Flood storage and detention area start decision-making method, system and equipment based on swarm intelligent optimization algorithm and medium

Abstract

The invention discloses a flood storage and detention area start decision-making method, system, equipment and medium based on a swarm intelligent optimization algorithm, and the method comprises the steps: collecting hydrological engineering, spatial geography and historical case data through a multi-source data integration architecture, and forming structured data with a unified time scale after cutting, abnormal value processing and standardization; and the uncertainty scene generation architecture is used for flood scene prejudgment. Outputting a flood routing trend and an uncertainty interval based on an LSTM deep learning model, and automatically generating various disturbance scenes of flood peak flow, peak present time and downstream risk levels; and the swarm intelligence optimization decision-making architecture is used for solving an optimal scheme. All scenes are traversed through an Arctic Halorot optimization algorithm, the total flood diversion cost is minimized under the conditions of hard constraint and soft constraint, and a flood diversion critical point and an optimal starting scheme are output through dynamic iteration. According to the method, the dynamic adaptability and the decision robustness are remarkably improved, the multi-target balance and the business landing are considered, and the flood control risk and the disaster loss are effectively reduced.

Description

TECHNICAL FIELD

[0001] The application relates to a flood storage area enabling decision, in particular to a flood storage area enabling decision method, system, equipment and medium based on a swarm intelligence optimization algorithm. BACKGROUND

[0002] The flood storage area is the last line of defense in the flood control system of a basin, and the core of its enabling decision is to accurately balance the two conflicting goals of "flood control safety guarantee" and "regional loss control". The complexity of this decision lies in the fact that, on the one hand, excess flood is stored to avoid major disasters such as downstream embankment collapse, and on the other hand, the property loss and personnel relocation cost in the flood storage area are minimized. With the intensification of global climate change, the randomness of flood evolution caused by extreme weather events is further increased, which further increases the difficulty of decision-making. As the "bottom card" of the flood control safety system of a basin, the scientific and accurate enabling of the flood storage area is the key to guaranteeing flood control safety and minimizing disaster losses.

[0003] At present, the traditional flood storage area flood diversion decision mode mainly relies on fixed hydrological thresholds and artificial experience, and there are two core challenges in dealing with uncertain flood scenarios: first, the dynamic adaptability and robustness are insufficient. Static thresholds are difficult to respond to dynamic evolution of flood and changes in future risks, which may lead to decision lag or misallocation of flood; at the same time, the traditional method lacks the ability to analyze the disturbance of key parameters, and cannot effectively deal with various uncertain scenarios, so the decision robustness is poor. Second, it lacks intelligent optimization ability. The existing mode cannot efficiently integrate multi-source data, and cannot quickly and automatically generate an optimal flood diversion scheme that takes into account the interests of upstream and downstream, and the decision-making process still highly depends on expert experience, which cannot meet the modernization needs of precise regulation of the basin flood control. SUMMARY

[0004] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art and provide a flood storage area enabling decision method, system, equipment and medium based on a swarm intelligence optimization algorithm, aiming to solve the intelligent flood diversion decision problem under uncertain flood scenarios.

[0005] The technical scheme of the present application is:

[0006] A flood storage area intelligent enabling decision method based on a swarm intelligence optimization algorithm, comprising the following steps:

[0007] Step 1, the range of the basin where the flood storage area is located is determined, and multi-source data in the basin is collected, all data containing multiple time scales, and the data time scales are kept uniform;

[0008] Step 2, data clipping: clip all types of data collected in step 1 according to the time dimension, so that the start and end times of all hydro-meteorological data and spatial data are consistent, ensuring that the data match in time span;

[0009] Step 3: Outlier removal and completion of the data obtained from Step 2, and data standardization processing;

[0010] Step 4: Flood evolution trend prediction, build LSTM deep learning model, use historical hydrological data and real-time monitoring data after preprocessing in Step 3 as training samples, input model training until convergence, use the trained model to output the change trend of upstream and downstream water level / flow in T hours, and generate the forecast uncertainty interval of flood peak flow and peak time;

[0011] Step 5: Uncertainty scenario generation, automatically generate N types of key parameter disturbance scenarios based on the above uncertainty interval, including but not limited to different over-forecast situations of flood peak flow, advance or delay of peak time, and changes in risk level of downstream protection objects;

