A dam flood discharge control method based on precipitation

Abstract

This invention provides a dam flood discharge control method based on precipitation, relating to the field of water management technology. The method includes: determining the upstream watershed of the dam and dividing the watershed into multiple regions; calculating the increase in dam storage caused by precipitation in each time period based on the precipitation in each region, and further predicting and calculating the dam storage at the end of each time period; and adjusting the flood discharge volume in real time based on the dam storage at the end of the time period. This method helps improve the accuracy of dam storage volume prediction, enabling real-time dynamic prediction of dam storage volume, facilitating precise and reasonable flood discharge control of the dam, and achieving accurate early warning and rapid response to water-related disasters.

Description

Technical Field

[0001] This invention relates to the field of water management technology, and in particular to a method for controlling dam flood discharge based on precipitation. Background Technology

[0002] Water-related disasters caused by sudden rainfall (such as floods, flash floods, and dam failures) are characterized by their suddenness, destructiveness, and wide impact. Therefore, building an efficient and accurate water-related disaster early warning system is one of the core tasks of disaster prevention and mitigation. With the development of monitoring technology, a multi-source monitoring data system has been established, covering various data sources such as remote sensing, meteorology, ground sensors, and social media, providing rich data support for disaster early warning. Among these, dams are important physical nodes in disaster prevention and mitigation. Therefore, monitoring and forecasting hydrological information related to dams are crucial means of water-related disaster early warning, water flow regulation, and reducing disaster losses.

[0003] Traditional early warning systems for dam-related water conservancy disasters primarily rely on two core technological approaches: hydrological and hydraulic models based on physical mechanisms, and threshold methods and early machine learning models based on statistical laws. Hydrological and hydraulic models are based on physical laws such as watershed hydrological cycles and water flow, simulating disaster evolution through mathematical equations. Early machine learning models, on the other hand, rely on historical disaster data to uncover statistical correlations or simple nonlinear relationships between data points, enabling disaster early warning. However, with increasingly complex underlying surface conditions in watersheds (such as accelerated urbanization and changes in land use) and frequent extreme disaster events, the limitations of traditional early warning methods have become increasingly apparent, failing to meet the practical needs of accurate early warning and rapid response to water conservancy disasters in the new era. Furthermore, the heterogeneity of multi-source monitoring data (different spatiotemporal scales and formats) presents challenges to data fusion in traditional methods, hindering the full utilization of data value and resulting in insufficient and incomplete perception of the watershed's state, further impacting the accuracy and timeliness of early warnings. Summary of the Invention

[0004] To address the aforementioned technical problems in the existing technology, this invention aims to provide a dam flood discharge control method based on precipitation. This method considers the diversity of underlying surface conditions in the upstream basin of the dam and effectively utilizes multi-source monitoring data. Based on comprehensive and real-time perception of hydrological information in the upstream basin, it performs high-precision dynamic prediction of the dam's water storage capacity and conducts precise and reasonable flood discharge control of the dam based on the prediction results.

[0005] Specifically, the technical solution is as follows: A dam flood discharge control method based on precipitation, comprising:

[0006] Determine the upstream watershed of the dam and divide the watershed into... One region;

[0007] Calculated based on precipitation in each region Increased water storage in dams due to rainfall during a given period ;

[0008] Based on water storage Predictive calculation The water level in the dam at the end of the period ;

[0009] Based on water storage Adjust the flood discharge volume in real time.

[0010] Preferably, the watershed is divided into The specific methods for dividing the region include:

[0011] For natural waterways, the flow velocity at each location is measured at a preset first interval; for waterways on artificial underlying surfaces, the flow velocity at each location is measured at a preset second interval.

[0012] Based on the flow velocity at each location in the waterway, calculate the time period when the water flow at each location converges into the dam, starting from time 0.

[0013] If the water flows into the dam at the same time, their corresponding locations will be assigned to the same area.

[0014] Preferably, the calculation is based on the precipitation in each region. Increased water storage in dams due to rainfall during a given period ,include:

[0015] Within each time period, each region is divided into sub-regions based on precipitation.

