An AI algorithm-based water regime abnormality risk prediction method and system

Abstract

The application relates to the technical field of water regime prediction, in particular to a water regime abnormality risk prediction method and system based on an AI algorithm. The method comprises the following steps: according to water level abnormality conditions of each monitoring point, an abnormal monitoring point is selected from each monitoring point; according to the water level change intensity and the danger degree of each upstream monitoring point of the abnormal monitoring point in an abnormal period, and in combination with the distance from the abnormal monitoring point, the influence degree of the abnormal monitoring point in the abnormal period caused by upstream rainfall is determined, and then in combination with the length of an abnormal river section where the abnormal monitoring point is located, a water regime abnormality factor of the abnormal monitoring point is obtained; and the current water level of each monitoring point is fused according to the water regime abnormality factor of each monitoring point, so that the current water level of the river is obtained. Considering that different monitoring points are influenced differently, the water regime abnormality conditions of each monitoring point are fused, the data reliability of the water level of the river can be ensured, the water level accuracy of the river is improved, and the accuracy of the water regime risk early warning is improved.

Description

Technical Field

[0001] This invention relates to the field of hydrological forecasting technology, specifically to a method and system for predicting hydrological anomaly risks based on AI algorithms. Background Technology

[0002] River water level forecasting allows for early assessment of flood disasters, providing guiding data for disaster warning and prevention. Anomaly risk forecasting refers to analyzing and predicting abnormal water changes (such as water levels) to predict future water levels, thereby enabling risk assessment based on these predicted levels.

[0003] Current methods for predicting river levels using real-time water levels obtain the average water level from all monitoring points along the river. However, because different monitoring points are located in different parts of the river (e.g., upstream, midstream, downstream), they are affected by different factors. For example, the time lag in water level changes between different monitoring points and the varying degrees to which each point is affected by upstream rainfall result in different hydrological capacities. Directly using the average water level from all monitoring points without considering the different influences on each point affects the reliability of the river level data, leading to inaccurate water level readings and reducing the accuracy of flood risk warnings. Summary of the Invention

[0004] To address the technical problem of low reliability in existing river water level data, the present invention aims to provide a method and system for predicting water situation anomalies based on AI algorithms. The specific technical solution adopted is as follows:

[0005] In a first aspect of the present invention, a method for predicting water situation anomaly risks based on AI algorithms is provided, comprising:

[0006] Based on the abnormal water level at each monitoring point, abnormal monitoring points were selected from all monitoring points;

[0007] Based on the impact of the drastic and dangerous water level changes at each upstream monitoring point on the abnormal monitoring point during the abnormal period, and combined with the distance from the abnormal monitoring point, the degree of influence of upstream rainfall on the abnormal monitoring point during the abnormal period is determined.

[0008] By integrating the degree of influence of the abnormal monitoring points and combining it with the length of the abnormal river section where the abnormal monitoring points are located, the hydrological anomaly factor of the abnormal monitoring points is obtained.

[0009] The current water level of the river is obtained by integrating the current water level at each monitoring point with the abnormal water conditions at each monitoring point.

[0010] In an exemplary embodiment, the screening process for the anomaly monitoring points includes:

[0011] Based on the degree of water level anomaly at each monitoring point during each abnormal period, and combined with the duration of each abnormal period, the degree of water backlog at each monitoring point during each abnormal period is obtained; the degree of water backlog is positively correlated with both the degree of water level anomaly and the duration.

[0012] The degree of water backflow at each monitoring point during all abnormal periods is integrated, and the slope and river width at each monitoring point are combined to obtain the degree of danger at each monitoring point; the degree of danger is positively correlated with the degree of water backflow and slope, and negatively correlated with the river width.

[0013] Based on the degree of danger at each monitoring point, abnormal monitoring points are selected from all monitoring points.

[0014] In an exemplary embodiment, the process of obtaining the abnormal time period includes:

[0015] The water level at the monitoring point is obtained at each moment in the historical time period. The moment corresponding to the water level exceeding the water level safety threshold is regarded as an abnormal moment, and consecutive abnormal moments constitute the abnormal time period of the monitoring point.

[0016] The process of obtaining the degree of water level anomaly during the abnormal period includes:

[0017] Obtain the super-safe water level at each abnormal moment within the abnormal period, wherein the super-safe water level is the difference between the water level and the safe water level threshold;

[0018] The super-safe water level at each abnormal moment within the abnormal period is fitted with a straight line, and the slope of the fitted straight line is taken as the degree of water level abnormality within the abnormal period.

[0019] In an exemplary embodiment, the process of obtaining the degree of water level change includes:

[0020] The method obtains the percentage of peaks, the mean of peak-to-peak values, and the mean of peak-to-peak time intervals in the water level sequence during the target rainfall period. The target rainfall period is a rainfall period associated with an abnormal period. The water level sequence is composed of the water level at each moment during the target rainfall period. The peak-to-peak value is the water level difference between a peak and its nearest trough in the water level sequence. The peak-to-peak time interval is the time interval between two adjacent peaks.

[0021] The severity of water level changes during abnormal periods is obtained based on the proportion of wave peaks, the average peak-to-peak value, and the average time interval between wave peaks. The severity of water level changes is positively correlated with the proportion of wave peaks and the average peak-to-peak value, and negatively correlated with the average time interval between wave peaks.

