Flood prevention material supply method in combination with flood disaster evolution time sequence

Abstract

The invention discloses a flood prevention material supply method in combination with a flood disaster evolution time sequence, and relates to the technical field of flood prevention material supply, and the method comprises the steps: obtaining a historical flood disaster data set through querying a historical flood disaster database; acquiring a real-time flood disaster data set through a sensor and monitoring equipment; a flood disaster evolution prediction model is generated in combination with the hydrological model and a machine learning algorithm; generating a flood disaster prediction result; determining a material allocation demand by using the flood disaster prediction result; and determining an optimal transportation route based on the material allocation demand in combination with a logistics path optimization algorithm. The technical problem that flood prevention materials cannot be supplied in time due to the fact that flood disasters cannot be accurately predicted in the prior art is solved, and the technical effects of accurately predicting the flood disasters and efficiently supplying the flood prevention materials are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of flood control material supply, and in particular to a method for supplying flood control materials in combination with the time sequence of flood disaster evolution. Background Art

[0002] In today's society, floods are a frequent and destructive natural disaster that poses a serious threat to people's lives, property and social and economic development. With the acceleration of climate change and urbanization, the frequency and intensity of floods are on the rise. In the process of dealing with floods, the timely supply of flood control materials is crucial. However, existing technologies have many shortcomings in flood disaster prediction. Traditional disaster prediction methods often rely on limited historical data and simple models, and it is difficult to fully consider various complex factors, such as the temporal and spatial changes in rainfall, the dynamic changes in river water levels, the diversity of topography and landforms, and the differences in land use types. The interaction of these factors makes the formation and evolution of floods extremely complicated, resulting in the inability of traditional prediction methods to meet the requirements of accurate prediction. Due to the inaccurate prediction of flood disasters, the supply of flood control materials often cannot match the actual demand.

[0003] Existing technologies have the technical problem that flood disasters cannot be accurately predicted, resulting in the inability to supply flood control materials in a timely manner. Summary of the invention

[0004] The present application provides a method for supplying flood control materials in combination with the time series of flood disaster evolution, which is used to solve the technical problem in the prior art that flood disasters cannot be accurately predicted, resulting in the inability to supply flood control materials in a timely manner.

[0005] In view of the above problems, the present application provides a method for supplying flood control materials in combination with the time sequence of flood disaster evolution, the method comprising:

[0006] By querying the historical flood disaster database, the historical flood disaster data set is obtained; by deploying sensors and monitoring equipment in disaster-prone areas, the real-time flood disaster data set is obtained; using the historical flood disaster data set, combined with the hydrological model and machine learning algorithm, a flood disaster evolution prediction model is generated; the real-time flood disaster data set is input into the flood disaster evolution prediction model to generate flood disaster prediction results; using the flood disaster prediction results, combined with the existing material reserves, the demand for allocation of materials is determined; based on the demand for allocation of materials, combined with the logistics path optimization algorithm, a supply plan for flood control materials is formulated to determine the best transportation route from the material reserve center to the disaster-stricken area.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] By querying the historical flood disaster database, historical flood disaster data sets are obtained; by deploying sensors and monitoring equipment in disaster-prone areas, real-time flood disaster data sets are obtained; combining hydrological models and machine learning algorithms, a flood disaster evolution prediction model is generated; the flood disaster evolution prediction model is input to generate flood disaster prediction results; using the flood disaster prediction results, combined with the existing material reserves, the demand for allocating materials is determined; based on the demand for allocating materials, combined with the logistics path optimization algorithm, a supply plan for flood control materials is formulated to determine the best transportation route from the material reserve center to the disaster-stricken area. The technical effect of achieving accurate flood disaster prediction and efficient flood control material supply is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a method for supplying flood control materials in combination with the time sequence of flood disaster evolution provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of a flow chart of generating a flood disaster evolution prediction model in a flood control material supply method combined with a flood disaster evolution time series provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] The present application provides a method for supplying flood control materials in combination with the time series of flood disaster evolution, so as to solve the technical problem in the prior art that flood disasters cannot be accurately predicted, resulting in the inability to supply flood control materials in a timely manner.

[0013] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0014] Examples, such as Figure 1 As shown, the present application provides a method for supplying flood control materials in combination with the time sequence of flood disaster evolution, the method comprising:

[0015] Step S100: Obtain a historical flood disaster dataset by querying a historical flood disaster database.

[0016] Specifically, first, relevant data is collected by querying the historical flood disaster database, which constitutes the historical flood disaster dataset. This dataset contains rich information, including rainfall data in different periods, changes in river water levels, topographic features, land use types, and the actual affected area and flood depth after each flood disaster. These historical data are crucial to understanding the occurrence patterns and characteristics of flood disasters, and will serve as the basis for subsequent model construction and analysis.

[0017] Step S200: Acquire a real-time flood disaster dataset by deploying sensors and monitoring equipment in disaster-prone areas.