[0016] Within each time period, the total precipitation for each region is calculated based on the precipitation in each sub-region, using the following formula:

[0017] ;

[0018] In the formula, The region number, for The sub-region number within the region. for During the period Total precipitation in the region for During the period Precipitation in sub-regions for The area of ​​the sub-region for During the period The number of sub-regions within the region;

[0019] Calculate water storage The formula is as follows:

[0020] ;

[0021] In the formula, for Regional precipitation loss rate ;

[0022] The prediction calculation The water level in the dam at the end of the period The formula is as follows:

[0023] ;

[0024] In the formula, This represents the initial water volume of the dam. This refers to the runoff flowing into the dam under non-precipitation conditions. The duration of a single time period. for The amount of water discharged from the dam during the specified period.

[0025] Furthermore, it also includes precipitation loss rate. The method for determining this is as follows:

[0026] Soil moisture meters were installed in each area to measure soil moisture. ;

[0027] For each region, construct the precipitation loss rate. With soil moisture Functional relationship The formula is as follows:

[0028] ;

[0029] In the formula, for The saturated hydraulic conductivity of the soil in the region, for The soil moisture saturation value in the region, for The wetting peak suction of the soil in the region;

[0030] Constructing a sample set The formula is as follows:

[0031] ;

[0032] ;

[0033] ;

[0034] In the formula, For sample set The serial number of the sample. For sample set The number of samples in the middle; for Samples from the area, including soil moisture And each Precipitation in sub-regions ; For the first The amount of water stored in the dam due to precipitation in each sample;

[0035] Humidity saturation value for each area and soil moisture and every Precipitation in sub-regions As input, with the corresponding water storage volume For the target output, perform few-sample meta-learning:

[0036] From saturated hydraulic conductivity and humidification peak suction power Starting with the initial values, calculate the loss function for the sample set. The formula is as follows:

[0037] ;

[0038] In the formula, It is the L2 norm. To make the first The output obtained by learning from a few samples;

[0039] Minimize the loss function and adjust to confirm the saturated hydraulic conductivity of each region. and humidification peak suction power ;

[0040] For each region, based on real-time soil moisture Through functional relationships Calculate precipitation loss rate .

[0041] Preferably, based on the water storage volume Real-time adjustment of flood discharge volume, specifically including:

[0042] If the predicted water storage volume exists... If the maximum warning value is exceeded, all valves will be opened to release floodwater;

[0043] If the predicted water storage volume exists... If the water level is below the minimum warning value, all valves will be closed to store water.

[0044] Compared with existing technologies, the technical solution provided by this invention considers the time delay relationship between precipitation and the increase in water storage in each region after division, divides the upstream basin of the dam, and makes full use of spatiotemporal scale information such as precipitation area, dam location, precipitation time, and water storage increase time. This can further realize real-time dynamic prediction of dam water storage, which helps to carry out precise and reasonable flood discharge control of the dam and realize accurate early warning and rapid response to water disasters. Attached Figure Description

[0045] Figure 1 This is a schematic flowchart of a dam flood discharge control method in one embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram of the result of watershed division in one embodiment of the present invention.

[0047] Figure 3 This is a precipitation distribution map of a watershed during a certain period in one embodiment of the present invention.

[0048] Figure 4 This is a schematic diagram showing the result of dividing a watershed into sub-regions during a certain period in one embodiment of the present invention. Detailed Implementation

[0049] The technical solutions provided by the present invention will be further described in detail below with reference to the embodiments and accompanying drawings.

[0050] like Figure 1 As shown, this embodiment provides a dam discharge control method based on precipitation. First, a satellite map of the upstream area of ​​the dam to be controlled is acquired using the BeiDou Navigation Satellite System or GPS. Rivers are identified from the satellite map, and watersheds between the river and adjacent rivers are drawn based on contour lines. All watersheds are connected on the satellite map to determine the upstream watershed. To facilitate subsequent regional division, the satellite map of the watershed is simplified to a line drawing, retaining only icons indicating the dam's location, lines connecting the watersheds, and the river's channel trajectory.

