Flood early warning decision analysis system and method based on artificial intelligence

Abstract

The invention discloses a flood early warning decision analysis system and method based on artificial intelligence, and belongs to the technical field of flood early warning. Constructing a monitoring area database, wherein the monitoring area database comprises monitoring area rainfall data and area flood control data; based on the monitored area database, calculating the total precipitation of the monitored area within a preset time; calculating a risk value of flood occurrence of the monitored area; constructing a regional flood model; updating model parameters, and optimizing the regional flood model; when the deviation sum of the predicted value of the regional water level and the actual value of the regional water level reaches the minimum, finishing regional flood model optimization; when the risk value of the flood occurrence of the monitored area is greater than or equal to a risk threshold value, predicting the water level of each area of the monitored area through an area flood model; early warning lists are generated, and the early warning lists of all areas in the monitoring area are gathered and fed back to a monitoring area administrator; therefore, regional and refined early warning of flood disasters is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of flood warning, and in particular to a flood warning decision analysis system and method based on artificial intelligence. Background Art

[0003] At present, we can use meteorological satellites, radars, weather stations and other equipment to monitor indicators such as atmospheric circulation, precipitation, and wind speed, and predict precipitation conditions and the possibility of floods in the monitored area in the future. However, due to the different topography and landforms in different areas of the same region, the possibility of floods is also different. Therefore, how to issue flood warnings for different areas in the same region has become a problem we need to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a flood warning decision analysis system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A flood warning decision analysis method based on artificial intelligence, the method comprises the following steps:

[0007] Step S1: constructing a monitoring area database, wherein the monitoring area database includes precipitation data and regional flood control data in the monitoring area;

[0008] Step S2: Based on the monitoring area database, calculate the total precipitation in the monitoring area within a preset time; calculate the risk value of flood occurrence in the monitoring area;

[0009] Step S3: construct a regional flood model; update model parameters and optimize the regional flood model; when the deviation between the regional water level forecast value and the regional water level actual value reaches the minimum, the regional flood model optimization ends;

[0010] Step S4: When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model; a warning list is generated, and the warning lists of all areas in the monitored area are summarized and fed back to the administrator of the monitored area.

[0011] Furthermore, the specific implementation process of step S1 includes:

[0012] Step S1.1: Obtain the regional precipitation intensity and duration of regional precipitation within the preset time in the monitoring area through weather forecast; obtain regional precipitation intensity data and regional precipitation duration data through the regional meteorological bureau; measure regional water level data through a water level meter; measure regional soil moisture data through a soil moisture meter;

[0013] Step S1.2: Based on the distribution of hydrological stations in the monitoring area, construct a regional set S = [s g |g∈[1,G]}, where s g represents the g-th region, G represents the total number of regions in which the monitoring area is divided;

[0014] A monitoring area database is constructed, and the monitoring area database is used to store the monitoring area precipitation data and regional flood control data; the monitoring area precipitation data includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data within the region.

[0015] Furthermore, the specific implementation process of step S2 includes:

[0016] Step S2.1: Based on the precipitation data in the monitored area, construct a precipitation duration set TI = {T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time;

[0017] The precipitation intensity in the monitored area changes with time, and the duration of precipitation will also change accordingly. Therefore, in order to obtain the total precipitation data more accurately, a precipitation duration set is constructed based on the precipitation intensity in the monitored area.

[0018] Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Indicates the precipitation intensity of the mth region within the preset time;

[0019] Step S2.2: Based on the flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of times floods occurred in the monitoring area when the total precipitation in the monitoring area in the historical data was less than or equal to G, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area;

[0020] Construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of flood occurrence in the region as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line;

[0021] Step S2.3: Obtain the regional precipitation intensity and regional precipitation duration in the monitored area within the preset time in the weather forecast, calculate the total precipitation in the monitored area, and substitute the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

[0022] The total precipitation in the monitored area within the preset time is obtained through weather forecast, so as to make a preliminary estimate of the total precipitation in the monitored area; the topography of different regions is different. For example, mountainous areas with steep terrain and low vegetation coverage are more prone to floods, and plains with flat terrain and poor drainage are also easily affected by floods. Therefore, it is necessary to make a preliminary prediction of floods based on the probability of floods occurring at different total precipitation in the region's historical data; when the risk is too great, it is necessary to conduct a specific analysis of the specific areas in the monitored area, and increase the frequency of regional flood control data collection in order to enhance the accuracy of the prediction.