[0022] In one exemplary embodiment, the process of obtaining the degree of influence includes:

[0023] Based on the severity of water level changes, the degree of danger, and the distance to each upstream monitoring point during the abnormal period, an impact index corresponding to each upstream monitoring point is obtained; the impact index is positively correlated with the severity of water level changes and the degree of danger, and negatively correlated with the distance.

[0024] By integrating the impact indicators corresponding to all upstream monitoring points of the abnormal monitoring point, the degree of impact of upstream rainfall on the abnormal monitoring point during the abnormal period can be obtained.

[0025] In an exemplary embodiment, the process of obtaining the hydrological anomaly factors includes:

[0026] The hydrological carrying capacity of the abnormal monitoring points is obtained based on the degree of impact of the abnormal monitoring points during each abnormal period and the time interval between adjacent abnormal periods; the hydrological carrying capacity is inversely correlated with the degree of impact and positively correlated with the time interval between adjacent abnormal periods.

[0027] Based on the difference in hydrological carrying capacity of the abnormal monitoring points and the length of the abnormal river segment where the abnormal monitoring point is located, the hydrological anomaly factor of the abnormal monitoring point is obtained; the hydrological anomaly factor is positively correlated with the difference and the length of the abnormal river segment where the abnormal monitoring point is located; the difference is the difference between the overall level of hydrological carrying capacity and the hydrological carrying capacity of the abnormal monitoring point, and the overall level of hydrological carrying capacity is the average hydrological carrying capacity of all abnormal monitoring points in the abnormal river segment where the abnormal monitoring point is located.

[0028] In an exemplary embodiment, the process of obtaining the current water level of the river includes:

[0029] The influence weight of water level at each monitoring point is obtained from the water condition anomaly factors at each monitoring point.

[0030] Based on the influence weight of water level at each monitoring point, the current water level at each monitoring point is weighted and summed to obtain the current water level of the river.

[0031] In an exemplary embodiment, among the water condition anomaly factors of each monitoring point, the water condition anomaly factors of the monitoring points other than the abnormal monitoring points are set to preset values, and the preset values ​​are less than the water condition anomaly factors of the abnormal monitoring points.

[0032] In one exemplary embodiment, the abnormal river segment consists of a series of abnormal monitoring points.

[0033] In a second aspect of the present invention, a water situation anomaly risk prediction system based on AI algorithm is provided, comprising: a memory and a processor; the memory is connected to the processor; the memory is used to store program instructions; the processor is used to implement the above-described water situation anomaly risk prediction method based on AI algorithm when the program instructions are executed.

[0034] This invention offers the following advantages: First, based on the abnormal water levels at each monitoring point, abnormal monitoring points are selected. Then, focusing on the influence of upstream monitoring points on these abnormal points, and considering the length of the abnormal river segment where each monitoring point is located, hydrological anomaly factors are derived. These hydrological anomaly factors are considered important parameters affecting the current river water level. The current river water level is then obtained by merging the hydrological anomaly factors of each monitoring point with the current water levels of all monitoring points. Therefore, this invention does not directly obtain the current river water level by averaging the current water levels of all monitoring points. Instead, it emphasizes the different influences on different monitoring points, ensuring the reliability of the river water level data, improving the accuracy of river water level data, and thus enhancing the accuracy of flood risk warnings. Attached Figure Description

[0035] Figure 1 This is a flowchart of a method for predicting water situation anomalies based on AI algorithms, provided in one embodiment of the present invention;

[0036] Figure 2 This is a flowchart of the screening process for anomaly monitoring points provided in one embodiment of the present invention;

[0037] Figure 3 This is a flowchart illustrating the process of obtaining the degree of water level change according to one embodiment of the present invention;

[0038] Figure 4 This is a flowchart illustrating the process of obtaining the degree of influence according to one embodiment of the present invention;

[0039] Figure 5 This is a flowchart of the process for obtaining hydrological anomaly factors according to an embodiment of the present invention;

[0040] Figure 6 This is a flowchart of the process for obtaining the current water level of a river, provided in one embodiment of the present invention. Detailed Implementation

[0041] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. All data and information collected in this application have been obtained with full consent.

[0043] This embodiment provides an AI-based method for predicting hydrological anomalies and risks. The specific scenario addressed is one where rivers are widely distributed, precipitation distribution often exhibits local variations, and the hydrological carrying capacity differs at different locations within the river. Upstream rainfall significantly impacts downstream hydrological changes. Changes in upstream precipitation propagate downstream, leading to rises in downstream water levels. Downstream areas are easily affected by upstream rainfall, causing water level fluctuations and affecting the assessment of hydrological anomaly risks at monitoring points. This embodiment can determine the hydrological anomaly factors for each monitoring point, obtain comprehensive water level data, and improve the accuracy of hydrological anomaly risk warnings.

[0044] The hydrological characteristics of a river (such as water level and flow velocity) vary along its longitudinal direction. The speed, depth, and width of the flow can differ at different locations within a river, potentially leading to anomalies in the hydrological conditions. For example, in some sections of a river, water levels may rise rapidly due to greater flow resistance, while in other sections, there may be no significant change.