[0012] Step 6: Use sliding window method to block process the time series data after preprocessing in Step 3, thus creating input and output pairs of the algorithm;

[0013] Step 7: Build group intelligence optimization algorithm decision model, determine decision variables and constraint conditions;

[0014] Step 8: Based on the optimization algorithm decision model determined in Step 7, set parameters: population size M, maximum evaluation times FES;

[0015] Step 9: Multi-scenario optimization solution: input the uncertainty scenarios generated in Step 5 and the decision sample set divided in Step 6 into the algorithm, the algorithm traverses all scenarios, and solves the optimal solution corresponding to the minimum value of the objective function under the premise of meeting the constraint conditions in Step 7;

[0016] Step 10: Dynamic iterative optimization and output of flood diversion scheme: real-time data linkage receives monitoring data of the basin, when the deviation between measured value and predicted value is large, automatically trigger Steps 4 to 9, recalculate and iteratively update the flood diversion scheme, and finally output the structured decision results.

[0017] Further: in Step 1, it mainly includes hydrological engineering data, spatial geographic data and historical case data. Among them, the hydrological engineering data covers the measured water level and flow of upstream and downstream, flood process prediction data, maximum design flow of flood diversion gate and design water level of embankment, etc. Spatial geographic data includes digital elevation model of flood storage area, flood diversion gate and river vector data, and distribution information of downstream core protection objects; historical case data involves hydrological parameters, loss statistics and decision result records in historical flood diversion events.

[0018] Further: in step 3, the completed numerical data (such as water level, flow, loss value) is standardized to the interval (0, 1) to eliminate the influence of dimension, and the standardization formula is as follows:

[0019]

[0020] In the formula: X std is the standardized data; X is the original sample data; X min is the minimum value of this type of data; X max is the maximum value of this type of data.

[0021] Further: in step 6, the sliding window method is as follows:

[0022] First, set the front pointer (the starting position of the window) and the back pointer (the end position of the window), and the interval between the pointers is defined as the time step. This time step is determined by the number of data points contained, and the interval between the back pointer and the prediction target is defined as the prediction period.

[0023] Then, the front and back pointers move from the beginning to the end of the data sequence step by step, and a new data block is generated every time a data point is moved, until the back pointer exceeds the end of the data sequence. Finally, this process will produce a sample set for the training and decision of the swarm intelligence optimization algorithm.

[0024] Further: in step 7, define the decision variable: set the storage and detention basin activation state X i (X i =1 indicates that the i-th storage and detention basin is activated, X i =0 indicates that it is not activated), the dynamic value of the flood diversion flow Q (unit: m³ / s), the duration of the flood diversion T d (unit: h);

[0025] Further: in step 7, construct the objective function: take the minimum total cost of flood diversion as the core goal, and the calculation formula is as follows:

[0026]

[0027] In the formula: C down is the downstream disaster loss (calculated based on the inundation range and the value of the protected object); C in is the loss in the storage and detention basin (farmland / property loss); C trans is the personnel transfer cost; C eng is the engineering risk cost (such as repair cost of over-flowing outlet gate);

[0028] Further: in step 7, set hierarchical constraint conditions: hard constraints (must be met, otherwise impose a loss of double penalty coefficient): upstream embankment water level ≤ design water level, no submergence risk for downstream core protection objects, flood discharge through flood gate ≤ maximum design value; soft constraints (optimization satisfaction): flood storage capacity of flood storage area ≤ total storage capacity, no repeated flood diversion for adjacent flood storage areas.

[0029] Further: in step 8, in terms of algorithm selection, adopt the Arctic Sea Parrot optimization algorithm as the swarm intelligence optimization algorithm, solve through multi-scene optimization, and output the dynamic flood diversion critical point, over-limit tolerance space, and optimal flood diversion scheme.

[0030] An intelligent activation decision system for flood storage and detention areas based on swarm intelligence optimization algorithm, characterized in that: the system can be used to implement the intelligent activation decision method for flood storage and detention areas based on swarm intelligence optimization algorithm, specifically comprising:

[0031] Data collection module: responsible for collecting hydrological engineering data, spatial geographic data, and historical case data in the study basin, supporting import and storage of multiple formats of data, and ensuring uniformity of data time scale.