[0023] Furthermore, the specific implementation process of step S3 includes:

[0024] Step S3.1: Extract region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value;

[0025] Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s gThe regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of

[0026] Large amounts of precipitation accumulated over a long period of time will lead to soil saturation, reduce the soil's infiltration capacity, and convert more precipitation into surface runoff, increasing the possibility of floods. When soil moisture reaches saturation, the soil's infiltration capacity tends to stabilize and no longer increases or decreases. Therefore, when calculating the expected growth value of regional soil moisture, a piecewise function is used to quantify the expected growth value of regional soil moisture. As the soil gradually approaches saturation, its ability to continue absorbing moisture will weaken, so the rate of moisture growth will gradually slow down. Therefore, the growth coefficient decreases as soil moisture increases.

[0027] Step S3.2: Optimize the regional flood model: fit the regional water level prediction value and the regional water level actual value in the training data. The fitting formula of the regional water level prediction value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HU i Indicates the actual value of the regional water level. represents the total deviation between the predicted value of regional water level and the actual value of regional water level in the ath training data, and A represents the total number of training data;

[0028] The model parameters μ1, μ2, and μ3 are continuously updated until the deviation and e between the predicted regional water level and the actual regional water level in all training data are minimized. The training is stopped and the optimal solution of the model parameters μ1, μ2, and μ3 is obtained, and the regional flood model optimization is completed.

[0029] When a region receives a large amount of precipitation in a short period of time, if the surface drainage system cannot drain the water in time, it will easily lead to an increase in runoff. Therefore, the greater the precipitation per unit time, that is, the greater the precipitation intensity, the faster the regional soil moisture increases, until the soil moisture reaches the saturation value and stops growing. In addition to precipitation intensity, the duration of precipitation is also an important factor affecting the severity of flood disasters. Long-term precipitation will continuously replenish surface and underground moisture, making the soil in a saturated state for a long time, increasing the possibility of floods and the severity of disasters.

[0030] Furthermore, the specific implementation process of step S4 includes:

[0031] When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model:

[0032] Extraction area g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is set, the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; the early warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator.

[0033] The output regional water level forecast value is compared with each level threshold. When the precipitation in the i-th region lasts for a period of time T i When the predicted water level in the subsequent area is greater than the first flood threshold, the water level level is a blue flood warning; when the precipitation in the ith area lasts for T i When the predicted water level in the next area is greater than the second flood threshold, the water level is a yellow flood warning. i When the predicted water level in the next area is greater than the third flood threshold, the water level level is an orange flood warning. i When the predicted water level in the rear area is greater than the fourth flood threshold, the water level level is a red flood warning; the time and water level reaching each level are recorded, and a warning list is generated.

[0034] A flood warning decision analysis system based on artificial intelligence, the system comprising: a collection and storage module, a risk module, a regional flood module and a feedback module;

[0035] The data acquisition unit obtains the regional precipitation intensity and the regional precipitation duration in the preset time of the monitoring area through the weather forecast; obtains the regional precipitation intensity data and the regional precipitation duration data through the regional meteorological bureau; measures the regional water level data through the water level meter; and measures the regional soil moisture data through the soil moisture measuring instrument;

[0036] The data storage unit constructs a region set S={s g |g∈[1,G]}, where s g represents the g-th region, G represents the total number of regions divided into regions in the monitoring area; constructs a monitoring area database, which is used to store precipitation data and regional flood control data in the monitoring area; the precipitation data in the monitoring area includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data in the region.

[0037] Further, the risk module includes a risk function unit and a prediction unit;

[0038] The risk function unit constructs a precipitation duration set TI={T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time;

[0039] Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Represents the precipitation intensity of the mth region within a preset time; based on flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of floods in the monitoring area when the total precipitation in the monitoring area is less than or equal to G in the historical data, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area; construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of regional flood occurrence as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line;

[0040] The prediction unit calculates the total precipitation in the monitored area by obtaining the regional precipitation intensity and the duration of regional precipitation in the monitored area within a preset time in the weather forecast, and substitutes the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

[0041] Further, the regional flood module includes a construction unit and an optimization unit;

[0042] The construction unit is constructed by extracting the region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value;

[0043] Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s g The regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of

[0044] The optimization unit optimizes the regional flood model by fitting the regional water level prediction value and the regional water level actual value in the training data. The fitting formula of the regional water level prediction value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HU i Indicates the actual value of the regional water level. represents the sum of the deviations between the predicted regional water level and the actual regional water level in the ath training data, and A represents the total number of training data; the model parameters μ1, μ2, and μ3 are continuously updated until the sum of the deviations e between the predicted regional water level and the actual regional water level in all training data reaches the minimum, and the training is stopped to obtain the optimal solution of the model parameters μ1, μ2, and μ3, and the regional flood model optimization is completed.