[0018] Specifically, various sensors and monitoring equipment deployed in disaster-prone areas are used to collect data related to flood disasters in real time, thereby obtaining real-time flood disaster data sets. These sensors and monitoring equipment can accurately measure the current rainfall, river water levels, real-time changes in topography and landforms, and the dynamic situation of land use types. The acquisition of real-time data can timely reflect the current disaster situation and provide the latest information for accurately predicting the development of flood disasters.

[0019] Step S300: Generate a flood disaster evolution prediction model by using the historical flood disaster dataset, combining a hydrological model and a machine learning algorithm.

[0020] Specifically, historical rainfall, historical river water levels, historical topography and land use types are extracted from the historical flood disaster dataset. The hydrological model processes and analyzes these data based on the principles and equations of hydrology. The hydrological model takes into account factors such as the rainfall-runoff process, the movement of water under different topographic conditions, and the impact of land use types on water infiltration and surface runoff. By establishing a mathematical model to simulate the formation and evolution process of flood disasters, preliminary flood disaster prediction results are generated.

[0021] The random forest algorithm is used. The random forest is an ensemble learning algorithm that consists of multiple decision trees. The preliminary prediction results generated by the hydrological model are compared with the actual affected area and actual flood depth in the historical flood disaster data set, and the error between the two is calculated. For each decision tree, the random forest algorithm randomly extracts samples and features from the historical data to construct a decision tree. In the process of constructing the decision tree, the split node is selected according to the importance of the feature. Then, through the integration of multiple decision trees, the prediction results of each tree are comprehensively considered. In each iteration, the parameter adjustment direction of the decision tree is calculated according to the error. By comparing the difference between the predicted value and the actual value, the mean square error (MSE) or other appropriate error metrics are used to evaluate the performance of the model. According to the back propagation of the error, the node splitting parameters and leaf node values ​​of the decision tree are adjusted. After multiple iterations, the random forest algorithm continuously optimizes the structure and parameters of the decision tree, so that the entire random forest model can better fit the historical data. In this way, the random forest algorithm can capture the complex patterns and nonlinear relationships in the historical flood disaster data, and improve the accuracy and generalization ability of the flood disaster evolution prediction model.

[0022] The final flood disaster evolution prediction model combines the physical basis of the hydrological model and the powerful learning ability of the random forest algorithm. It can accurately predict the development trend of flood disasters based on the new input data, providing a scientific basis for flood control decisions and material allocation.

[0023] Step S400: inputting the real-time flood disaster dataset into the flood disaster evolution prediction model to generate a flood disaster prediction result.

[0024] Specifically, the real-time flood disaster data set obtained by sensors and monitoring equipment deployed in disaster-prone areas is input into the generated flood disaster evolution prediction model to generate flood disaster prediction results, which specifically include the actual disaster-stricken area and the actual flood depth. First, the real-time flood disaster data set contains key information such as real-time rainfall, real-time river water level, real-time topography, and real-time land use type. These real-time data reflect the current state of the disaster. When these real-time data are input into the flood disaster evolution prediction model, the model will analyze and process these data according to its internal algorithms and parameters. The model will comprehensively consider the relationship between various factors, such as how rainfall affects river water levels, and how topography and land use types affect the distribution and accumulation of water flow. Through comprehensive consideration and calculation of these factors, the model finally outputs the prediction results of the actual disaster-stricken area and the actual flood depth. This prediction result helps to understand the scope and extent of damage that may be caused by flood disasters in advance, and provides important decision-making basis for subsequent material allocation and rescue operations. For example, if the forecast results show that the actual disaster-stricken area in a certain area is large and the flood depth is deep, then more flood prevention materials, such as sandbags, water pumps, etc., can be prepared in advance, and more rescue forces can be arranged to go to the area to reduce the losses caused by the disaster.

[0025] Step S500: Determine the demand for allocating materials by using the flood disaster prediction results and combining them with the existing material reserves.

[0026] Specifically, make full use of the actual affected area and actual flood depth in the flood disaster prediction results, and combine the existing material reserves to accurately determine the demand for allocation of materials. First, based on the actual affected area and actual flood depth, make a more accurate assessment of the severity of the affected area. The size of the affected area is directly related to the scope of the area that needs protection and rescue, while the flood depth reflects the impact and destructive power of the flood. For areas with a large affected area and a deep flood depth, more flood control materials are needed to cope with it. Based on this assessment, further analyze the demand for different materials. For example, a large affected area requires a large number of sandbags to build flood control works to prevent the further spread of the flood. Areas with a larger actual flood depth have higher requirements for the power and number of pumps to remove the accumulated water faster. At the same time, a detailed inventory of existing material reserves should be taken. This includes a comprehensive inventory of the types, quantities and qualities of various materials such as sandbags, pumps, emergency food, drinking water and medical supplies. Compare the predicted material needs with the existing material reserves. If the existing material reserves can meet the predicted needs, then the materials will be allocated according to the urgency and importance of the affected area. However, if there is a shortage of supplies, they need to be prioritized according to their importance and urgency. Medical supplies are usually the most critical because they are directly related to the life and health of the affected people, so they should be given priority when supplies are in short supply. Emergency food and drinking water are also indispensable, as they are necessities for maintaining the basic survival of the affected people. Flood control materials such as sandbags and water pumps are reasonably allocated according to the actual conditions of the affected areas. By comprehensively considering the actual affected area and actual flood depth in the flood disaster prediction results and the existing material reserves, the demand for allocation of materials can be scientifically and accurately determined, thereby providing strong support for subsequent material supply and rescue work, and minimizing the loss of life and property caused by flood disasters.

