRIPTION OF DRAWINGS

[0057] The accompanying drawings are included to provide a further understanding of the application, and constitute a part of the specification, illustrate the application, and are used to explain the application together with the embodiments of the application, and do not constitute a limitation on the application.

[0058] Figure 1 The step schematic diagram of the basin collaborative prediction optimization method of water conservancy big data. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0060] In the first embodiment: a basin collaborative prediction optimization system of water conservancy big data is provided, which includes: a monitoring deployment and time synchronization unit, an identifier binding and feature flow generation unit, a data capture and matrix construction unit, a difference calculation and matrix update unit, and a collaborative prediction and alarm generation unit;

[0061] The monitoring deployment and time synchronization unit is used for setting the basin monitoring station and deploying the Internet of Things sensor, dividing the monitoring area within the coverage range of the basin monitoring station, so that each Internet of Things sensor collects hydrological data in the same monitoring area, and is also used for triggering an initialization instruction through the upstream starting basin monitoring station to realize the synchronization calibration of the time stamp of each Internet of Things sensor;

[0062] An identification binding and feature flow generating unit is configured to bind a unique identification to each flow basin monitoring site and write site information, and to configure an initial synchronization timestamp. When a flood event occurs in a monitoring area, the identification binding and feature flow generating unit triggers Internet of Things sensors in the monitoring area to collect data and generate a hydrological feature flow with a synchronization timestamp.

[0063] The identification binding and feature flow generating unit includes a regional device group module and an initial timestamp module.

[0064] The regional device group module is configured to count all Internet of Things sensors deployed in different monitoring areas to form a regional device group.

[0065] The initial timestamp module is configured to configure an initial synchronization timestamp.

[0066] A data capturing and matrix constructing unit is configured to capture hydrological data of all Internet of Things sensors in a flood event based on a hydrological feature flow and a monitoring area, and to establish a full-basin data matrix with the hydrological feature flow as a horizontal dimension and the monitoring area as a vertical dimension.

[0067] The data capturing and matrix constructing unit includes a monitoring period generating module and a monitoring feature label generating module.

[0068] The monitoring period generating module is configured to divide monitoring periods of different monitoring areas under the hydrological feature flow.

[0069] The monitoring feature label generating module is configured to capture hydrological data of all Internet of Things sensors generated under a monitoring flood event to generate a monitoring feature label.

[0070] A difference calculation and matrix updating unit is configured to calculate horizontal and vertical state differences with full-basin data matrix elements as the center, and to update matrix elements according to maximum differences.

[0071] The difference calculation and matrix updating unit includes a state replacement module and a matrix updating module.

[0072] The state replacement module is configured to replace a horizontal state and a vertical state. The horizontal state refers to a state of a monitoring area itself when a monitoring period is replaced, and the vertical state refers to double states between two monitoring areas when the monitoring area is replaced to a downstream.

[0073] The matrix updating module is configured to perform a gradient updating of a full-basin data matrix to a downstream.

[0074] A collaborative prediction and alarm generating unit is configured to remove a first row and a first column of the full-basin data matrix to obtain a downstream site collaborative prediction matrix, to calculate a collaborative prediction correlation between sites to determine a prediction range of a flood event, and to select a maximum continuous range segment after removing monitoring sites in the prediction range to form an alarm range and send the alarm range to a staff terminal.

[0075] Referring to Figure 1 In the second embodiment, a watershed collaborative forecasting optimization method for water conservancy big data is provided, which is applicable to the first embodiment. The method comprises the following steps:

[0076] Parameter setting:

[0077] Watershed area: 1000 , set 5 monitoring sites , divide each site into 2 monitoring areas , deploy a total of Internet of Things sensors;

[0078] Initialize the synchronization timestamp node (time interval 1 hour), set the collaborative forecasting threshold to 0.7;

[0079] The secondary flood rating influence parameters include rainfall, runoff coefficient, and soil moisture.