[0032] Data preprocessing module: includes data cropping submodule, outlier processing submodule, and standardization submodule; the data cropping submodule realizes uniformity of data time span, the outlier processing submodule completes outlier elimination and completion, and the standardization submodule standardizes the data to the interval (0, 1).

[0033] Uncertainty scenario generation module: based on the LSTM deep learning model, outputs the flood evolution trend and uncertainty interval, automatically generates multiple types of key parameter disturbance scenarios, and covers multiple unfavorable flood scenarios.

[0034] Data block module: uses the sliding window method to divide time series data into blocks, sets time step and prediction period parameters, generates algorithm input and output pairs, and supports automatic division and storage of sample sets.

[0035] Swarm intelligence optimization decision module: system core module, including model construction submodule and algorithm solving submodule; the model construction submodule defines decision variables, objective function, and hierarchical constraint conditions; the algorithm solving submodule uses the Arctic Sea Parrot optimization algorithm to traverse the uncertainty scenarios to solve the optimal flood diversion scheme, and outputs the dynamic flood diversion critical point and over-limit tolerance space.

[0036] Dynamic iterative optimization module: interfaces with the real-time monitoring network of the basin, receives real-time data in real time, judges whether the data deviation exceeds the threshold, and automatically triggers scheme recalculation and iterative update.

[0037] The decision result output module outputs the dynamic threshold table, the flood diversion operation instruction and the risk prompt in the form of a table and text, supports result export, and is suitable for a flood control scheduling business scenario.

[0038] A computer device comprises a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor executes the computer program to implement the steps of the group intelligence optimization algorithm-based impoundment and detention basin enabling decision method.

[0039] A computer readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the group intelligence optimization algorithm-based impoundment and detention basin enabling decision method.

[0040] The present application has the following advantages:

[0041] 1. The present application fuses deep learning and group intelligence optimization algorithm, constructs an intelligent regulation and control system coupled with an uncertainty scenario, accurately determines a flood diversion critical point and an over-limit tolerance space, and, in combination with a multi-objective constraint balance logic, realizes an iteration optimization of a flood diversion scheme based on multi-source data driving, provides intelligent support for scientific and accurate enabling of an impoundment and detention basin, and promotes intelligent breakthroughs in basin flood control scheduling.

[0042] 2. The present application has dynamic adaptability and significantly improved robustness: the group intelligence optimization algorithm traverses an uncertainty scenario to generate a dynamic flood diversion threshold, replacing a traditional fixed threshold, avoiding decision lag when the upstream does not exceed the threshold but the downstream is affected, preventing excessive loss when the future risk is relieved after exceeding the threshold, and adapting to multiple flood scenarios, with improved decision robustness compared with traditional modes.

[0043] 3. The present application achieves intelligent balance of multiple objectives: through hierarchical constraint design, in combination with a target function of "minimum total flood diversion cost", the algorithm can automatically balance the safety of downstream core protection objects, loss control in the impoundment and detention basin, and multiple objectives such as engineering operation risk. This optimization not only ensures the safety of core protection objects, but also maximally reduces the loss in the area.

[0044] 4. The present application considers both decision efficiency and accuracy: the system is connected with a real-time monitoring network of the basin, and when the deviation between the measured water level or flow and the forecast exceeds the threshold, the system can automatically trigger the whole process of flood evolution prediction and scheme optimization iteration. This real-time response mechanism significantly improves the decision speed and can timely adapt to the dynamic changes of the flood, thereby reducing the risks caused by decision lag.

[0045] 5、The service of the application has strong landing: the output results include a structured dynamic threshold table, a flood diversion operation instruction and a risk prompt, which can be directly exported and connected to the dispatching business system. At the same time, the system module design conforms to the actual dispatching logic, forming a closed loop from data collection to scheme output, improving the business adaptation efficiency and meeting the actual needs of flood control and dispatching.

[0046] 6、The application constructs an intelligent decision-making system that integrates multi-source data and has dynamic optimization capability, which is the key to improving the scientific nature and adaptability of the decision-making of the flood storage and detention area, easy to implement and has good economic benefits.

[0047] 7、The application is supported by the National Key R&D Program of the Ministry of Science and Technology of the People's Republic of China "Major Natural Disasters and Public Safety" Key Special Project "Key Technologies for Three-dimensional Perception and Intelligent Decision-making of Flood Storage and Detention Areas", project number [2024YFC3012300]. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The figure is the overall system architecture and data flow diagram of the embodiment of the application.