[0027] Step S600: Based on the material allocation demand and in combination with the logistics path optimization algorithm, a supply plan for flood control materials is formulated to determine the best transportation route from the material reserve center to the disaster-stricken area.

[0028] Specifically, according to the demand for allocating materials, it is necessary to clarify the types and quantities of materials that need to be transported, as well as the degree of emergency demand for different materials in each disaster-stricken area. This is the basis for formulating a supply plan, because the characteristics and urgency of different materials will affect the priority and method of transportation. Next, the logistics path optimization algorithm is used. The algorithm will comprehensively consider multiple factors to determine the best transportation route. It will collect the geographical location information of the material storage center and the disaster-stricken area, including the coordinates of the starting and ending points, the structure of the road network, and the road capacity, such as the grade, width, carrying capacity of different roads, and possible traffic congestion. In addition, the type and capacity of transportation tools are also important factors. Different transportation tools have differences in transportation speed, cargo capacity, etc. The algorithm will match them according to the actual situation. When calculating the best transportation route, the algorithm will aim to minimize transportation costs, time and risks. Transportation costs include fuel consumption, vehicle loss, etc.; transportation time is directly related to whether the materials can be delivered to the disaster-stricken area in time. For emergency materials, the time factor is more critical; transportation risks include uncertain factors such as road damage and bad weather. The algorithm will weigh these factors through complex calculations and analysis to find the best transportation route. For example, for high-priority disaster-stricken areas that urgently need medical supplies, priority is given to routes with better road conditions, shorter distances and relatively smooth traffic, even if the transportation cost of this route is slightly higher. For some non-emergency materials, cost factors can be considered to a certain extent and a more economical transportation route can be selected. Through the calculation of the logistics path optimization algorithm, a detailed flood control material supply plan is formulated, and the departure time, transportation tools, transportation routes and estimated arrival time of each batch of materials are clarified. This ensures that flood control materials can be efficiently and accurately transported from the material storage center to the disaster-stricken area, providing strong material support for flood control and disaster relief work, and minimizing the losses caused by flood disasters. By combining the allocation of material needs and the logistics path optimization algorithm, a scientific and reasonable flood control material supply plan is formulated, the best transportation route is determined, the efficiency and effectiveness of material transportation are improved, and flood disasters are better dealt with.

[0029] In a possible manner, step S100 further includes:

[0030] Step S110: preprocessing and standardizing the historical flood disaster dataset, wherein the historical flood disaster dataset includes historical rainfall, historical river water level, historical topography, historical land use type, actual disaster-stricken area, and actual flood depth, wherein the historical rainfall, historical river water level, historical topography, and historical land use type serve as input data, and the actual disaster-stricken area and actual flood depth serve as output data.

[0031] Specifically, historical rainfall, historical river levels, historical topography and historical land use types are determined as input data. For historical rainfall data, the integrity and accuracy of the data are checked, and some outliers or missing values ​​are removed. For example, if the rainfall data for a certain period of time deviates significantly from the normal range and is verified to be a measurement error, it will be removed. At the same time, the rainfall data is sorted and classified, and statistics and analysis are performed according to different time intervals (such as daily rainfall, monthly rainfall, annual rainfall, etc.).

[0032] The preprocessing of historical river water level data is similar, checking the measurement accuracy and consistency of water level data, calibrating and integrating data from different measurement sites. In addition, seasonal changes and long-term trends in river water levels are considered, and useful features are extracted through data analysis methods.

[0033] Historical topographic data exist in different formats and accuracies, and need to be unified, including converting terrain elevation data into a unified coordinate system and unit, and calculating and sorting out parameters such as terrain slope and aspect. At the same time, combined with geographic information system (GIS) technology, topographic data can be visualized and analyzed to better understand its characteristics and changes.

[0034] The processing of historical land use type data requires standardization of different land use classifications. For example, land use types such as farmland, urban areas, and forests should be clearly defined and coded so that these data can be used accurately in subsequent analysis.

[0035] The actual disaster-stricken area and actual flood depth are regarded as output data. For the actual disaster-stricken area data, we will accurately measure and count it, and verify and calibrate it with the help of multiple sources such as satellite remote sensing images and field survey data. The actual flood depth data is obtained from measurements at water level monitoring stations, on-site measurements or model simulation results, which also require quality control and verification.

[0036] In terms of standardization, all data are normalized or standardized to make them comparable and consistent. For example, data such as rainfall and river water levels are converted into standardized numerical ranges so that they can be better processed and compared in subsequent model training and analysis. Through such preprocessing and standardization, the quality and availability of historical flood disaster datasets are improved, providing a reliable data foundation for the subsequent construction of flood disaster evolution prediction models.