[0051] Delineating the upstream basin of a dam requires combining on-site data collection, including: determining the waterway trajectory based on satellite maps; measuring the flow velocity at each location along the waterway at predetermined intervals; calculating the time period at which the water flows into the dam, starting from time 0, based on the flow velocity at each location; and assigning the corresponding locations to the same area if the water flows into the dam during the same time period.

[0052] Specifically, such as Figure 2As shown, the waterway trajectory of each river in the entire basin is determined based on satellite maps. For the natural waterway portion of each river, the flow velocity at each location is measured at a preset first interval. For the waterway portion of each river on the artificial underlying surface (a river channel with an artificially modified bottom and slope, such as a concrete bottom, a stone bottom, or an artificial canal), the flow velocity at each location is measured at a preset second interval. Generally, to balance data accuracy and workload, the first interval is set to 5 to 10 kilometers. If the dam is located in a relatively flat plain in the middle and lower reaches of the river, the first interval can be set to 10 kilometers or even larger; if the dam is located in a relatively steep mountainous area in the middle and upper reaches of the river, the first interval can be set to 5 kilometers or even smaller; if the dam is located at the boundary between plains and mountains, the first interval can be adaptively adjusted according to the rate of elevation change of each river channel. Due to the varying roughness of the underlying surface, the flow velocity in natural waterways is generally 60% to 90% of that in waterways on artificial underlying surfaces. Therefore, under the same slope, a first interval distance is set with reference to the natural river channel, and a second interval distance is set to 60% to 90% of the first interval distance. If the length of the waterway on the artificial underlying surface is between 1 and 3 kilometers, the flow velocity in the middle of that waterway is measured; if the length of the waterway on the artificial underlying surface is less than 1 kilometer, it is ignored. For any waterway, after measuring the flow velocity and distance from the dam at each location, the time required for the water to flow from the current location into the dam can be calculated. This embodiment adopts a staged prediction of the change in the dam's water volume. Correspondingly, each waterway trajectory can be divided into zones according to the time required for the water to flow from each location into the dam, and each location with the same required time is assigned to the same area. That is, based on the flow velocity at each location in the waterway, the time period when the water flow at each location enters the dam is calculated with time 0 as the starting point. Locations with the same time period when the water flow enters the dam are divided into the same area. A total of 6 areas are divided in the entire watershed, including area 1, area 2, area 3, area 4, area 5 and area 6.

[0053] like Figure 3 As shown, precipitation at a certain moment gradually decreases from north to south, with a gradient of 10 mm / h, decreasing from 50 mm / h in the precipitation center in the north of the basin to 10 mm / h at the precipitation edge. It is evident that there is a significant discrepancy between the precipitation distribution map and the defined regions. Furthermore, because real-time precipitation distribution is constantly changing, these discrepancies also change continuously and are difficult to describe mathematically using linear or simple mapping relationships. Therefore, as... Figure 4As shown, to ensure accurate precipitation statistics, this embodiment combines the regional division results and precipitation distribution map to further divide each region in real time, resulting in more sub-regions. Taking the divided region 4 as an example, after further division, it forms 5 sub-regions with real-time precipitation of 50 mm / h, 40 mm / h, 30 mm / h, 20 mm / h, and 10 mm / h. Region 1, since its entire area falls within the 20 mm / h precipitation range, is only divided into 1 sub-region. Regions 2, 3, 5, and 6 are divided into 3, 5, 4, and 3 sub-regions, respectively. In summary, a total of 21 sub-regions are divided within the entire watershed for the current time period.

[0054] After dividing the region into sub-regions, the total precipitation for each region is calculated based on the precipitation in each sub-region within each time period, using the following formula:

[0055] ;

[0056] In the formula, The region number, for The sub-region number within the region. for During the period Total precipitation in the region for During the period Precipitation in sub-regions for The area of ​​the sub-region for During the period The number of sub-regions within the region. To ensure the comprehensiveness and accuracy of precipitation statistics for the entire region and all time periods, the length of the sliding window for the forecast period needs to be determined based on the number of sub-regions. That is, the forecast period should be based on the length of the number of sub-regions.