[0049] Figure 2 The figure is the group intelligence optimization algorithm decision-making model and solution flowchart of the embodiment of the application. DETAILED DESCRIPTION

[0050] Embodiment one: see Figure 1 -- Figure 2 In the figure,

[0051] An intelligent decision-making method for the activation of a flood storage and detention area based on a group intelligence optimization algorithm, which realizes accurate decision-making under uncertain flood scenarios through a closed-loop process of "multi-source data integration-uncertainty scenario generation-group intelligence optimization-decision result output", specifically including the following steps:

[0052] Step 1, determine the range of the basin where the flood storage and detention area is located, collect multi-source data in the basin, all data need to contain multiple time scales, and the data time scales are kept uniform;

[0053] Step 2, data cropping: crop all types of data collected in step 1 according to the time dimension, so that the start and end times of all hydro-meteorological data and spatial data are consistent, ensuring that the data match in time span;

[0054] Step 3, remove outliers and complete the data obtained in step 2, and perform data standardization processing;

[0055] Step 4, flood trend prediction, build LSTM deep learning model, use historical hydrological data and real-time monitoring data as training samples, input model training until convergence, use trained model to output future T hours of upstream and downstream water level / flow trend, and generate flood peak flow and peak time prediction uncertainty interval;

[0056] Step 5, uncertainty scenario generation, automatically generate N types of key parameter disturbance scenarios based on the above uncertainty interval, including but not limited to different over-forecast situations of flood peak flow, advance or delay of peak time, and changes in downstream protection object risk level;

[0057] Step 6, use sliding window method to block process the time series data preprocessed in step 3, to create input and output pairs of the algorithm;

[0058] Step 7, build group intelligence optimization algorithm decision model, determine decision variables and constraints;

[0059] Step 8, based on the optimization algorithm decision model determined in step 7, set parameters: population size M, maximum evaluation times FES;

[0060] Step 9, multi-scenario optimization solution: input uncertainty scenarios generated in step 5 and decision sample set divided in step 6 into the algorithm, the algorithm traverses all scenarios, and solves the minimum value of the objective function corresponding to the optimal solution under the premise of meeting the constraints in step 7;

[0061] Step 10, dynamic iterative optimization and output of flood diversion scheme: real-time data linkage receives monitoring data of the basin, when the deviation between measured value and predicted value is large, automatically trigger steps 4 to 9, recalculate and iteratively update the flood diversion scheme, and finally output structured decision results.

[0062] Multi-source data mainly includes hydrological engineering data, spatial geographic data and historical case data. Among them, hydrological engineering data covers measured water level and flow, flood process prediction data, maximum design flow of flood diversion gate, and design water level of embankment; spatial geographic data includes digital elevation model of flood storage area, flood diversion gate and river vector data, and distribution information of downstream core protection objects; historical case data involves hydrological parameters, loss statistics and decision result records in historical flood diversion events. In actual application, all data time scales should be unified.

[0063] The cropped and completed numerical data (such as water level, flow, and loss value) is standardized to the interval (0, 1) to eliminate the dimension effect, and the standardization formula is as follows:

[0064]

[0065] In the formula: X std is the standardized data; X is the original sample data; X min is the minimum value of the data; X max is the maximum value of the data.

[0066] The 3σ principle is adopted for outlier rejection; the linear interpolation method is adopted for data completion, and for the missing short-time data, the effective data at adjacent time is used to fill in.

[0067] The LSTM deep learning model controls the information time sequence flow through the synergistic effect of the forgetting gate, the input gate and the output gate, and the core formula is as follows:

[0068] Forgetting gate:

[0069]

[0070] Input gate:

[0071]

[0072]

[0073] Cell state update:

[0074]

[0075] Output gate:

[0076] ,

[0077] Predicted output:

[0078]

[0079] In the formula: is the corresponding weight matrix; is the corresponding bias term; , is the hidden layer output and cell state at t-1 time; is the hyperbolic tangent activation function; is the input data at t time; is the predicted value at t time.