[0037] In a possible implementation, step S200 further includes:

[0038] Step S210: preprocessing and standardizing the real-time flood disaster dataset, wherein the real-time flood disaster dataset includes real-time rainfall, real-time river water level, real-time topography, and real-time land use type.

[0039] Specifically, the real-time flood disaster data set is preprocessed and standardized. This data set includes real-time rainfall, real-time river water level, real-time topography and real-time land use type. For real-time rainfall data, the accuracy and integrity of real-time data collection and transmission must be ensured first. The rainfall data obtained by sensors and monitoring equipment will have some noise or interference, so filtering and denoising are required. The median filtering method is used to remove abnormal data points caused by equipment errors or environmental interference. At the same time, the rainfall data is analyzed in time series to better understand its changing trends and periodicity. The preprocessing of real-time river water level data is equally important. The data collected by the water level sensor is calibrated and corrected in real time to eliminate possible measurement bias. In addition, the sudden change of water level data needs to be considered. If the water level changes significantly in a short period of time, the accuracy of the data needs to be further verified. Due to the sudden rise and fall of floods or sensor failures, the abnormal situation of the current water level can be better judged by comparing and analyzing with historical river water level data. In the processing of real-time topographic data, since the changes of topography are relatively slow, but local changes may occur in some special cases (such as geological disasters, etc.), the real-time topographic data are compared and analyzed with the historical topographic data. If there are obvious differences, the data source and accuracy are further verified. At the same time, the real-time topographic data are integrated and updated with the basic topographic data in the geographic information system (GIS) to ensure the most accurate topographic information. For real-time land use type data, the changes in land use are updated in a timely manner, including changes in land use types caused by urban construction, land development or natural disasters. The latest land use information is obtained through satellite remote sensing images, geographic information system data, and field surveys, and integrated into the real-time flood disaster data set. In the process of data integration, data from different sources need to be processed consistently and standardized for subsequent analysis and application. In the standardization processing link, all real-time flood data are standardized. This includes converting data of different units and magnitudes into a unified standard format, such as converting rainfall from millimeters to centimeters, converting river water levels from meters to feet (if necessary), etc. At the same time, the data is normalized so that its value range is within a certain interval, which is convenient for comparison and analysis between different data. Through such preprocessing and standardization, the real-time flood disaster data set is more accurate, reliable and easy to analyze, providing strong support for subsequent flood disaster prediction and emergency decision-making.

[0040] In one possible implementation, Figure 2 As shown, step S300 also includes:

[0041] Step S310: In combination with the hydrological model, the historical rainfall, historical river water level, historical topography, and historical land use type are input to generate preliminary flood disaster prediction results.

[0042] Step S320: Compare the preliminary flood disaster prediction result with the actual disaster-affected area and the actual flood depth in the historical flood disaster dataset to generate a comparison result.

[0043] Step S330: Based on the comparison result, the parameters of the hydrological model are adjusted through a machine learning algorithm to generate the flood disaster evolution prediction model.

[0044] Specifically, historical rainfall, historical river levels, historical topography, and historical land use types are input into the hydrological model, which performs calculations based on the basic principles of hydrology and mathematical formulas. For historical rainfall data, the model incorporates it into the analysis of precipitation input. The intensity, duration, and distribution of rainfall in different regions are carefully considered. Through calculations, the amount of rainwater infiltrating the surface and the amount of surface runoff are determined. These runoffs are affected by topography and land use types. Historical river water level data provides the model with key information on the initial state of the river, and the model uses this data to calculate the flow, flow rate, and the start of water level changes in the river. Combined with topographic features, such as the guiding and obstruction of water flow by different terrains such as valleys and plains, the model can predict the direction and accumulation of water flow. Historical topographic data plays a vital role in predictions, and the model analyzes factors such as the ups and downs, slope, and slope direction of the terrain. In highlands, water may flow faster, while in low-lying areas, water is prone to accumulate. Changes in topography will directly affect the flow path and inundation range of floods, thereby affecting the actual affected area. Historical land use type data will also be fully considered by the model. Different land use types have different permeability and water storage capacity. For example, forests and grasslands have a high infiltration rate and can absorb part of the rainwater, reducing surface runoff; while urban construction land has a low infiltration rate, and rainwater is more likely to form surface runoff. The model will adjust the distribution of water flow on the surface according to these characteristics, thereby affecting the depth of the flood. By comprehensively analyzing these input data, the hydrological model can simulate the formation and development process of the flood, and then predict the actual affected area and actual flood depth. This preliminary flood disaster prediction result provides a basis for subsequent comparison with the actual affected area and actual flood depth in the historical flood disaster dataset, which will help to further optimize the model and improve the accuracy of the prediction.