[0057] Based on the total precipitation in each region within each time period, and taking into account the time required for precipitation in each region to flow into the dam, calculate... Increased water storage in dams due to rainfall during a given period The formula is as follows:

[0058] ;

[0059] In the formula, for Regional precipitation loss rate Since regions are defined based on the time required for runoff to flow into reservoirs from different locations, the water flowing into reservoirs within the same time period represents rainfall from different regions at different times. For the current time period, the water flowing into the reservoir from the nearest region represents the rainfall for that current time period, while the water flowing into the reservoir from the farthest region represents the rainfall for the time period corresponding to the start of this forecast. Therefore, it's easy to understand that, reflected in the parameter subscripts, the sum of the region number and the time period number is fixed at [value missing]. (that is ), for Total precipitation in the two regions during the period for During the period Total precipitation in the region.

[0060] Predictive calculation The water level in the dam at the end of the period The formula is as follows:

[0061] ;

[0062] In the formula, This represents the initial water volume of the dam. This refers to the runoff flowing into the dam under non-precipitation conditions. The duration of a single time period. for The amount of water discharged from the dam during the specified period.

[0063] Among them, the precipitation loss rate is a parameter that varies over time due to the previous total precipitation and the surface type in the region. In this embodiment, it is updated in real time using the following method.

[0064] Soil moisture meters were installed in each area to measure soil moisture. Due to the large area, therefore throughout... Multiple soil moisture meters were evenly distributed throughout the area, and the average value was calculated to obtain... Soil moisture in the area The soil moisture content of the hardened surface was set at 100%.

[0065] Soil moisture meters were installed in each area to measure soil moisture. For each region, a precipitation loss rate was constructed. With soil moisture Functional relationship The formula is as follows:

[0066] ;

[0067] In the formula, for The saturated hydraulic conductivity of the soil in the region, for The soil moisture saturation value in the region, for The soil moisture peak suction within the region. Soil moisture meters can be resistive, capacitive, time-domain reflectometry (TDAR), frequency-domain reflectometry (FDR), negative pressure, or piezoresistive. Furthermore, based on the soil moisture meter, the soil moisture saturation value is measured using either the constant head method or the variable head method.

[0068] Constructing a sample set The formula is as follows:

[0069] ;

[0070] ;

[0071] ;

[0072] In the formula, For sample set The serial number of the sample. For sample set The number of samples in the middle; for Samples from the area, including soil moisture And each Precipitation in sub-regions ; For the first The amount of water stored in dams due to precipitation in each sample.

[0073] Humidity saturation value for each area and soil moisture and every Precipitation in sub-regions As input, with the corresponding water storage volume For the target output, perform few-sample meta-learning:

[0074] Based on the water storage volume mentioned above The calculation formula, from saturated hydraulic conductivity and humidification peak suction power Starting with the initial values, calculate the loss function for the sample set. The formula is as follows:

[0075] ;

[0076] In the formula, It is the L2 norm. To make the first The output obtained by learning from a few samples.

[0077] Minimize the loss function and adjust to confirm the saturated hydraulic conductivity of each region. and humidification peak suction power For each region, based on real-time soil moisture... Through functional relationships Calculate precipitation loss rate .

[0078] Optionally, few-shot meta-learning can be performed using Model-Agnostic Meta-Learning (MAML) algorithms or Memory-Augmented Neural Networks (MANN).

[0079] Finally, based on the water storage volume Real-time adjustment of flood discharge volume, specifically including: if there is a predicted water storage volume... If the water level exceeds the highest warning level, all valves will be opened to release floodwater; if the predicted water storage capacity exists... If the water level is below the minimum warning level, all valves will be closed to store water. Furthermore, if the predicted water level rises too rapidly over a given period, the discharge can be increased in advance to ensure sufficient reserves to cope with subsequent flood peaks, thereby further reducing the possibility of dam failure and other water-related disasters.

[0080] As can be seen from the above embodiments and accompanying drawings, compared with the prior art, the technical solution provided by the present invention considers the time delay relationship between precipitation in each region and the increase in water storage in the dam after division, divides the upstream basin of the dam, and makes full use of spatiotemporal scale information such as precipitation area, dam location, precipitation time, and water storage increase time. This can further realize real-time dynamic prediction of dam water storage, which helps to carry out precise and reasonable flood discharge control of the dam and realize accurate early warning and rapid response to water disasters.