[0080] Implementation steps:

[0081] Step S1: Set up watershed monitoring sites and deploy Internet of Things sensors in the watershed, divide the monitoring areas of each watershed monitoring site, count the Internet of Things sensors in the monitoring areas to collect hydrological data including secondary flood rating influence parameters and water level values, and based on the upstream and downstream relationship of watershed runoff, trigger the initialization instruction through the upstream starting watershed monitoring site to realize the synchronization calibration of the timestamp of each Internet of Things sensor;

[0082] For example, set up watershed monitoring sites, deploy Internet of Things sensors at each watershed monitoring site, divide the monitoring areas of the watershed monitoring sites, and count the Internet of Things sensors deployed in each monitoring area. The Internet of Things sensors are used to collect hydrological data in the watershed and add a synchronization timestamp. The hydrological data includes secondary flood rating influence parameters and water level values in the monitoring area;

[0083] Based on the upstream and downstream relationship of watershed runoff, the synchronization of the timestamp of each Internet of Things sensor is calibrated, the initialization instruction for synchronization calibration is set, and the initialization instruction is triggered through the upstream starting watershed monitoring site and distributed to each downstream watershed monitoring site to synchronize the timestamp of each Internet of Things sensor;

[0084] For example: trigger the timestamp synchronization instruction at the upstream site of the watershed, and the downstream site receives the instruction to complete the sensor time calibration.

[0085] Step S2: Bind a unique identifier to each watershed monitoring site and write site information, configure an initial synchronization timestamp, and when a flood event occurs in the monitoring area, trigger the IoT sensors in the monitoring area to collect data and generate a hydrological characteristic flow with a synchronization timestamp;

[0086] For example, a unique identifier is bound to each watershed monitoring site and site information is written. The site information includes:

[0087] Any m-th watershed monitoring station is denoted as , record the flood event with flood number n as , any e-th monitoring area is recorded as , any r-th IoT sensor is recorded as , then the statistical monitoring area All IoT sensors deployed within the area constitute the regional device group , R represents the total number of IoT sensors;

[0088] The synchronization timestamp of configuration initialization is , Indicates the gth synchronization time node in the initialized synchronization timestamp, G indicates the total number of synchronization time nodes in the initialized synchronization timestamp; when a flood event occurs in the monitoring area When the regional device group is triggered Each IoT sensor collects hydrological data and generates a hydrological feature stream with synchronized time stamps ,and , q represents the number of triggers;

[0089] For example: when monitoring area ( Flood signal at site When an event occurs, the regional device group is triggered 20 sensors collect data to generate hydrological characteristic flow (Contains 5 time nodes).

[0090] Step S3: Based on the hydrological characteristic flow and monitoring area, the hydrological data of all IoT sensors in the flood event are captured, and monitoring feature labels containing the monitoring period, influencing parameter clusters, and water level change values ​​are generated. The data matrix of the entire watershed is established with the hydrological characteristic flow as the horizontal dimension and the monitoring area as the vertical dimension;

[0091] For example, based on the hydrological characteristics of the flow and monitoring area, the monitoring of flood events is captured The hydrological data of all IoT sensors generated at the time of generation are used to generate monitoring feature labels, which are recorded as ,in, Represents hydrological characteristic flow Lower monitoring area monitoring period, representing the monitoring period monitoring area adopted in the monitoring period, monitoring period monitoring area water level change value in the monitoring period, the water level change value being the difference between the water level value in the monitoring area at the starting synchronous time node in the monitoring period and the water level value in the monitoring area at the ending synchronous time node in the monitoring period, if the water level change value is positive, it indicates that the water level in the monitoring area decreases, if the water level change value is negative, it indicates that the water level in the monitoring area increases;

[0092] establish a full-basin data matrix with the hydrological characteristic flow as the horizontal dimension and the monitoring area as the vertical dimension, and the full-basin data matrix is when monitoring the flood event , the full-basin data matrix generated is , the matrix element corresponding to the qth row and the e th column in the full-basin data matrix is denoted as ;

[0093] For example: capturing the full-basin data in the event, constructing a 5x2 full-basin matrix , the matrix elements including the monitoring period T, the parameter cluster U and the water level change value d.

[0094] Step S4: taking the full-basin data matrix element as the center, calculating the horizontal (the state difference of the area itself when the monitoring period is replaced) and vertical (the double state difference between the areas when the area is replaced downstream) state differences, and updating the matrix element according to the maximum difference;

[0095] For example, based on the full-basin data matrix, selecting the matrix element as the matrix center, calculating the horizontal state difference and the vertical state difference of the matrix element , the horizontal state difference referring to the state difference of the monitoring area itself when the monitoring period is replaced, and the vertical state difference referring to the double state difference between the monitoring area and the downstream monitoring area when the monitoring area is replaced downstream, the double referring to the two dimensions of the monitoring period and the monitoring area:

[0096] ;

[0097] In the formula, represents the horizontal state difference, represents the vertical state difference, represents the monitoring period Substitute to monitoring period Monitoring area The self-state difference, and the self-state difference value is the influence parameter cluster of the sub-flood rate And The number of the same influence parameters contained in the intersection set between

[0098] Select the maximum value in the lateral state difference and the longitudinal state difference If , update the matrix element and let the matrix element be 0, if , update the matrix element and let the matrix element be 1;

[0099] For example: taking the matrix element as the center, calculate the lateral difference (the parameter cluster change of the period ) and the longitudinal difference (the parameter cluster of the area and the water level change), update the matrix element to 1 (if ).