[0045] The preliminary flood disaster prediction results obtained by the hydrological model are comprehensively compared with the actual disaster-stricken area and actual flood depth in the historical flood disaster data set to generate a critical comparison result. First, the predicted disaster-stricken area and predicted flood depth in the preliminary flood disaster prediction results are extracted. These are simulated and calculated by the model based on the input data such as historical rainfall, historical river water level, historical topography and land use type. At the same time, the actual disaster-stricken area and actual flood depth data that actually occurred in the past are accurately obtained from the historical flood disaster data set. Next, a detailed comparison is carried out. In terms of the disaster-stricken area, the difference between the predicted disaster-stricken area and the actual disaster-stricken area is accurately calculated. If this difference is a positive number, it means that the disaster-stricken area predicted by the model is larger than the actual disaster-stricken area, suggesting that the model has over-predicted in some aspects; conversely, if the difference is a negative number, it means that the disaster-stricken area predicted by the model is smaller than the actual disaster-stricken area, which reflects that the model is insufficient in prediction. For the flood depth, a similar method is also used to accurately calculate the difference between the predicted flood depth and the actual flood depth. Through such a comprehensive and accurate comparison, a clear and definite comparison result is obtained. This comparison result can be presented in the form of specific numerical values, such as the difference between area and depth, or in the form of relative proportions, such as the ratio of the predicted disaster-affected area to the actual disaster-affected area. This comparison result clearly shows the accuracy and deviation of the preliminary flood disaster prediction model, and provides a solid basis for adjusting the parameters of the hydrological model through means such as the random forest algorithm.

[0046] According to the comparison results obtained in step S320, the random forest machine learning algorithm is used to adjust the parameters of the hydrological model, thereby generating a flood disaster evolution prediction model. The random forest algorithm first deeply analyzes the differences between the predicted values ​​and the actual values ​​in the comparison results, and these differences provide the direction and basis for the algorithm to improve. Then, the algorithm adjusts many parameters in the hydrological model. The parameters closely related to rainfall infiltration rate, surface runoff coefficient, water flow resistance, etc. are carefully optimized. During the adjustment process, the random forest algorithm performs parallel calculations by constructing multiple decision trees. Each decision tree is trained based on different samples and features in the historical flood disaster data set. The algorithm randomly selects a part of the data and features to construct each decision tree, which can increase the diversity and stability of the model. Then, by combining the prediction results of multiple decision trees, a more accurate parameter adjustment direction is obtained. For each iteration, the random forest algorithm evaluates the performance of the current parameter combination based on the error calculated by the comparison results. If the error is large, the algorithm will automatically adjust the parameter value, reconstruct the decision tree and make predictions until the optimal parameter combination that minimizes the error is found. By repeating this process, the random forest algorithm can gradually optimize the parameters of the hydrological model. With the continuous optimization of the parameters, the simulation ability of the hydrological model for flood disasters has been significantly improved. It can better combine historical rainfall, historical river water levels, historical topography and land use types and other factors to accurately reflect the occurrence and evolution of flood disasters. Finally, after many iterations and optimizations, a flood disaster evolution prediction model was successfully generated. This model not only has the physical basis of the hydrological model, but also integrates the powerful learning and optimization capabilities of the random forest algorithm. It can provide more accurate and reliable flood disaster prediction results and provide strong support for flood prevention and disaster reduction work.

[0047] In a possible implementation, step S320 further includes:

[0048] Step S321: Calculate the difference between the predicted disaster-affected area and the actual disaster-affected area in the preliminary flood disaster prediction result, and the difference between the predicted flood depth and the actual flood depth, and generate an error value.

[0049] Step S322: partitioning the disaster-stricken area based on its geographical features and population density, and generating partitioning results, wherein the partitioning results include high-priority areas, medium-priority areas, and low-priority areas.

[0050] Step S323: According to the partitioning result, weighted processing is performed on the error value of each area, wherein the error value weight of the high priority area is greater than the error value weight of the medium priority area, and the error value weight of the medium priority area is greater than the error value weight of the low priority area.

[0051] Step S324: Based on the weighted error value, generate weighted error feedback of each partition as the comparison result.

[0052] Specifically, first of all, the two key indicators of the predicted disaster-stricken area and predicted flood depth included in the preliminary flood disaster prediction results are clarified. For the predicted disaster-stricken area, it is compared with the actual disaster-stricken area obtained from the historical flood disaster data set to find the numerical difference between the two. This process requires precise measurement and comparison, taking into account the boundaries and scope of the disaster-stricken area to ensure the accuracy of the calculation. Similarly, for the predicted flood depth, it is also carefully compared with the actual flood depth. The data of the actual flood depth is collected at multiple monitoring points or different areas, and then the difference is calculated one by one with the predicted flood depth. Whether it is the disaster-stricken area or the flood depth, the calculated difference will be an important part of the error value. These error values ​​reflect the degree of deviation of the preliminary prediction results from the actual situation. They are key indicators for evaluating the accuracy of the prediction and will provide an important basis for subsequent model adjustment and optimization, helping to continuously improve the prediction model so that it can more accurately simulate the actual situation of flood disasters and provide more reliable support for flood prevention and disaster reduction work.