[0080] The specific steps of the sliding window method are as follows:

[0081] First, a front pointer (the starting position of the window) and a back pointer (the ending position of the window) are defined. The interval between the pointers is defined as the time step, determined by the number of data points included. The interval between the back pointer and the prediction target is defined as the prediction period. Next, the front and back pointers move progressively from the beginning to the end of the data sequence, generating a new data block for each data point moved, until the back pointer exceeds the end of the data sequence. Ultimately, this process produces a sample set for training and decision-making in swarm intelligence optimization algorithms.

[0082] The decision variables are defined as follows: Let X be the activation status of the flood storage and detention area. i (X) i =1 indicates that the i-th flood storage and detention area is activated, X i =0 indicates that it is not enabled), dynamic value of flood diversion flow Q (unit: m³ / s), flood diversion duration T d (Unit: h);

[0083] The objective function focuses on minimizing the total cost of flood diversion, and the calculation formula is as follows:

[0084] In the formula: C down Downstream disaster losses (calculated based on the inundation area and the value of the protected objects); C in For losses within the flood storage and detention area (farmland / property loss); C trans For personnel relocation costs; C eng For engineering risk costs (such as maintenance costs due to overflow exceeding the design limit of the valve);

[0085] Constraints are divided into hard constraints and soft constraints: Hard constraints (must be met, and a penalty coefficient of doubled loss will be imposed if they are violated) include upstream dike water level ≤ design water level, no risk of flooding of downstream core protected objects, and flood diversion gate flow ≤ maximum design value; Soft constraints (optimization required) include flood storage capacity of flood storage and detention areas ≤ total reservoir capacity, and no duplicate flood diversion when adjacent flood storage and detention areas are activated.

[0086] The swarm intelligence optimization algorithm selected is the Arctic puffin optimization algorithm, which consists of two main phases: aerial flight (exploration phase) and underwater foraging (development phase). During aerial flight, Arctic puffins typically fly in formation or groups at low altitudes to capture underwater food resources. Their position update equation is as follows:

[0087]

[0088]

[0089]

[0090] In the formula, r is a random integer between 1, 2, ..., M-1, excluding i. is the i-th candidate solution in the current population; is a candidate solution randomly selected from the current population; L(D) is a random number generated by Levy flight; D is the dimension; is a random number following the standard normal distribution.

[0091] Diving is a key strategy for Arctic terns when hunting, as they quickly change direction to speed up food capture:

[0092]

[0093]

[0094] In this strategy, Arctic terns adjust displacement by a speed factor S to adapt to different foraging needs. To obtain the best results, the algorithm combines the candidate positions of the two stages, and the combination process is as follows:

[0095]

[0096]

[0097]

[0098] In the formula, sort is the new According to the fitness value from small to large, the new population is selected .

[0099] During underwater foraging, Arctic terns often use the aggregation foraging strategy in their foraging behavior. The position update equation of this cooperative foraging behavior is as follows:

[0100]

[0101] Where the cooperation factor F = 0.5 adjusts the foraging behavior of Arctic terns. Variables r1, r2, r3 are random integers between 1 and M-1. As the hunting progresses, the terns may feel that the food resources in the current area are exhausted and need to change their position to find more underwater food. The position update equation is as follows:

[0102]

[0103]

[0104] In the formula, FES represents the maximum number of evaluations, and fe represents the current number of evaluations. rand is a random number. When Arctic terns find predators nearby, they will use the evasion strategy, and the position update equation is as follows:

[0105]

[0106] In the formula, the random number β is uniformly distributed between 0 and 1. In summary, the parrot adopts multiple strategies when foraging underwater, and the algorithm generates a new solution by combining candidate positions in the above strategies to obtain the best result. The combination equation is described as follows:

[0107]

[0108]

[0109]

[0110] The scene optimization solution process is as follows: first, randomly generate M groups of decision variables (X i , Q, T d ), and filter the feasible solutions according to the hard constraint conditions; then, for each uncertainty scenario, calculate the objective function value Z of the feasible solution; update the population position and filter the global optimal individual through the Arctic parrot optimization algorithm; when the maximum evaluation number is reached, stop iteration, and finally output the dynamic flood diversion critical point, overrun tolerance space and optimal flood diversion scheme corresponding to each scene. The detailed logic and key steps of the above whole process are visualized as shown in Figure 2 .