[0100] Step S5: remove the first row and the first column of the whole basin data matrix to obtain a downstream site cooperative prediction matrix, calculate the cooperative prediction correlation between sites to determine the prediction range of the flood event, select the maximum continuous range segment after removing the monitoring sites in the prediction range to form an alarm range and send it to the staff port;

[0101] After the update of each matrix element in the exemplary whole basin data matrix , remove the matrix elements in the first row and the first column of the whole basin data matrix to obtain a downstream site cooperative prediction matrix, denoted as ; based on the downstream site cooperative prediction matrix, calculate the cooperative prediction correlation between monitoring sites and optimize the prediction range of the flood event :

[0102] ;

[0103] In the formula, represents the cooperative prediction correlation between monitoring sites and monitoring sites under the flood event , and represents the downstream site cooperative prediction matrix and the downstream site cooperative prediction matrix the number of values 1 contained after Boolean logical and operation between indicates the downstream station cooperative prediction matrix and the downstream station cooperative prediction matrix the number of values 1 contained after Boolean logical or operation between

[0104] preset cooperative prediction threshold, if the cooperative prediction correlation is greater than or equal to the cooperative prediction threshold, the monitoring station is associated with the monitoring station is recorded as a flood event prediction range , the monitoring station in the prediction range is de-duplicated, and the monitoring station in the largest continuous range segment after de-duplication is selected to form an alarm range and sent to a staff port;

[0105] For example, the first row and the first column are removed to obtain the downstream matrix , the cooperative prediction correlation of the station is calculated (> threshold ), the station is recorded as a prediction range , and the continuous alarm range is determined to be after de-duplication and sent to a staff port.

[0106] It should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device.

[0107] Finally, it should be noted that: the above only describes the preferred embodiments of the present application, and is not used to limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, and those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A basin collaborative forecasting optimization method based on water conservancy big data, characterized by: The method comprises the following steps: Step S1: Set up watershed monitoring stations in the basin and deploy IoT sensors. Divide the monitoring area of ​​each watershed monitoring station. Count the IoT sensors in the monitoring area to collect hydrological data including secondary flood rate influencing parameters and water level values. At the same time, based on the upstream and downstream relationship of the basin runoff, trigger the initialization instruction through the upstream starting watershed monitoring station to achieve synchronous calibration of the timestamps of each IoT sensor. Step S2: Bind a unique identifier to each watershed monitoring site and write site information, configure an initial synchronization timestamp, and when a flood event occurs in the monitoring area, trigger the IoT sensors in the monitoring area to collect data and generate a hydrological characteristic flow with a synchronization timestamp; Step S3: Based on the hydrological characteristic flow and monitoring area, the hydrological data of all IoT sensors in the flood event are captured, and monitoring feature labels containing the monitoring period, influencing parameter clusters, and water level change values ​​are generated. The data matrix of the entire watershed is established with the hydrological characteristic flow as the horizontal dimension and the monitoring area as the vertical dimension; Step S4: Calculate the horizontal and vertical state differences with the digitized matrix element of the entire watershed as the center, and update the matrix element according to the maximum value of the difference; The horizontal direction indicates the state of the region itself when the monitoring period changes, and the vertical direction indicates the dual state between the two regions when the region changes downstream; Step S5: Remove the first row and first column of the whole basin data matrix to obtain the downstream site collaborative forecast matrix, calculate the inter-site collaborative forecast correlation to determine the forecast range of the flood event, remove duplicate monitoring sites within the forecast range, select the largest continuous range segment to form the alarm range and send it to the staff port; The specific implementation process of step S1 includes: Establishing watershed monitoring stations, deploying IoT sensors at each monitoring station, dividing the coverage area of ​​the monitoring stations into monitoring regions, and counting the IoT sensors deployed in each monitoring region. The IoT sensors are used to collect hydrological data within the watershed and attach synchronized timestamps. The hydrological data includes influencing parameters for secondary flood determination and water level values ​​within the monitoring region. Based on the upstream and downstream relationship of the watershed runoff, the timestamp synchronization of each IoT sensor is calibrated, and the initialization instruction of the synchronization calibration is set. The initialization instruction is triggered by the starting watershed monitoring station upstream of the watershed runoff and distributed to each watershed monitoring station downstream to instruct the synchronization calibration of the timestamps of each IoT sensor; The specific implementation process of step S2 includes: Bind a unique identifier to each watershed monitoring site and write site information, which includes: Let any m-th watershed monitoring station be denoted as W m , the flood event with flood number n is recorded as H n , any e-th monitoring area is recorded as Z e , denote any r-th IoT sensor as S r , then the statistical monitoring area Z e All IoT sensors deployed in the area constitute the regional device group O(Z e )={S r |r∈[1, R]}, R represents the total number of IoT sensors; The synchronization timestamp of configuration initialization is P = {t g |g∈[1,G]},t g Indicates the gth synchronization time node in the initialized synchronization timestamp, G indicates the total number of synchronization time nodes in the initialized synchronization timestamp; when a flood event occurs in the monitoring area Z e When triggering the regional device group O(Z e ) in which each IoT sensor collects hydrological data and generates a hydrological characteristic flow P with a synchronized time stamp. q , and P q =P, q represents the number of trigger times.