[0053] The disaster-affected areas are divided into zones based on their geographical features and population density. Geographical features include the ups and downs of the terrain, the distribution of rivers, etc., while population density reflects the concentration of people in different areas. Based on these factors, the disaster-affected areas are divided into high-priority areas, medium-priority areas, and low-priority areas. High-priority areas are low-lying and densely populated areas, medium-priority areas are relatively flat areas with a certain population distribution, and low-priority areas are sparsely populated areas with relatively high terrain.

[0054] According to the partition results, the error values ​​of each area are weighted. Because different areas have different importance, the error value weight of high priority areas is greater than that of medium priority areas, and the error value weight of medium priority areas is greater than that of low priority areas. This means that the error of high priority areas has a greater impact on the overall evaluation, and more attention is paid to the prediction accuracy of these areas.

[0055] Based on the weighted error values, the weighted error feedback of each partition is generated and used as the comparison result. This comparison result comprehensively considers the different characteristics and importance of the disaster-stricken areas, and can more comprehensively and accurately reflect the performance of the prediction model in different areas. Through such comparison results, the model can be adjusted and optimized more targetedly to improve the prediction accuracy of flood disasters, so as to better support flood prevention and disaster relief work, reasonably allocate resources, and protect people's lives and property.

[0056] In a possible implementation, step S500 further includes:

[0057] Step S510: Based on the flood disaster prediction results, combined with the population density and infrastructure needs of each area, the material demand of each area is calculated. The materials include sandbags, water pumps, emergency food, drinking water and medical supplies.

[0058] Step S520: Based on the current inventory status of the material reserve, the materials are prioritized according to the degree of urgency, and the material allocation demand is generated.

[0059] Specifically, the severity of the disaster in different regions is assessed based on the actual affected area and actual flood depth in the flood disaster prediction results. For areas with larger affected areas and deeper flood depths, the flood threat they face is more severe. For example, areas with deeper flood depths are more likely to have water inflow into houses and flooded roads, which requires more sandbags to build waterproof fortifications to prevent floods from further invading buildings and important facilities. At the same time, the demand for pumps in areas with severe waterlogging will increase accordingly, so as to drain the water as soon as possible and reduce the impact of floods on infrastructure and residents' lives. Considering the population density of each area, areas with high population density will have a greater demand for emergency food, drinking water and medical supplies. Because more people need to maintain basic living needs and deal with possible health problems, taking emergency food as an example, the amount of food required to meet the survival of residents is calculated based on the population size and the expected time of disaster. For drinking water, in addition to meeting daily drinking needs, water demand for sanitation and cleaning needs must also be considered. In terms of medical supplies, densely populated areas will have a corresponding increase in demand for medical supplies such as medicines, bandages, and disinfectants due to frequent contact between people and the increase in injuries and illnesses that may be caused by floods. In addition, infrastructure demand is also one of the key factors in calculating the demand for materials. Areas with weak infrastructure are more vulnerable to flood damage, and the demand for materials such as sandbags and pumps is more urgent to strengthen the protection and repair of infrastructure. For example, some old buildings need more sandbags to reinforce the foundation and doors and windows to prevent floods from soaking and causing the collapse of houses. At the same time, areas with poor infrastructure lack effective drainage systems, which requires more pumps to remove the accumulated water and restore normal production and living order. By comprehensively considering the actual disaster-stricken area and actual flood depth in the flood disaster prediction results, the population density of each region, and infrastructure demand, etc., the specific demand for materials such as sandbags, pumps, emergency food, drinking water and medical supplies in each region is comprehensively and accurately calculated, providing a scientific and reasonable basis for the subsequent material allocation and supply, ensuring that when flood disasters occur, the required materials can be provided to the disaster-stricken areas in a timely and effective manner to protect the safety of people's lives and property and basic living needs.

[0060] Based on the current inventory of material reserves, materials are prioritized according to the degree of urgency, and then the demand for allocation of materials is generated. A comprehensive inventory of the current quantity, quality and expiration date of various materials in the material reserve is taken. First, medical supplies are given the highest priority, because injuries and diseases are common in flood disasters, and medical supplies such as medicines, bandages, and medical equipment are essential for treating the wounded and preventing the spread of diseases. Even if other materials in the inventory are relatively sufficient, as long as there is a shortage of medical supplies, they should be allocated and replenished first. Secondly, emergency food and drinking water are classified as secondary priorities. Flood disasters cause traffic disruptions and supply chain obstructions, and the basic living security of the affected people depends on the supply of emergency food and drinking water. Ensuring sufficient food and clean drinking water is the key to maintaining the life, health and physical strength of the affected people. Then, for flood control materials such as sandbags and pumps, they are evaluated based on the actual situation in the affected area and the predicted flood development trend. If some areas are facing a direct threat of floods and there are limited sandbags and pumps in stock, their priority will be increased accordingly. Through this priority classification based on the degree of urgency, we can clearly understand the allocation needs of different materials. When materials are limited, we can reasonably arrange the allocation order and give priority to meeting the most urgent and critical needs, so as to minimize the impact of floods and waterlogging disasters on people’s lives and property and ensure the smooth progress of flood prevention and disaster relief work.