[0111] In step 10, the real-time data linkage frequency is every 30 minutes. By connecting the real-time monitoring network of the basin (such as water level station, flow station, and rainfall station), the measured data is received, the deviation rate of the measured value and the predicted value in step 4 is calculated, and when the deviation is large, steps 4-9 are automatically triggered to recalculate and iteratively update the flood diversion scheme. The output results include dynamic threshold table, flood diversion operation instructions and risk prompt, support Excel, PDF format export, and can be directly connected to the flood control dispatching business system.

[0112] Example: Taking a typical storage and detention basin in a certain basin as the research object, the total storage capacity V total of the storage and detention basin is 120 million m³, the maximum design flow capacity Q max of the flood diversion gate is 1900 m³ / s, and the core protection objects downstream are towns and traffic trunks.

[0113] The multi-source data of the basin from 2018 to 2023 (the time scale is unified to 1 hour) is collected, and after cutting, abnormal value processing and standardization, it is input into the LSTM model for training, and the future 3-hour flood evolution trend and flood peak flow prediction uncertainty interval [1980, 2420] m³ / s is output; based on the interval, 5 typical scenes are generated, and the most unfavorable scene is “flood peak exceeds the prediction by 20% (2640 m³ / s) + peak appears 1 hour in advance + downstream high risk”; 25 sample sets are generated by using the sliding window method, and are input into the Arctic sea bird optimization algorithm (population size M=80, maximum evaluation number FES=1000) for optimization, and the optimal scheme is obtained: 2 flood storage areas are started, the flood diversion flow is 1850 m³ / s, the duration is 7 hours, and the total flood diversion cost is 880,000 yuan; the dynamic flood diversion critical point is the water level H=61.8 m and the flow Q=2650 m³ / s.

[0114] The scheme of the application is compared with the traditional fixed threshold scheme (fixed flood diversion trigger water level H fix =62.0m, flood diversion flow Q fix =1900m³ / s, lasting 8 hours), and the evaluation indexes are as follows:

[0115] Evaluation index Invention scheme Traditional fixed threshold scheme Flood diversion total cost (ten thousand yuan) 880 1230 Downstream core area inundation area (km²) 0 2.8 Decision response time (minutes) 12 55 Nash efficiency coefficient (NSE) 0.96 0.82

[0116] Embodiment two provides an intelligent starting decision system for a flood storage area based on a swarm intelligence optimization algorithm, which can be used to implement the intelligent starting decision method for a flood storage area based on a swarm intelligence optimization algorithm, and specifically comprises:

[0117] A data collection module is responsible for collecting hydrological engineering data, spatial geographic data and historical case data in the research basin, supports the import and storage of multi-format data (such as Shp, CSV and TIF), and ensures the uniformity of the data time scale.

[0118] A data preprocessing module includes a data cutting submodule, an abnormal value processing submodule and a standardization submodule; the data cutting submodule realizes the uniformity of the data time span, the abnormal value processing submodule completes the abnormal value elimination and completion, and the standardization submodule standardizes the data to the interval (0, 1).

[0119] An uncertainty scenario generation module is based on an LSTM deep learning model, outputs the flood evolution trend and the uncertainty interval, automatically generates multiple key parameter perturbation scenes, and covers multiple unfavorable flood scenes.

[0120] A data blocking module divides the time series data by using a sliding window method, sets the time step and prediction period parameters, generates algorithm input and output pairs, and supports automatic division and storage of sample sets.

[0121] Swarm intelligence optimization decision module: the core module of the system, including model construction submodule and algorithm solving submodule; the model construction submodule defines decision variables, objective function and hierarchical constraint conditions; the algorithm solving submodule uses the Arctic sea parrot optimization algorithm to traverse the uncertainty scene to solve the optimal flood diversion scheme, and outputs the dynamic flood diversion critical point and the over-limit tolerance space.

[0122] Dynamic iterative optimization module: connected to the real-time monitoring network of the basin, real-time receiving of measured data, judgment of whether the data deviation exceeds the threshold, automatic triggering of scheme recalculation and iterative update.

[0123] Decision result output module: output the dynamic threshold table, flood diversion operation instruction and risk prompt in the form of table and text, support result export, and adapt to the flood control dispatch business scenario.