[0045] Multiple monitoring points are arranged along the river's flow direction. In one exemplary embodiment, the monitoring points are evenly spaced, with the distance between adjacent monitoring points set according to actual needs, such as 1 km. The number of monitoring points is determined by the spacing between them and the length of the river to be monitored. A 0.5 km radius around each monitoring point is designated as its monitoring area. The positions of each monitoring point are sequentially marked according to the river's flow direction, thereby determining the upstream and downstream relationships of each monitoring point.

[0046] A water level sensor, such as a radar water level sensor, is installed at each monitoring point to obtain the water level at each point. The sampling frequency of the water level sensor is set according to actual needs, such as once per minute.

[0047] This embodiment is used to obtain the current water level of a river. A historical time period is set relative to the current moment; the length of this historical time period is determined by actual needs, and this embodiment uses the most recent week as an example. Additionally, the rainfall periods for each monitoring area of ​​the river within the historical time period are obtained from the meteorological bureau. The slope of each monitoring point is determined using a slope meter / tilt meter or GPS device, and the river width at each monitoring point is obtained.

[0048] For the river to be monitored, the water level safety threshold of the river is determined. The water level safety threshold of the river is the highest water level within the safe range of the river, which is determined by the specific conditions of the river. In this embodiment, the water level safety threshold is 10 meters.

[0049] like Figure 1 As shown in the figure, the water situation anomaly risk prediction method based on AI algorithm provided in this embodiment includes the following steps:

[0050] Step S1: Based on the abnormal water level at each monitoring point, select the abnormal monitoring points from among the monitoring points;

[0051] Step S2: Based on the impact of the drastic water level changes and the degree of danger of each upstream monitoring point on the abnormal monitoring point during the abnormal period, and combined with the distance from the abnormal monitoring point, determine the degree of impact of upstream rainfall on the abnormal monitoring point during the abnormal period.

[0052] Step S3: Combine the influence degree of the abnormal monitoring points with the length of the abnormal river section where the abnormal monitoring points are located to obtain the water condition anomaly factor of the abnormal monitoring points;

[0053] Step S4: The current water level of each monitoring point is obtained by fusing the water anomaly factors of each monitoring point with the current water level of each monitoring point.

[0054] The following description, in conjunction with the accompanying drawings, explains each step in detail.

[0055] Step S1: Based on the abnormal water level at each monitoring point, select the abnormal monitoring points from among the monitoring points.

[0056] Rivers are widely distributed, and precipitation distribution often exhibits local variations. Therefore, it is necessary to analyze the abnormal water level changes at each monitoring point to identify which monitoring points exhibit anomalies, thereby improving the accuracy of subsequent risk predictions. Based on the abnormal water level conditions at each monitoring point, abnormal monitoring points are selected from among them.

[0057] In one exemplary embodiment, such as Figure 2 As shown, the following is a specific screening process for anomaly monitoring points:

[0058] Step S11: Based on the degree of water level anomaly at each monitoring point during each abnormal period, and combined with the duration of each abnormal period, obtain the degree of backwater at each monitoring point during each abnormal period.

[0059] First, identify the anomalous periods within the historical timeframe. It should be understood that anomalous periods are determined based on abnormal water levels at monitoring points; therefore, each monitoring point has its corresponding anomalous period, and these anomalous periods may differ between different monitoring points.

[0060] For any given monitoring point, the process for obtaining the corresponding abnormal time period is as follows: Obtain the water level at each moment within the historical time period for that monitoring point. The moments corresponding to water levels exceeding the safe water level threshold are considered abnormal moments for that monitoring point. Finally, consecutive abnormal moments constitute the abnormal time period for that monitoring point, thus obtaining several abnormal time periods corresponding to that monitoring point. Simultaneously, for any abnormal time period for that monitoring point, calculate the water level at each abnormal moment within that time period minus the safe water level threshold. The difference between the water level at each abnormal moment and the safe water level threshold is taken as the excess safe water level at that abnormal moment within that time period. In other words, the excess safe water level is the difference between the water level at the abnormal moment and the safe water level threshold. Therefore, the greater the water level exceeds the safe water level threshold at an abnormal moment, the greater the excess safe water level at that moment.

[0061] Rivers naturally have the function of channeling water flow. When the water level rises, the river's flow velocity and volume will change, which may cause the water level to gradually return to normal. For example, after heavy rainfall ends, the water in the basin will gradually shift downstream or to other areas, and the water level will naturally drop and return to normal. Therefore, if the water level exceeding the safe threshold continues to increase at each abnormal moment during the abnormal period, that is, the actual water level gradually moves away from the safe water level threshold, it indicates that the water flow channeling function of the river section or area where the monitoring point is located is poor, and it cannot effectively remove excess water, making flooding more likely. In an exemplary embodiment, the trend of the water level exceeding the safe threshold during the abnormal period is obtained by: performing a linear fitting on the water level exceeding the safe threshold at each abnormal moment during the abnormal period, obtaining the slope of the fitted line, and using the slope of the fitted line as the degree of water level anomaly during the abnormal period. It should be understood that when the slope is greater than 0, it indicates that the water level exceeding the safe level during the abnormal period is on an increasing trend, indicating a higher degree of water level anomaly; the more severe the increasing trend, the higher the degree of water level anomaly. When the slope is equal to 0, it indicates that the water level exceeding the safe level during the abnormal period is in a stable state. When the slope is less than 0, it indicates that the water level exceeding the safe level during the abnormal period is on a decreasing trend, indicating a lower degree of water level anomaly; the stronger the decreasing trend, the lower the degree of water level anomaly. Therefore, in terms of the slope value, the larger the slope, the higher the degree of water level anomaly during the abnormal period.