2. The basin collaborative forecasting optimization method based on water conservancy big data according to claim 1 is characterized in that: The specific implementation process of step S3 includes: Based on the hydrological characteristics of the flow and monitoring area, capture the flood events H n The hydrological data of all IoT sensors generated at the time of generation are used to generate monitoring feature labels, which are recorded as P q :Z e =[U[T(P q |Z e )],d[T(P q |Z e )]], where T(P q |Z e ) represents the hydrological characteristic flow P q Lower monitoring area Z e During the monitoring period, U[T(P q |Z e )] represents the monitoring period T(P q |Z e ) Monitoring area Z e The cluster of influencing parameters used in the secondary flood calibration, d[T(P q |Z e )]]Monitoring period T(P q |Z e ) Monitoring area Z e The water level change value within the monitoring period is the difference between the water level value in the monitoring area at the start synchronization time node within the monitoring period and the water level value in the monitoring area at the end synchronization time node within the monitoring period. If the water level change value is a positive value, it means that the water level in the monitoring area has dropped, and if the water level change value is a negative value, it means that the water level in the monitoring area has increased; Establish a basin-wide data matrix with hydrological characteristic flow as the horizontal dimension and monitoring area as the vertical dimension, and divide the basin monitoring stations W m In monitoring flood events H n The whole basin data matrix generated when m |H n ), then the whole basin data matrix V(W m |H n ) The matrix element corresponding to the qth row and the eth column is denoted as P q :Z e .

3. The basin collaborative forecasting optimization method based on water conservancy big data according to claim 1 is characterized in that: The specific implementation process of step S4 includes: Based on the data matrix of the entire basin, select the matrix element P q :Z e is the center of the matrix, calculate the matrix element P q :Z e The lateral state difference and longitudinal state difference of the monitoring period T(P q |Z e ) Monitoring area Z when changing e The longitudinal state difference refers to the monitoring area Z e When the monitoring area is replaced downstream, the monitoring area Z e The dual status difference between the downstream monitoring area and the monitoring period refers to the two dimensions of monitoring area: Where, K1(P q :Z e ) represents the lateral state difference, K2(P q :Z e ) represents the longitudinal state difference, NUM{U[T(P q |Z e )]→U[T(P q+1 |Z e )]} represents the monitoring period T(P q |Z e ) is replaced to the monitoring period T(P q+1 |Z e ) when monitoring area Z e The difference of its own state, and the difference of its own state value is the influence parameter cluster U[T(P q |Z e )] and U[T(P q+1 |Z e )] contains the number of identical influencing parameters in the intersection set; Select the maximum value MAX{K1(P q :Z e →P q+1 :Z e ), K2(P q :Z e →P q+1 :Z e+1 )}, if MAX{K1(P q :Z e →P q+1 :Z e ), K2(P q :Z e →P q+1 :Z e+1 )}=K1(P q :Z e →P q+1 :Z e ), then for the matrix element P q+1 :Z e+1 Update and let the matrix element P q+1 :Z e+1 is 0, if MAX{K1(P q :Z e →P q+1 :Z e ), K2(P q :Z e →P q+1 :Z e+1 )}=K2(P q :Z e →P q+1 :Z e+1 ), then for the matrix element P q+1 :Z e+1 Update and let the matrix element P q+1 :Z e+1 is 1.