[0081] Furthermore, by further dividing the region into sub-regions based on precipitation, the total precipitation in each region at each time period can be accurately calculated, improving the accuracy of dam water storage prediction. In the process of updating precipitation loss rate, soil moisture is detected in real time by soil moisture meter, and considering the diversity of underlying surface conditions, the soil moisture of hardened ground is set at 100%, which can further improve the accuracy of dam water storage prediction, which is conducive to accurate early warning and rapid response to water disasters.

Claims

1. A method for controlling dam flood discharge based on precipitation, characterized in that, include: Determine the upstream watershed of the dam and divide the watershed into... One region; Calculated based on precipitation in each region Increased water storage in dams due to rainfall during a given period ; Based on water storage Predictive calculation The water level in the dam at the end of the period ; Based on water storage Adjust the flood discharge volume in real time.

2. The dam flood discharge control method based on precipitation as described in claim 1, characterized in that, The division of the watershed into The specific methods for dividing the region include: For natural waterways, the flow velocity at each location is measured at a preset first interval; for waterways on artificial underlying surfaces, the flow velocity at each location is measured at a preset second interval. Based on the flow velocity at each location in the waterway, calculate the time period when the water flow at each location converges into the dam, starting from time 0. If the water flows into the dam at the same time, their corresponding locations will be assigned to the same area.

3. The dam flood discharge control method based on precipitation as described in claim 2, characterized in that, The calculation is based on the precipitation in each region. Increased water storage in dams due to rainfall during a given period ,include: Within each time period, each region is divided into sub-regions based on precipitation. Within each time period, the total precipitation for each region is calculated based on the precipitation in each sub-region, using the following formula: ； In the formula, The region number, for The index of the sub-region within the region. for During the period Total precipitation in the region for During the period Precipitation in sub-regions for The area of ​​the sub-region for During the period The number of sub-regions within the region; Calculate water storage The formula is as follows: ； In the formula, for Regional precipitation loss rate ; The prediction calculation The water level in the dam at the end of the period The formula is as follows: ； In the formula, This represents the initial water volume of the dam. This refers to the runoff flowing into the dam under non-precipitation conditions. The duration of a single time period. for The amount of water discharged from the dam during the specified period.

4. The dam flood discharge control method based on precipitation as described in claim 3, characterized in that, It also includes precipitation loss rate The method for determining this is as follows: Soil moisture meters were installed in each area to measure soil moisture. ; For each region, construct the precipitation loss rate. With soil moisture Functional relationship The formula is as follows: ； In the formula, for The saturated hydraulic conductivity of the soil in the region, for The soil moisture saturation value in the region, for The wetting peak suction of the soil in the region; Constructing a sample set The formula is as follows: ； ； ； In the formula, For sample set The serial number of the sample. For sample set The number of samples in the middle; for Samples from the area, including soil moisture And each Precipitation in sub-regions ; For the first The amount of water stored in the dam due to precipitation in each sample; Humidity saturation value for each area and soil moisture and every Precipitation in sub-regions As input, with the corresponding water storage volume For the target output, perform few-sample meta-learning: From saturated hydraulic conductivity and humidification peak suction power Starting with the initial values, calculate the loss function for the sample set. The formula is as follows: ； In the formula, It is the L2 norm. To make the first The output obtained by learning from a few samples; Minimize the loss function and adjust to confirm the saturated hydraulic conductivity of each region. and humidification peak suction power ; For each region, based on real-time soil moisture Through functional relationships Calculate precipitation loss rate .

5. A dam flood discharge control method based on precipitation as described in any one of claims 1 to 4, characterized in that, According to the water storage Real-time adjustment of flood discharge volume, specifically including: If the predicted water storage volume exists... If the maximum warning value is exceeded, all valves will be opened to release floodwater; If the predicted water storage volume exists... If the water level is below the minimum warning value, all valves will be closed to store water.

Images