[0062] The duration of each abnormal period at the monitoring point is obtained. The longer the abnormal period, the greater the risk of flooding due to ineffective drainage during that period. Therefore, based on the degree of water level anomaly at the monitoring point during each abnormal period, combined with the duration of each abnormal period, the degree of backwater at the monitoring point during each abnormal period is determined. The degree of backwater characterizes the severity of water level rise or accumulation after water flow is obstructed. The degree of backwater is the opposite of water flow drainage capacity; the higher the degree of backwater at the monitoring point during a particular abnormal period, the worse the water flow drainage capacity during that abnormal period.

[0063] The higher the degree of water level anomaly during the abnormal period, the worse the water flow management capacity during that period, and the higher the degree of water backlog. Similarly, the longer the duration of the abnormal period, the worse the water flow management capacity and the higher the degree of water backlog. Therefore, the degree of water backlog during an abnormal period is positively correlated with both the degree of water level anomaly and the duration of the abnormal period. In an exemplary embodiment, a specific method for quantifying the degree of water backlog is given below:

[0064] ;

[0065] in, This indicates the degree of water backlog during the q-th abnormal time period. This represents the degree of water level anomaly during the q-th abnormal period (i.e., the slope of the fitted line of the water level exceeding the safe level during the q-th abnormal period). This represents the duration of the q-th abnormal period. `norm` represents the normalization function, where, due to... The value can be positive, 0, or negative. Therefore, in this embodiment, the normalization method is the sigmoid function.

[0066] Step S12: Combine the backwater level of each monitoring point during all abnormal periods, and combine the slope and river width at each monitoring point to obtain the danger level of each monitoring point.

[0067] In one exemplary embodiment, the average level of water backlog at a monitoring point across all abnormal periods is calculated by integrating the water backlog levels at that monitoring point during all abnormal periods. A higher level of water backlog at a monitoring point during each abnormal period indicates a poorer water flow management capacity and a higher level of danger for that monitoring point. Therefore, the level of danger at a monitoring point is positively correlated with the average level of water backlog at that monitoring point across all abnormal periods.

[0068] The steeper the slope at a monitoring point, the greater the inclination of the river channel, the faster the water flow, and the greater the danger—in other words, the higher the danger level of the monitoring point. Therefore, the danger level of a monitoring point is positively correlated with its slope. Conversely, the narrower the river channel at a monitoring point, the faster the water flow, leading to more concentrated floodwaters and more severe damage—in other words, the higher the danger level of the monitoring point. Therefore, the danger level of a monitoring point is inversely correlated with its river channel width.

[0069] Therefore, the hazard level of the monitoring point is obtained based on the average backwater level during all abnormal periods, the slope of the monitoring point, and the width of the river channel. In an exemplary embodiment, a specific method for quantifying the hazard level is given below:

[0070] ;

[0071] in, This indicates the degree of danger at the j-th monitoring point. This represents the slope of the j-th monitoring point. Let represent the width of the river channel at the j-th monitoring point, and exp represent an exponential function with the natural constant e as the base. This represents the average level of backwater at the j-th monitoring point during all abnormal periods.

[0072] Step S13: Based on the degree of danger of each monitoring point, select abnormal monitoring points from the monitoring points.

[0073] Step S12 yields the hazard level of each monitoring point. A higher hazard level indicates a higher probability of flooding and a more abnormal monitoring point. Therefore, in an exemplary embodiment, a hazard level threshold is preset. This threshold is set according to actual judgment needs. If the judgment is strict, the hazard level threshold can be set relatively low; in this embodiment, 0.6 is used as an example. The hazard level of each monitoring point is compared with this hazard level threshold. Monitoring points with a hazard level greater than or equal to the threshold are designated as abnormal monitoring points. Abnormal monitoring points represent monitoring points with large water level fluctuations and weak drainage capacity. The monitoring area corresponding to each abnormal monitoring point is designated as a danger zone. Furthermore, all other monitoring points besides the abnormal monitoring points are defined as normal monitoring points.

[0074] Step S2: Based on the impact of the drastic water level changes and the degree of danger of each upstream monitoring point on the abnormal monitoring point during the abnormal period, and combined with the distance from the abnormal monitoring point, determine the degree of influence of upstream rainfall on the abnormal monitoring point during the abnormal period.

[0075] Step S1 involves identifying anomalous monitoring points from each monitoring point, i.e., identifying several hazardous areas from each monitoring region. Since a river is a continuous system, not all monitoring areas are risk areas. Analyzing anomalous monitoring points helps identify areas more prone to flooding.

[0076] For any given anomaly monitoring point, identify all upstream monitoring points. It should be understood that other anomaly monitoring points may exist among the upstream monitoring points of that anomaly monitoring point.