[0124] Embodiment three: a computer device is provided, including a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executes the program to realize the steps of the above-mentioned intelligent enabling decision method of the flood storage and detention area based on swarm intelligence optimization algorithm. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), ready-to-program gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the computer device, suitable for implementing one or more instructions, and specifically suitable for loading and executing one or more instructions to realize the corresponding method process or corresponding function; the processor described in the embodiment of the present application can be used for the operation of the intelligent enabling decision method of the flood storage and detention area based on swarm intelligence optimization algorithm.

[0125] In an embodiment, a computer readable storage medium is provided, and the computer program is stored in the computer readable storage medium. The computer readable storage medium is a memory device in the computer device, and is used to store programs and data. The computer readable storage medium can include a built-in storage medium in the computer device, and can also include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space, and the storage space stores an operating system of the computer device, and stores one or more instructions suitable for being loaded and executed by the processor, and the instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory.

[0126] The one or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the intelligent enabling decision method of the flood storage and detention basin based on the swarm intelligence optimization algorithm in the above-described embodiments.

[0127] It should be understood by those skilled in the art that the embodiments of the present application can be provided as methods, systems, devices and media. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0128] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The device that implements the functions specified in one or more flows and / or blocks.

[0129] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0131] Those skilled in the art will realize that the embodiments described herein are for illustrative purposes only and that various modifications and changes in light thereof will be apparent to those skilled in the art without departing from the scope and spirit of the application. It is therefore understood that this application can encompass all such modifications and changes as fall within the scope of the appended claims.

Claims

1. A decision-making method for activating flood storage and detention areas based on swarm intelligence optimization algorithms, comprising the following steps: Step 1: Determine the watershed area where the flood storage and detention area is located, and collect multi-source data within the watershed. All data must include multiple time scales, and the data time scales must be consistent. Step 2, Data cropping: Cropping the various types of data collected in Step 1 according to the time dimension to ensure that the start and end times of all hydrological and meteorological data and spatial data are consistent, and to ensure that the data match in terms of time span; Step 3: Remove and complete outliers from the data obtained in Step 2, and perform data standardization. Step 4: Flood evolution trend prediction. Construct an LSTM deep learning model, using the preprocessed historical hydrological data and real-time monitoring data from Step 3 as training samples. Train the model until convergence, and then use the trained model to output future trends. T The hourly trend of upstream and downstream water level / flow rate changes is generated, along with the uncertainty range for forecasting peak flow and peak time. Step 5: Uncertainty scenario generation. Based on the aforementioned uncertainty range, the scenario is automatically generated. N Key parameter disturbance scenarios include, but are not limited to: different situations of exceeding the forecast for flood peak flow, the advance or delay of peak occurrence time, and changes in the risk level of downstream protected objects; Step 6: The time series data preprocessed in Step 3 is divided into blocks using the sliding window method to create the input-output pairs of the algorithm. Step 7: Construct a swarm intelligence optimization algorithm decision model and determine the decision variables and constraints; Step 8: Based on the optimization algorithm decision model determined in Step 7, set the parameters: population size is... M The maximum number of evaluations is FES ; Step 9, Optimization and Solution in Multiple Scenarios: Input the uncertain scenarios generated in Step 5 and the decision sample set divided in Step 6 into the algorithm. The algorithm traverses all scenarios and, under the premise of satisfying the constraints in Step 7, finds the optimal solution corresponding to the minimum value of the objective function. Step 10, Dynamic Iterative Optimization and Output of Flood Diversion Scheme: Real-time data linkage receives monitoring data from the basin. When the deviation between the measured value and the forecast value is large, steps 4 to 9 are automatically triggered to recalculate and iteratively update the flood diversion scheme, and finally output structured decision results.

2. The decision-making method for the activation of flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 1, the multi-source data includes hydrological engineering data, spatial geographic data, and historical case data. The hydrological engineering data covers information such as measured water levels and flows in upstream and downstream areas, flood forecast data, maximum design flow rates at flood diversion gates, and design water levels at dikes. The spatial geographic data includes digital elevation models of flood storage and detention areas, vector data of flood diversion gates and river channels, and distribution information of downstream core protected objects. The historical case data involves hydrological parameters, loss statistics, and decision-making records from historical flood diversion events.

3. The decision-making method for the activation of flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 3, the trimmed and completed numerical data is standardized to the (0, 1) interval to eliminate the influence of dimensions. The standardization formula is as follows: ， In the formula: X std The data is standardized. X This is the original sample data; X min It is the minimum value of this type of data; X max This represents the maximum value of this type of data.