[0045] Furthermore, when the risk value of flood in the monitored area is greater than or equal to the risk threshold, the feedback module predicts the water level of each area in the monitored area through the regional flood model: extracting area s g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is set, the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; the early warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator.

[0046] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: in a flood warning decision analysis system and method based on artificial intelligence provided by the present invention, a monitoring area database is constructed, and the monitoring area database includes precipitation data of the monitoring area and regional flood control data; based on the monitoring area database, the total precipitation of the monitoring area within a preset time is calculated; the risk value of flood occurrence in the monitoring area is calculated; a regional flood model is constructed; the model parameters are updated to optimize the regional flood model; when the sum of the deviations between the regional water level predicted value and the regional water level actual value reaches the minimum, the regional flood model optimization ends; when the risk value of flood occurrence in the monitoring area is greater than or equal to the risk threshold, the water level of each area in the monitoring area is predicted by the regional flood model; a warning list is generated, and the warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator; thereby realizing regionalized and refined warning of flood disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1It is a structural schematic diagram of a flood early warning decision analysis system based on artificial intelligence of the present invention;

[0049] Figure 2 It is a schematic diagram of the steps of a flood warning decision analysis method based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] See also Figure 1-Figure 2 , the present invention provides a technical solution:

[0052] A flood warning decision analysis method based on artificial intelligence, the method comprising the following steps:

[0053] Step S1: construct a monitoring area database, which includes precipitation data and regional flood control data in the monitoring area;

[0054] Step S2: Based on the monitoring area database, calculate the total precipitation in the monitoring area within a preset time; calculate the risk value of flood occurrence in the monitoring area;

[0055] Step S3: construct a regional flood model; update model parameters and optimize the regional flood model; when the deviation between the regional water level forecast value and the regional water level actual value reaches the minimum, the regional flood model optimization ends;

[0056] Step S4: When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model; a warning list is generated, and the warning lists of all areas in the monitored area are summarized and fed back to the administrator of the monitored area.

[0057] Specifically, the specific implementation process of step S1 includes:

[0058] Step S1.1: Obtain the regional precipitation intensity and duration of regional precipitation within the preset time in the monitoring area through weather forecast; obtain regional precipitation intensity data and regional precipitation duration data through the regional meteorological bureau; measure regional water level data through a water level meter; measure regional soil moisture data through a soil moisture meter;

[0059] Step S1.2: Based on the distribution of hydrological stations in the monitoring area, construct a regional set S = {s g |g∈[1,G]}, where s grepresents the g-th region, G represents the total number of regions in which the monitoring area is divided;

[0060] Construct a monitoring area database, which is used to store the monitoring area precipitation data and regional flood control data; the monitoring area precipitation data includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data within the region.

[0061] Furthermore, the specific implementation process of step S2 includes:

[0062] Step S2.1: Based on the precipitation data in the monitored area, construct a precipitation duration set TI = {T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time;

[0063] For example: the precipitation intensity in the detection area within the preset time is {140mm / day, 150mm / day, 60mm / day, 0mm / day, 200mm / day, 30mm / day, 140mm / day}, and the corresponding precipitation duration is {3 days, 2 days, 2 days, 0 days, 2 days, 4 days, 2 days}.

[0064] Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Indicates the precipitation intensity of the mth region within the preset time;

[0065] Step S2.2: Based on the flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of times floods occurred in the monitoring area when the total precipitation in the monitoring area in the historical data was less than or equal to G, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area;

[0066] Construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of flood occurrence in the region as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line;

[0067] Step S2.3: Obtain the regional precipitation intensity and regional precipitation duration in the monitored area within the preset time in the weather forecast, calculate the total precipitation in the monitored area, and substitute the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

[0068] For example: total precipitation G = 1640 mm. Substituting the total precipitation into the function, the risk value of total precipitation of 1640 mm is 0.9. Assuming the risk threshold is 0.6, floods are likely to occur in this area.