[0077] For any upstream monitoring point of the anomaly monitoring point, during any rainfall period, determine the severity of water level changes at that upstream monitoring point during that rainfall period. In an exemplary embodiment, such as... Figure 3 As shown, the following is a process for obtaining the degree of drastic water level changes:

[0078] Step S21: Obtain the percentage of peaks, the average peak value, and the average peak time interval in the water level sequence during the target rainfall period.

[0079] Determine the water levels at various times during the rainfall period to form a water level sequence for that period. Obtain the peaks and troughs in the water level sequence to determine the number of peaks. Calculate the ratio of the number of peaks to the total number of water levels in the sequence, which is the percentage of peaks in the water level sequence. A larger percentage of peaks indicates more frequent and drastic changes in water level during the rainfall period, reflecting greater volatility and a higher degree of water level fluctuation.

[0080] For any given peak, find the nearest trough and calculate the water level difference between the peak and trough. This difference is taken as the peak-to-peak value of the peak. This process yields the peak-to-peak values ​​for each peak during the rainfall period. Then, calculate the average of these peak-to-peak values ​​as the mean. A larger mean mean value indicates a greater amplitude of water level change during the rainfall period, signifying more drastic and dramatic changes in water level.

[0081] For each wave peak within a rainfall period, the time interval between two adjacent wave peaks is obtained as the wave peak time interval. This yields the time intervals for each wave peak within the rainfall period. The average of these time intervals is then calculated as the mean wave peak time interval. A smaller mean wave peak time interval indicates a faster rate of water level change and more drastic water level fluctuations within the rainfall period.

[0082] Step S22: Based on the proportion of wave peaks, the average value of peak values, and the average time interval between wave peaks, the degree of water level change during abnormal periods is obtained.

[0083] The above analysis shows that the severity of water level changes is positively correlated with the proportion of wave crests and the mean of their peak values, and negatively correlated with the mean of the time interval between wave crests. In an exemplary embodiment, a specific method for quantifying the severity of water level changes is given below:

[0084] ;

[0085] in, This indicates the degree of drastic change in water level at the s-th upstream monitoring point of the w-th anomaly monitoring point during the e-th rainfall period. This represents the number of peaks in the water level sequence of the s-th upstream monitoring point at the w-th anomaly monitoring point during the e-th rainfall period. This represents the total number of water levels in the water level sequence of the s-th upstream monitoring point at the w-th anomaly monitoring point during the e-th rainfall period. Let represent the mean of the peak-to-peak values ​​in the water level sequence of the s-th upstream monitoring point at the w-th anomaly monitoring point during the e-th rainfall period. This represents the average time interval between peaks in the water level sequence of the s-th upstream monitoring point at the w-th anomaly monitoring point during the e-th rainfall period.

[0086] Since rainfall affects river water levels, there is a temporal correlation between rainfall periods and abnormal periods. Therefore, based on the temporal relationship between rainfall periods and abnormal periods within a historical timeframe, the rainfall period associated with the abnormal period at the monitoring point is determined as the target rainfall period for the abnormal period. Specifically, since water levels only change after rainfall, for the q-th rainfall period at the w-th abnormal monitoring point, the start time of this q-th rainfall period (i.e., the first moment) and the start times of each rainfall period at the s-th upstream monitoring point of the w-th abnormal monitoring point are obtained. Then, the start time of the rainfall period preceding the start time of the q-th abnormal period and having the shortest time interval with it is obtained. This rainfall period is defined as the target rainfall period associated with the q-th abnormal period relative to the s-th upstream monitoring point of the w-th abnormal monitoring point. Thus, the target rainfall period for each upstream monitoring point of the w-th abnormal monitoring point corresponding to the q-th rainfall period of the w-th abnormal monitoring point is obtained. Furthermore, the degree of water level change during the target rainfall period at the s-th upstream monitoring point of the w-th abnormal monitoring point is taken as the degree of water level change during the q-th abnormal period at the w-th abnormal monitoring point relative to the s-th upstream monitoring point.

[0087] When the water level changes more drastically during rainfall in the upstream area and the water flow diversion capacity is poor, it will cause the water level at the upstream monitoring point to rise sharply and the water flow speed to increase. These changes may affect the downstream area after a certain period of time, increasing the possibility of flooding downstream.

[0088] Therefore, based on the impact of the drastic and dangerous water level changes at each upstream monitoring point on the abnormal monitoring point during the abnormal period, and combined with the distance from the abnormal monitoring point, the degree of influence of upstream rainfall on the abnormal monitoring point during the abnormal period is determined. In an exemplary embodiment, such as... Figure 4 As shown, the following is a specific process for obtaining the degree of influence:

[0089] Step S23: Based on the severity of water level changes, the degree of danger, and the distance from each upstream monitoring point to the abnormal monitoring point during the abnormal period, obtain the impact index corresponding to each upstream monitoring point.