4. The basin collaborative forecasting optimization method based on water conservancy big data according to claim 3 is characterized in that: The specific implementation process of step S5 includes: The whole basin data matrix V(W m |H n ) is updated, the whole basin digitization matrix V(W m |H n ) in the first row and first column, and obtain the downstream station collaborative forecast matrix, which is recorded as V -1 (W m |H n ); Based on the collaborative forecast matrix of downstream stations, calculate the correlation between the collaborative forecasts of monitoring stations and optimize the flood event H n Forecast range: Where, F(W m |H n →W m+1 |H n ) indicates that in flood event H n Lower monitoring station W m With monitoring site W m+1 The correlation degree of collaborative forecast between -1 (W m |H n )∩V -1 (W m+1 |H n )] represents the downstream site collaborative forecast matrix V -1 (W m |H n ) and the downstream station collaborative forecast matrix V -1 (W m+1 |H n ) after the Boolean logic operation, NUM[V -1 (W m |H n )∪V -1 (W m+1 |H n )] represents the downstream site collaborative forecast matrix V -1 (W m |H n ) and the downstream station collaborative forecast matrix V -1 (W m+1 |H n ) after the Boolean logical OR operation; Preset collaborative forecast threshold, if collaborative forecast correlation F(W m |H n →W m+1 |H n ) is greater than or equal to the collaborative forecast threshold, the monitoring station W m With monitoring site W m+1 Record flood event H n The forecast range Y(H n ), for the forecast range Y(H n ) are deduplicated, and after deduplication, the monitoring sites with the largest continuous range are selected to form the alarm range and sent to the staff port.

5. A watershed collaborative forecasting and optimization system for water conservancy big data, which executes the watershed collaborative forecasting and optimization method according to any one of claims 1 to 4, characterized in that: The system includes: a monitoring deployment and time synchronization unit, an identification binding and feature flow generation unit, a data capture and matrix construction unit, a difference calculation and matrix update unit, and a collaborative forecast and alarm generation unit; The monitoring deployment and time synchronization unit is used to set up watershed monitoring sites and deploy IoT sensors, divide the monitoring area within the coverage area of ​​the watershed monitoring sites so that each IoT sensor collects hydrological data within the same monitoring area, and is also used to trigger an initialization instruction through the upstream starting watershed monitoring site to achieve synchronous calibration of the timestamps of each IoT sensor; The identifier binding and characteristic flow generation unit is used to bind a unique identifier to each watershed monitoring site and write site information. It is also used to configure the initialization synchronization timestamp. When a flood event occurs in the monitoring area, it triggers the IoT sensors in the monitoring area to collect data and generate a hydrological characteristic flow with a synchronization timestamp. The data capture and matrix construction unit captures the hydrological data of all IoT sensors during a flood event based on the hydrological characteristic flow and the monitoring area, and establishes a basin-wide data matrix with the hydrological characteristic flow as the horizontal dimension and the monitoring area as the vertical dimension; The difference calculation and matrix update unit is used to calculate the horizontal and vertical state differences with the digitized matrix elements of the entire watershed as the center, and update the matrix elements according to the maximum value of the difference; The collaborative forecast and alarm generation unit is used to remove the first row and first column of the data matrix of the entire basin to obtain the collaborative forecast matrix of the downstream site, calculate the collaborative forecast correlation between sites to determine the forecast range of the flood event, remove duplicate monitoring sites within the forecast range, select the largest continuous range segment to form the alarm range and send it to the staff port.

6. The basin collaborative forecasting and optimization system for water conservancy big data according to claim 5 is characterized in that: The identification binding and feature flow generation unit includes a regional device group module and an initialization timestamp module; The regional device group module is used to count all IoT sensors deployed in different monitoring areas to form regional device groups; The initialization timestamp module is used to configure the initialization synchronization timestamp.

7. The basin collaborative forecasting and optimization system for water conservancy big data according to claim 5 is characterized in that: The data capture and matrix construction unit includes a monitoring period generation module and a monitoring feature label generation module; The monitoring period generation module is used to divide the monitoring periods of different monitoring areas under the hydrological characteristic flow; The monitoring feature label generation module is used to capture the hydrological data of all IoT sensors generated during flood monitoring events to generate monitoring feature labels.

8. The basin collaborative forecasting and optimization system for water conservancy big data according to claim 5 is characterized in that: The difference calculation and matrix update unit includes a state replacement module and a matrix update module; The state replacement module is used to replace the horizontal state and the vertical state. The horizontal state refers to the state of the monitoring area itself when the monitoring period is replaced. The vertical state refers to the dual state between the two monitoring areas when the monitoring area is replaced downstream. The matrix update module is used to perform downstream gradient updates on the digitized matrix of the entire river basin.

Images