[0069] Furthermore, the specific implementation process of step S3 includes:

[0070] Step S3.1: Extract region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value;

[0071] Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s g The regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of

[0072] Step S3.2: Optimize the regional flood model: fit the regional water level prediction value and the regional water level actual value in the training data. The fitting formula of the regional water level prediction value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HU i Indicates the actual value of the regional water level. represents the total deviation between the predicted value of regional water level and the actual value of regional water level in the ath training data, and A represents the total number of training data;

[0073] The model parameters μ1, μ2, and μ3 are continuously updated until the deviation and e between the predicted regional water level and the actual regional water level in all training data are minimized. The training is stopped and the optimal solution of the model parameters μ1, μ2, and μ3 is obtained, and the regional flood model optimization is completed.

[0074] Furthermore, the specific implementation process of step S4 includes:

[0075] When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model:

[0076] Extraction area g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is set, the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; the early warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator.

[0077] A flood warning decision analysis system based on artificial intelligence, the system includes: a collection and storage module, a risk module, a regional flood module and a feedback module;

[0078] The data collection unit obtains the regional precipitation intensity and duration of regional precipitation within the preset time in the monitoring area through weather forecasts; obtains regional precipitation intensity data and regional precipitation duration data through the regional meteorological bureau; measures regional water level data through a water level meter; and measures regional soil moisture data through a soil moisture measuring instrument;

[0079] The data storage unit constructs a regional set S = {s g |g∈[1,G]}, where s gRepresents the g-th region, G represents the total number of regions in the monitoring area; constructs a monitoring area database, which is used to store the monitoring area precipitation data and regional flood control data; the monitoring area precipitation data includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data within the region.

[0080] Further, the risk module includes a risk function unit and a prediction unit;

[0081] The risk function unit constructs a precipitation duration set TI={T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time;

[0082] Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Represents the precipitation intensity of the mth region within a preset time; based on flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of floods in the monitoring area when the total precipitation in the monitoring area is less than or equal to G in the historical data, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area; construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of regional flood occurrence as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line;

[0083] The prediction unit calculates the total precipitation in the monitored area by obtaining the regional precipitation intensity and regional precipitation duration in the monitored area within a preset time in the weather forecast, and substitutes the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

[0084] Furthermore, the regional flood module includes a construction unit and an optimization unit;

[0085] Construct the unit by extracting the region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value;

[0086] Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s g The regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of

[0087] Optimization unit, optimizes the regional flood model: fits the regional water level forecast value and the regional water level actual value in the training data, and the fitting formula of the regional water level forecast value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HU i Indicates the actual value of the regional water level. represents the sum of the deviations between the predicted regional water level and the actual regional water level in the ath training data, and A represents the total number of training data; the model parameters μ1, μ2, and μ3 are continuously updated until the sum of the deviations e between the predicted regional water level and the actual regional water level in all training data reaches the minimum, and the training is stopped to obtain the optimal solution of the model parameters μ1, μ2, and μ3, and the regional flood model optimization is completed.

[0088] Furthermore, in the feedback module, when the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted by the regional flood model: extracting area s g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is set, the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; the early warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator.

[0089] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0090] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A flood warning decision analysis method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: constructing a monitoring area database, wherein the monitoring area database includes precipitation data and regional flood control data in the monitoring area; Step S2: Based on the monitoring area database, calculate the total precipitation in the monitoring area within a preset time; calculate the risk value of flood occurrence in the monitoring area; Step S3: construct a regional flood model; update model parameters and optimize the regional flood model; when the deviation between the regional water level forecast value and the regional water level actual value reaches the minimum, the regional flood model optimization ends; Step S4: When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model; a warning list is generated, and the warning lists of all areas in the monitored area are summarized and fed back to the administrator of the monitored area.