[0090] The greater the drastic the water level change and the worse the water flow conduction capacity, the more the abnormal monitoring point is affected by upstream rainfall during the abnormal period, and the higher the impact index. Similarly, the higher the level of danger, the worse the water flow conduction capacity, and the more the abnormal monitoring point is affected by upstream rainfall during the abnormal period, the higher the impact index. The greater the distance between the upstream monitoring point and the abnormal monitoring point (the distance is not a straight-line distance between the two monitoring points, but a distance along the river), the weaker the impact on the abnormal monitoring point, and the weaker the impact of upstream rainfall on the abnormal monitoring point during the abnormal period, the lower the impact index. Therefore, the impact index is positively correlated with the drasticness of water level change and the level of danger, and inversely correlated with distance. In an exemplary embodiment, a specific quantification method for the impact index is given below:

[0091] ;

[0092] in, This represents the impact index of the w-th abnormal monitoring point during the q-th abnormal period relative to the s-th upstream monitoring point. This indicates the degree of drastic change in water level at the w-th anomalous monitoring point during the q-th anomalous period, relative to the s-th upstream monitoring point. This indicates the danger level of the s-th upstream monitoring point of the w-th anomaly monitoring point. This represents the distance between the w-th anomaly monitoring point and its s-th upstream monitoring point.

[0093] Step S24: Integrate the impact indicators corresponding to all upstream monitoring points of the abnormal monitoring point to obtain the degree of impact of upstream rainfall on the abnormal monitoring point during the abnormal period.

[0094] For the q-th anomalous time period of the w-th anomalous monitoring point, calculate the average value of the impact index of the w-th anomalous monitoring point during the q-th anomalous time period relative to each upstream monitoring point. This average value is used as the degree of influence of upstream rainfall on the w-th anomalous monitoring point during the q-th anomalous time period. The calculation formula is as follows:

[0095] ;

[0096] in, Let S represent the degree of influence of upstream rainfall on the w-th anomalous monitoring point during the q-th anomalous period, and let S represent the number of upstream monitoring points for the w-th anomalous monitoring point. This allows us to obtain the degree of influence of upstream rainfall on the w-th anomalous monitoring point during each anomalous period.

[0097] Step S3: Combine the influence degree of the abnormal monitoring points with the length of the abnormal river section where the abnormal monitoring points are located to obtain the water condition anomaly factor of the abnormal monitoring points.

[0098] Step S2 obtains the impact of a single rainfall event on the upstream monitoring point of the abnormal monitoring point. Even if the impact of a single rainfall event on the abnormal monitoring point is small, if rainfall occurs frequently over a continuous period, the changes in water level and flow will gradually accumulate, increasing the risk of flooding. For example, the water level at a certain abnormal monitoring point may rise slightly after an upstream rainfall event, but if the upstream area continues to experience rainfall in the following period, the gradual rise in water level may exceed the capacity of the abnormal monitoring point, thereby triggering a more severe flood. Therefore, for any abnormal monitoring point, the impact level of the abnormal monitoring point is integrated with the length of the abnormal river segment in which the abnormal monitoring point is located to obtain the hydrological anomaly factor of the abnormal monitoring point. In an exemplary embodiment, such as... Figure 5 As shown, the following is a specific process for obtaining hydrological anomaly factors:

[0099] Step S31: Based on the degree of influence of the abnormal monitoring point in each abnormal period and the time interval between adjacent abnormal periods, the hydrological carrying capacity of the abnormal monitoring point is obtained.

[0100] For the w-th anomalous monitoring point, the higher the impact of the w-th anomalous monitoring point during each anomalous period, the more affected the w-th anomalous monitoring point is by upstream rainfall, and the weaker the hydrological carrying capacity of the w-th anomalous monitoring point. Hydrological carrying capacity is inversely correlated with the degree of impact. Conversely, the shorter the time interval between two adjacent anomalous periods for the w-th anomalous monitoring point, the less time the rainfall at the w-th anomalous monitoring point has to dissipate moisture, increasing the pressure on the hydrological system and weakening the hydrological carrying capacity of the w-th anomalous monitoring point. Therefore, hydrological carrying capacity is positively correlated with the time interval between adjacent anomalous periods. In an exemplary embodiment, the following is a method for quantifying the hydrological carrying capacity of the w-th anomalous monitoring point:

[0101] ;

[0102] in, This represents the hydrological carrying capacity of the w-th anomaly monitoring point. This represents the time interval between the q-th anomalous period at the w-th anomalous monitoring point and its preceding anomalous period, i.e., the (q-1)-th anomalous period, where Q represents the number of anomalous periods at the w-th anomalous monitoring point. It should be understood that since there are no anomalous periods before the first anomalous period in the time series, the time interval between the first anomalous period and its preceding anomalous periods is not calculated; that is, the first anomalous period in the time series is not included in the calculation of hydrological carrying capacity.

[0103] Using the above process, the current hydrological carrying capacity of each abnormal monitoring point is obtained. A value of 1 is set as the current hydrological carrying capacity of each normal monitoring point. This yields the water level carrying capacity of each monitoring point in the river. Then, by arranging the water level carrying capacities of each monitoring point in the river in order from upstream to downstream, a sequence of the current river channel's hydrological carrying capacity is obtained.

[0104] In this embodiment, along the river direction, consecutive adjacent abnormal monitoring points are grouped into abnormal river segments, that is, the danger zones corresponding to consecutive adjacent abnormal monitoring points are grouped into abnormal river segments, thus obtaining several abnormal river segments. It should be understood that the length of an abnormal river segment is the length of the superimposed area formed by the superposition of the danger zones corresponding to each of the included abnormal monitoring points along the river direction.