4. The decision-making method for activating flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 6, the specific steps of the sliding window method are as follows: First, set a front pointer and a back pointer. The interval between the pointers is defined as the time step, which is determined by the number of data points included. The interval between the back pointer and the prediction target is defined as the prediction period. Secondly, the front and rear pointers move gradually from the beginning to the end of the data sequence, generating a new data block each time they move a data point, until the rear pointer exceeds the end of the data sequence; Finally, this process will produce a sample set for training and decision-making in swarm intelligence optimization algorithms.

5. The decision-making method for the activation of flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 7, the decision variable is defined: assuming the flood storage and detention area is in operation. X i Dynamic values ​​of flood diversion flow Q Duration of flood diversion T d ; in, X i =1 indicates that the first [function name] is enabled. i A flood storage and detention area, X i =0 indicates that it is not enabled; dynamic value of flood diversion flow. Q The unit is: m³ / s, duration of flood diversion. T d The unit is: h; In step 7, the objective function is constructed: with the minimum total cost of flood diversion as the core objective, the calculation formula is as follows: ， In the formula: C down For downstream disaster losses; C in Losses within the flood storage and detention area; C trans For personnel relocation costs; C eng This refers to the cost of engineering risks.

6. The decision-making method for the activation of flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 7, hierarchical constraints are set: hard constraints must be met, and a penalty coefficient of doubled loss is applied if they are violated: upstream dike water level ≤ design water level, downstream core protected objects have no risk of flooding, and flood diversion gate overflow ≤ maximum design value; soft constraints are optimized to meet: flood storage capacity of flood storage and detention areas ≤ total reservoir capacity, and no duplicate flood diversion when adjacent flood storage and detention areas are activated.

7. The decision-making method for the activation of flood storage and detention areas based on swarm intelligence optimization algorithm according to claim 1, characterized in that: In step 8, regarding algorithm selection, the Arctic Puffin optimization algorithm is adopted as a swarm intelligence optimization algorithm. Through multi-scenario optimization, it outputs the dynamic flood diversion critical point, the over-limit tolerance space, and the optimal flood diversion scheme.

8. A decision-making system for the activation of flood storage and detention areas based on swarm intelligence optimization algorithms, characterized in that: This system can be used to implement the flood storage and detention area activation decision-making method based on swarm intelligence optimization algorithm as described in any one of claims 1 to 7, specifically including: Data collection module: responsible for collecting hydrological engineering data, spatial geographic data, and historical case data within the research basin, supporting the import and storage of multi-format data, and ensuring the consistency of data time scale; Data preprocessing module: includes data pruning submodule, outlier handling submodule, and standardization submodule; the data pruning submodule realizes the uniformity of data time span, the outlier handling submodule completes outlier removal and completion, and the standardization submodule standardizes the data to the (0,1) interval; Uncertainty Scenario Generation Module: Based on the LSTM deep learning model, it outputs the flood evolution trend and uncertainty range, and automatically generates multiple key parameter disturbance scenarios, covering a variety of adverse flood scenarios; Data segmentation module: It uses the sliding window method to segment time series data, sets the time step and prediction period parameters, generates algorithm input-output pairs, and supports automatic partitioning and storage of sample sets; Swarm intelligence optimization decision-making module: The core module of the system, which includes a model building submodule and an algorithm solving submodule; the model building submodule defines decision variables, objective functions and hierarchical constraints; the algorithm solving submodule adopts the Arctic Puffin optimization algorithm to traverse uncertain scenarios to solve the optimal flood diversion scheme and output the dynamic flood diversion critical point and the over-limit tolerance space. Dynamic Iterative Optimization Module: Connects to the real-time monitoring network of the watershed, receives measured data in real time, determines whether the data deviation exceeds the threshold, and automatically triggers the recalculation and iterative update of the scheme; Decision result output module: Outputs dynamic threshold tables, flood diversion operation instructions and risk warnings in tabular and text formats, supports result export, and is adapted to flood control scheduling business scenarios.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the flood storage and detention area activation decision method based on swarm intelligence optimization algorithm as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the decision-making method for the activation of flood storage and detention areas based on a swarm intelligence optimization algorithm as described in any one of claims 1 to 7.

Images