2. The flood warning decision analysis method based on artificial intelligence according to claim 1 is characterized in that: The specific implementation process of step S1 includes: Step S1.1: Obtain the regional precipitation intensity and duration of regional precipitation within the preset time in the monitoring area through weather forecast; obtain regional precipitation intensity data and regional precipitation duration data through the regional meteorological bureau; measure regional water level data through a water level meter; measure regional soil moisture data through a soil moisture meter; Step S1.2: Based on the distribution of hydrological stations in the monitoring area, construct a regional set S = {s g |g∈[1,G]}, where s g represents the g-th region, G represents the total number of regions in which the monitoring area is divided; A monitoring area database is constructed, and the monitoring area database is used to store the monitoring area precipitation data and regional flood control data; the monitoring area precipitation data includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data within the region.

3. The flood warning decision analysis method based on artificial intelligence according to claim 2 is characterized in that: The specific implementation process of step S2 includes: Step S2.1: Based on the precipitation data in the monitored area, construct a precipitation duration set TI = {T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time; Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Indicates the precipitation intensity of the mth region within the preset time; Step S2.2: Based on the flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of times floods occurred in the monitoring area when the total precipitation in the monitoring area in the historical data was less than or equal to G, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area; Construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of flood occurrence in the region as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line; Step S2.3: Obtain the regional precipitation intensity and regional precipitation duration in the monitored area within the preset time in the weather forecast, calculate the total precipitation in the monitored area, and substitute the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

4. The flood warning decision analysis method based on artificial intelligence according to claim 3 is characterized in that: The specific implementation process of step S3 includes: Step S3.1: Extract region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value; Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s g The regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of Step S3.2: Optimize the regional flood model: fit the regional water level prediction value and the regional water level actual value in the training data. The fitting formula of the regional water level prediction value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HU i Indicates the actual value of the regional water level. represents the total deviation between the predicted value of regional water level and the actual value of regional water level in the ath training data, and A represents the total number of training data; The model parameters μ1, μ2, and μ3 are continuously updated until the deviation and e between the predicted regional water level and the actual regional water level in all training data are minimized. The training is stopped and the optimal solution of the model parameters μ1, μ2, and μ3 is obtained, and the regional flood model optimization is completed.

5. The flood warning decision analysis method based on artificial intelligence according to claim 4 is characterized in that: The specific implementation process of step S4 includes: When the risk value of flood in the monitored area is greater than or equal to the risk threshold, the water level of each area in the monitored area is predicted through the regional flood model: Extraction Areas g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is set, the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; the early warning lists of all areas in the monitoring area are summarized and fed back to the monitoring area administrator.

6. A flood warning decision analysis system based on artificial intelligence, characterized in that: The system includes: a collection and storage module, a risk module, a regional flood module and a feedback module; The acquisition and storage module includes a data acquisition unit and a data storage unit, which are used to construct a monitoring area database, wherein the monitoring area database includes precipitation data and regional flood control data in the monitoring area; The risk module includes a risk function unit and a prediction unit, which calculates the total precipitation in the monitoring area within a preset time based on the monitoring area database; and calculates the risk value of flood occurrence in the monitoring area; The regional flood module includes a construction unit and an optimization unit, which are used to construct a regional flood model; update model parameters and optimize the regional flood model; when the deviation between the regional water level predicted value and the regional water level actual value reaches the minimum, the regional flood model optimization ends; The feedback module predicts the water level of each area in the monitoring area through the regional flood model when the risk value of flood occurrence in the monitoring area is greater than or equal to the risk threshold; generates an early warning list, and summarizes the early warning lists of all areas in the monitoring area and feeds them back to the monitoring area administrator.

7. The artificial intelligence-based flood warning decision analysis system according to claim 6, characterized in that: The acquisition and storage module includes a data acquisition unit and a data storage unit; The data collection unit obtains the regional precipitation intensity and regional precipitation duration within a preset time in the monitoring area through weather forecast; Obtain regional precipitation intensity data and regional precipitation duration data through regional meteorological bureaus; measure regional water level data through water level gauges; measure regional soil moisture data through soil moisture measuring instruments; The data storage unit constructs a region set S={s g |g∈[1,G]}, where s g represents the g-th region, G represents the total number of regions divided into regions in the monitoring area; constructs a monitoring area database, which is used to store precipitation data and regional flood control data in the monitoring area; the precipitation data in the monitoring area includes regional precipitation intensity, regional precipitation duration and flood data; the regional flood control data includes regional precipitation intensity data, regional soil moisture data, regional precipitation duration data and regional water level data in the region.