[0105] Step S32: Based on the differences in the hydrological carrying capacity of the abnormal monitoring points and the length of the abnormal river section where the abnormal monitoring points are located, obtain the hydrological anomaly factors of the abnormal monitoring points.

[0106] For the w-th anomalous monitoring point, the anomalous river segment in which the w-th anomalous monitoring point is located is determined, and the length of the anomalous river segment in which the w-th anomalous monitoring point is located is obtained. Since the anomalous river segment is composed of several anomalous monitoring points, the hydrological carrying capacity of all anomalous monitoring points (including the w-th anomalous monitoring point) in the anomalous river segment in which the w-th anomalous monitoring point is located is obtained. Then, the average hydrological carrying capacity of all anomalous monitoring points in the anomalous river segment in which the w-th anomalous monitoring point is located is calculated as the overall hydrological carrying capacity level corresponding to the w-th anomalous monitoring point.

[0107] The difference between the overall hydrological carrying capacity level corresponding to the w-th anomaly monitoring point and the hydrological carrying capacity of the w-th anomaly monitoring point is defined as the gap between the w-th anomaly monitoring point and the overall level. The larger the difference, the greater the gap between the hydrological carrying capacity of the w-th anomaly monitoring point and the overall level, the greater the hydrological pressure faced by the w-th anomaly monitoring point, and the larger the hydrological anomaly factor of the w-th anomaly monitoring point. The hydrological anomaly factor is positively correlated with the gap between the anomaly monitoring points.

[0108] The longer the abnormal river segment where the w-th anomaly monitoring point is located, the greater the hydrological pressure faced by the w-th anomaly monitoring point, and the greater the hydrological anomaly factor of the w-th anomaly monitoring point. The hydrological anomaly factor is positively correlated with the length of the abnormal river segment where the anomaly monitoring point is located.

[0109] In one exemplary embodiment, a specific quantification method for hydrological anomaly factors is given below:

[0110] ;

[0111] in, This represents the water condition anomaly factor at the w-th abnormal monitoring point. This represents the difference between the w-th anomaly monitoring points. This represents the length of the abnormal river segment where the w-th abnormal monitoring point is located.

[0112] Through the above process, the hydrological anomaly factors for each abnormal monitoring point are obtained. In an exemplary embodiment, the hydrological anomaly factors for other monitoring points besides the abnormal monitoring points, i.e., each normal monitoring point, are set to a preset value, and this preset value is less than the hydrological anomaly factor for each abnormal monitoring point. Based on the calculation formula for the hydrological anomaly factor of the abnormal monitoring points, the value 1 is set as the hydrological anomaly factor for each normal monitoring point. Thus, the hydrological anomaly factors for each monitoring point are obtained.

[0113] Step S4: The current water level of each monitoring point is obtained by fusing the water anomaly factors of each monitoring point with the current water level of each monitoring point.

[0114] For any given monitoring point, the greater the hydrological anomaly factor, the greater the weight should be given to the current water level data of that monitoring point, so as to ensure that the current water level of the river can better reflect the overall trend of the river's water level change, rather than relying solely on the water level of a specific monitoring point.

[0115] In one exemplary embodiment, such as Figure 6 As shown, the following is a specific process for obtaining the current water level of a river:

[0116] Step S41: Obtain the water level influence weight of each monitoring point from the water condition anomaly factors of each monitoring point.

[0117] Calculate the sum of the water level anomaly factors at all monitoring points, and then calculate the ratio of the water level anomaly factor at each monitoring point to this sum, which is used as the water level influence weight for each monitoring point, so that the sum of the water level influence weights for all monitoring points is 1.

[0118] Step S42: Based on the influence weight of water level at each monitoring point, the current water level at each monitoring point is weighted and summed to obtain the current water level of the river.

[0119] Based on the influence weight of water levels at each monitoring point, the current water levels at each monitoring point are weighted and summed to obtain the current river water level. The calculation formula is as follows:

[0120] ;

[0121] Where U represents the current water level of the river. This represents the weight of the water level influence at the j-th monitoring point. Let J represent the current water level at the j-th monitoring point, and J represent the total number of monitoring points.

[0122] By using a weighted average, we can reduce large water level fluctuations at local monitoring points caused by local factors, as well as reduce the impact of measurement errors or equipment failures at individual monitoring points. This results in a smoother water level that better represents the overall water level status and improves the accuracy of the overall data.

[0123] Safety warnings can be issued based on the current water level of the river. When the current water level exceeds the safety threshold, an alarm signal will be immediately output.

[0124] Furthermore, by continuously updating the current time using the above process, the real-time river water level is obtained, forming a real-time river water level data sequence. Subsequently, existing data prediction algorithms, such as the autoregressive moving average model, are used to predict the river water level for future times based on the real-time river water level data sequence. Water level warnings are then issued based on the predicted river water levels, completing the prediction of abnormal river conditions and risks.

[0125] This embodiment also provides a water situation anomaly risk prediction system based on AI algorithm, including: a memory and a processor; the memory is connected to the processor, and the memory is used to store program instructions; the processor is used to implement the steps in the above embodiment of the water situation anomaly risk prediction method based on AI algorithm when the program instructions are executed.

[0126] In one exemplary embodiment, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the embodiment of the AI-based algorithm-based water situation anomaly risk prediction method.