[0061] In a possible implementation, step S600 further includes:

[0062] Step S610: Based on real-time traffic data, use a dynamic path planning algorithm to preliminarily plan multiple material transportation routes.

[0063] Step S620: Based on the urgency of the materials and the priority of the disaster-stricken areas, the best solution is selected from multiple transportation routes through a multi-objective optimization algorithm.

[0064] Specifically, we make full use of real-time traffic data and use dynamic path planning algorithms to preliminarily plan multiple routes for material transportation. The dynamic path planning algorithm is an intelligent path planning method that can find a suitable path for material transportation based on real-time changing traffic conditions. First, the algorithm collects a wealth of real-time traffic data from a wide range of sources, including road sensors, traffic monitoring cameras, satellite navigation systems, and real-time information provided by traffic management departments. Next, the algorithm sets the material reserve center as the starting point and the disaster-stricken areas as the end points. Then, based on the collected real-time traffic data, the algorithm begins path planning, which takes into account multiple factors to calculate possible transportation routes. The algorithm uses graph theory to abstract the road network into a graph structure, where nodes represent intersections or important locations of roads, edges represent road segments connecting nodes, and edge weights are set according to traffic conditions. For example, sections with heavy traffic and slow driving speeds have larger weights, and vice versa. During the planning process, the algorithm will use a search strategy to find the path from the starting point to the end point. Taking the A algorithm as an example, it selects the next node to be expanded by comprehensively considering the actual cost from the starting point to the current node (such as driving time or distance) and the estimated cost from the current node to the end point (usually based on some heuristic function). In this way, the search process can move towards the target direction more specifically and improve the search efficiency. In this way, the dynamic path planning algorithm can flexibly adjust the path planning according to the changes in real-time traffic data, and provide a variety of feasible options for material transportation, so as to further select the optimal transportation route in the subsequent steps.

[0065] Based on the urgency of the materials and the priority of the disaster-stricken areas, a multi-objective optimization algorithm is used to determine the optimal solution from multiple transportation routes. First, the urgency of the materials is clarified. For example, medical materials are usually regarded as the most urgent materials because they are directly related to the life and health of the affected people, and delayed delivery may lead to serious consequences; emergency food and drinking water are also highly urgent, and they are necessities for maintaining the basic survival of the affected people. At the same time, the priority of the disaster-stricken areas is determined, and they are divided according to factors such as population density, degree of disaster, and infrastructure damage in the disaster-stricken areas. Areas with dense populations and severe disasters and severe infrastructure damage may be given higher priority. The multi-objective optimization algorithm considers multiple objectives at the same time to evaluate each transportation route. On the one hand, transportation time is a key goal. For emergency materials and high-priority disaster-stricken areas, it is crucial to minimize transportation time. The algorithm analyzes factors such as traffic conditions and road distance on each path to estimate the time required for transportation. On the other hand, transportation costs are also factors that need to be considered, including fuel consumption, vehicle wear and tear, and labor costs. In addition, the loss of materials will also be taken into consideration, such as the risk of deterioration of some perishable foods during transportation, or the possibility of damage to precision medical equipment during transportation. The multi-objective optimization algorithm will assign different weights to these goals according to the urgency of the materials and the priority of the disaster-stricken areas. For emergency materials and high-priority disaster-stricken areas, the weight of transportation time will be relatively high; while for some non-emergency materials, the weight of transportation cost will be appropriately increased. Then, the algorithm conducts a comprehensive evaluation of each transportation path and calculates a comprehensive fitness value. Through a comprehensive evaluation and comparison of multiple transportation paths, the multi-objective optimization algorithm finally selects the optimal solution. This optimal solution can balance multiple goals such as transportation time, transportation cost and material loss as much as possible under the premise of meeting the urgency of materials and the priority of the disaster-stricken areas, ensuring that materials can be delivered to the disaster-stricken areas efficiently, economically and safely, and providing strong support and guarantee for flood control and disaster relief work.

[0066] In a possible implementation, step S620 further includes:

[0067] Step S621: planning a backup path for the optimal solution. If the transportation path of the optimal solution is blocked by floods, the backup path is automatically activated.