8. The artificial intelligence-based flood warning decision analysis system according to claim 7, characterized in that: The risk module includes a risk function unit and a prediction unit; The risk function unit constructs a precipitation duration set TI={T m |m∈[1,M]}, where M represents the total duration of precipitation in the region within the preset time, T m Indicates the duration of precipitation in the mth region within the preset time; Based on the monitoring area database, extract the historical monitoring area precipitation data and calculate the total precipitation in the monitoring area within the preset time: Among them, P m Represents the precipitation intensity of the mth region within a preset time; based on flood data, calculate the risk value of flood occurrence: R = count(G) / COUNT, where count(G) represents the number of floods in the monitoring area when the total precipitation in the monitoring area is less than or equal to G in the historical data, and COUNT represents the total number of floods in the monitoring area in the historical precipitation data of the monitoring area; construct a coordinate system with the total precipitation of the region as the horizontal axis and the risk value of regional flood occurrence as the vertical axis, map the data point (G, R) on the coordinate system to obtain the coordinate point, connect the adjacent coordinate points in the order of the horizontal axis coordinate value from small to large to generate a broken line, and obtain the function corresponding to the broken line; The prediction unit calculates the total precipitation in the monitored area by obtaining the regional precipitation intensity and the duration of regional precipitation in the monitored area within a preset time in the weather forecast, and substitutes the total precipitation in the monitored area within the preset time into the function to calculate the risk value of flood occurrence in the monitored area.

9. The artificial intelligence-based flood warning decision analysis system according to claim 8, characterized in that: The regional flood module includes a construction unit and an optimization unit; The construction unit is constructed by extracting the region s g The regional flood control data is used as training data to construct the regional s g The regional flood model takes in regional precipitation, regional precipitation duration and regional soil moisture as input, and outputs the regional water level prediction value; Regional Flood Models: Among them, T i Indicates area g The duration of precipitation in the i-th region within the preset time, H i Indicates area g The duration of precipitation in the i-th region is T i The predicted water level of the area after H0 represents the area s g The predicted value of regional water level before precipitation, P i Indicates area g The duration of precipitation in the i-th region is T i The regional precipitation intensity value at time w i Indicates area g The duration of precipitation in the i-th region is T i Soil moisture value in the rear area, (μ2P i *T i +μ3w i ) represents area s g The regional water level in the i-th region has a rainfall duration T i The expected growth after that, w0 represents the area s g The regional soil moisture value before precipitation, I represents the total duration of regional precipitation within the preset time, α(w i ) represents area s g The duration of precipitation in the i-th region is T i The expected increase in regional soil moisture at , k(P i ) represents area s g In the area, the precipitation intensity value is P i Time i The growth coefficient of w1 represents the area s g The saturation value of regional soil moisture, μ1, μ2, and μ3 represent the saturation value of regional soil moisture. g Model parameters of The optimization unit optimizes the regional flood model by fitting the regional water level prediction value and the regional water level actual value in the training data. The fitting formula of the regional water level prediction value and the regional water level actual value is: Where e represents the sum of the deviations between the regional water level prediction value and the regional water level actual value in all training data, HI i Indicates the actual value of the regional water level. represents the sum of the deviations between the predicted regional water level and the actual regional water level in the ath training data, and A represents the total number of training data; the model parameters μ1, μ2, and μ3 are continuously updated until the sum of the deviations e between the predicted regional water level and the actual regional water level in all training data reaches the minimum, and the training is stopped to obtain the optimal solution of the model parameters μ1, μ2, and μ3, and the regional flood model optimization is completed.

10. The artificial intelligence-based flood warning decision analysis system according to claim 9, characterized in that: The feedback module predicts the water level of each area in the monitoring area through the regional flood model when the risk value of flood in the monitoring area is greater than or equal to the risk threshold: extract the area s g The real-time water level of the hydrological station and the real-time regional soil moisture and regional s g The predicted value of regional precipitation intensity and the predicted value of regional precipitation duration within the preset time of the weather station are input into the regional s g In the regional flood model, the output area s g The regional water level is expected to be g The water level threshold is recorded, and the time and water level level when the regional water level forecast value reaches the water level threshold are recorded, and an early warning list is generated in chronological order; The early warning lists of all areas in the monitoring area will be summarized and fed back to the monitoring area administrator.

Images