[0127] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0128] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for predicting water situation anomaly risks based on AI algorithms, characterized in that, include: Based on the abnormal water level at each monitoring point, abnormal monitoring points were selected from all monitoring points; Based on the impact of the drastic and dangerous water level changes at each upstream monitoring point on the abnormal monitoring point during the abnormal period, and combined with the distance from the abnormal monitoring point, the degree of influence of upstream rainfall on the abnormal monitoring point during the abnormal period is determined. By integrating the degree of influence of the abnormal monitoring points and combining it with the length of the abnormal river section where the abnormal monitoring points are located, the hydrological anomaly factor of the abnormal monitoring points is obtained. The current water level of the river is obtained by merging the current water level at each monitoring point with the abnormal water conditions at each monitoring point; The screening process for the abnormal monitoring points includes: determining the degree of water level anomaly at each monitoring point during each abnormal period, and combining this with the duration of each abnormal period; the degree of water level anomaly is positively correlated with both the degree and duration of the anomaly; integrating the degree of water level anomaly at each monitoring point across all abnormal periods, and combining this with the slope and river width at each monitoring point, to determine the degree of danger at each monitoring point; the degree of danger is positively correlated with the degree of water level anomaly and the slope, and negatively correlated with the river width; and selecting abnormal monitoring points from among the monitoring points based on their degree of danger.

2. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The process of obtaining the abnormal time period includes: The water level at the monitoring point is obtained at each moment in the historical time period. The moment corresponding to the water level exceeding the water level safety threshold is regarded as an abnormal moment, and consecutive abnormal moments constitute the abnormal time period of the monitoring point. The process of obtaining the degree of water level anomaly during the abnormal period includes: Obtain the super-safe water level at each abnormal moment within the abnormal period, wherein the super-safe water level is the difference between the water level and the safe water level threshold; The super-safe water level at each abnormal moment within the abnormal period is fitted with a straight line, and the slope of the fitted straight line is taken as the degree of water level abnormality within the abnormal period.

3. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The process of obtaining the degree of drastic water level change includes: The method obtains the percentage of peaks, the mean of peak-to-peak values, and the mean of peak-to-peak time intervals in the water level sequence during the target rainfall period. The target rainfall period is a rainfall period associated with an abnormal period. The water level sequence is composed of the water level at each moment during the target rainfall period. The peak-to-peak value is the water level difference between a peak and its nearest trough in the water level sequence. The peak-to-peak time interval is the time interval between two adjacent peaks. The severity of water level changes during abnormal periods is obtained based on the proportion of wave peaks, the average peak-to-peak value, and the average time interval between wave peaks. The severity of water level changes is positively correlated with the proportion of wave peaks and the average peak-to-peak value, and negatively correlated with the average time interval between wave peaks.

4. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The process of obtaining the degree of influence includes: Based on the severity of water level changes, the degree of danger, and the distance to each upstream monitoring point during the abnormal period, an impact index corresponding to each upstream monitoring point is obtained; the impact index is positively correlated with the severity of water level changes and the degree of danger, and negatively correlated with the distance. By integrating the impact indicators corresponding to all upstream monitoring points of the abnormal monitoring point, the degree of impact of upstream rainfall on the abnormal monitoring point during the abnormal period can be obtained.

5. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The process of obtaining the hydrological anomaly factors includes: The hydrological carrying capacity of the abnormal monitoring points is obtained based on the degree of impact of the abnormal monitoring points during each abnormal period and the time interval between adjacent abnormal periods; the hydrological carrying capacity is inversely correlated with the degree of impact and positively correlated with the time interval between adjacent abnormal periods. Based on the difference in hydrological carrying capacity of the abnormal monitoring points and the length of the abnormal river segment where the abnormal monitoring point is located, the hydrological anomaly factor of the abnormal monitoring point is obtained; the hydrological anomaly factor is positively correlated with the difference and the length of the abnormal river segment where the abnormal monitoring point is located; the difference is the difference between the overall level of hydrological carrying capacity and the hydrological carrying capacity of the abnormal monitoring point, and the overall level of hydrological carrying capacity is the average hydrological carrying capacity of all abnormal monitoring points in the abnormal river segment where the abnormal monitoring point is located.

6. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The process of obtaining the current water level of the river includes: The influence weight of water level at each monitoring point is obtained from the water condition anomaly factors at each monitoring point. Based on the influence weight of water level at each monitoring point, the current water level at each monitoring point is weighted and summed to obtain the current water level of the river.

7. The method for predicting water situation anomalies based on AI algorithms as described in claim 6, characterized in that, Among the water condition anomaly factors of each monitoring point, the water condition anomaly factors of the monitoring points other than the abnormal monitoring points are set to preset values, and the preset values ​​are less than the water condition anomaly factors of the abnormal monitoring points.

8. The method for predicting water situation anomalies based on AI algorithms as described in claim 1, characterized in that, The abnormal river section consists of a series of abnormal monitoring points.

9. A hydrological anomaly risk prediction system based on AI algorithms, characterized in that it includes: Memory and processor; The memory is connected to the processor; The memory is used to store program instructions; The processor is configured to implement the AI-based water situation anomaly risk prediction method according to any one of claims 1-8 when the program instructions are executed.

Images