[0068] Specifically, the optimal solution determined by the multi-objective optimization algorithm plans the backup path. Considering the uncertainty and destructiveness of floods, even if the optimal material transportation path is determined, the path may still be blocked by floods. Therefore, it is necessary to plan the backup path in advance to deal with such emergencies. When planning the backup path, a variety of factors are considered, similar to the optimal solution, including traffic conditions, road conditions, distance, etc. At the same time, the prediction of the possible impact range and degree of floods is combined to try to select those paths that are less affected by floods as backup paths. Once the transportation path of the optimal solution is blocked by floods, the backup path will be automatically activated. Through multiple channels such as real-time traffic monitoring systems, meteorological data, and feedback from on-site personnel, the status information of the transportation path is obtained in a timely manner. When it is monitored that the optimal path is impassable due to floods, it will automatically switch to the backup path to ensure that the material transportation will not be stagnant due to path blocking. For example, suppose the optimal transportation path passes through a low-lying section that is easily flooded, and through real-time monitoring, it is found that the section has been flooded and is impassable. At this time, the backup path will immediately come into play. The alternative route will bypass the affected section of road and choose other roads with higher terrain and relatively better road conditions to continue transporting supplies. This can minimize the interference of floods and waterlogging disasters on the transportation of supplies, ensure that supplies can be delivered to the affected areas continuously and stably, and provide solid material support for flood control and disaster relief work.

[0069] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0071] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for supplying flood control materials in combination with the time series of flood disaster evolution, characterized in that: The method comprises: Obtain historical flood disaster datasets by querying the historical flood disaster database; Obtain real-time flood disaster datasets through sensors and monitoring equipment deployed in disaster-prone areas; Using the historical flood disaster dataset, combined with a hydrological model and a machine learning algorithm, a flood disaster evolution prediction model is generated; Inputting the real-time flood disaster data set into the flood disaster evolution prediction model to generate flood disaster prediction results; Determine the need for allocating materials by using the flood disaster prediction results and combining them with the existing material reserves; Based on the material allocation needs and combined with the logistics path optimization algorithm, a supply plan for flood control materials is formulated to determine the best transportation route from the material reserve center to the disaster-stricken area.

2. The method according to claim 1, characterized in that By querying the historical flood disaster database, a historical flood disaster data set is obtained, and the method further includes: The historical flood disaster dataset is preprocessed and standardized, wherein the historical flood disaster dataset includes historical rainfall, historical river water level, historical topography, historical land use type, actual disaster-stricken area, and actual flood depth, wherein the historical rainfall, historical river water level, historical topography, and historical land use type serve as input data, and the actual disaster-stricken area and actual flood depth serve as output data.

3. The method according to claim 2, characterized in that By deploying sensors and monitoring equipment in disaster-prone areas, real-time flood disaster data sets are obtained, and the method further includes: The real-time flood disaster data set is preprocessed and standardized, wherein the real-time flood disaster data set includes real-time rainfall, real-time river water level, real-time topography, and real-time land use type.

4. The method according to claim 3, characterized in that Using the historical flood disaster dataset, combined with a hydrological model and a machine learning algorithm, a flood disaster evolution prediction model is generated, and the method further includes: Combined with the hydrological model, the historical rainfall, historical river water level, historical topography and land use type are input to generate preliminary flood disaster prediction results; Comparing the preliminary flood disaster prediction result with the actual disaster-affected area and the actual flood depth in the historical flood disaster dataset to generate a comparison result; Based on the comparison results, the parameters of the hydrological model are adjusted through a machine learning algorithm to generate the flood disaster evolution prediction model.

5. The method according to claim 4, characterized in that Comparing the preliminary flood disaster prediction result with the actual disaster-affected area and the actual flood depth in the historical flood disaster dataset to generate a comparison result, the method further comprising: Calculating the difference between the predicted disaster-affected area and the actual disaster-affected area in the preliminary flood disaster prediction result, and the difference between the predicted flood depth and the actual flood depth, to generate an error value; Based on the geographical features and population density of the disaster-stricken area, the disaster-stricken area is partitioned to generate partition results, wherein the partition results include high-priority areas, medium-priority areas, and low-priority areas; According to the partitioning result, weighted processing is performed on the error values ​​of each area, wherein the error value weight of the high priority area is greater than the error value weight of the medium priority area, and the error value weight of the medium priority area is greater than the error value weight of the low priority area; Based on the weighted error values, weighted error feedback of each partition is generated as the comparison result.

6. The method according to claim 5, characterized in that Using the flood disaster prediction results and combining the existing material reserves, the material demand for allocation is determined, and the method further includes: Based on the flood disaster prediction results, combined with the population density and infrastructure needs of each region, the material demand of each region is calculated, including sandbags, water pumps, emergency food, drinking water and medical supplies; According to the current inventory situation of material reserves, the materials are prioritized according to their urgency and the material allocation demand is generated.

7. The method according to claim 6, characterized in that Based on the material allocation demand, combined with the logistics path optimization algorithm, a flood control material supply plan is formulated to determine the best transportation route from the material reserve center to the disaster-stricken area. The method also includes: Based on real-time traffic data, a dynamic path planning algorithm is used to preliminarily plan multiple material transportation routes; Based on the urgency of the materials and the priority of the affected areas, the optimal solution is selected from multiple transportation routes through a multi-objective optimization algorithm.

8. The method according to claim 7, characterized in that Based on the urgency of the materials and the priority of the disaster-stricken areas, the optimal solution is selected from multiple transportation routes through a multi-objective optimization algorithm. The method further includes: A backup path is planned for the optimal solution. If the transportation path of the optimal solution is blocked by flood disasters, the backup path is automatically activated.

Images