Method for Assessing Water Supply Risk under Saltwater Intrusion in Tidal River Estuaries Based on Freshwater Storage and Diversion Scheduling

Abstract

The present invention discloses a method, system, device and storage medium for evaluating the water supply risk under the saltwater intrusion in a tidal estuary area based on the fresh water storage and diversion scheduling. The method includes: obtaining the daily target site data and historical saltwater intrusion disaster simulation data in the same time series in the research area; training the historical saltwater intrusion disaster simulation model with the target site data, and inputting the historical saltwater intrusion disaster simulation data into the trained historical saltwater intrusion disaster simulation model to obtain the long-sequence historical salinity simulation values; generalizing the water supply system in the research area and constructing a comprehensive water supply risk evaluation index system based on the fresh water storage and diversion scheduling; calculating the comprehensive water supply risk index based on the long-sequence historical salinity simulation values according to the comprehensive water supply risk evaluation index system; performing frequency analysis on the comprehensive water supply risk index, and dividing the water supply risk level and early warning level based on the set water supply risk threshold. The present invention quantifies the comprehensive water supply risk index based on the fresh water storage and diversion scheduling, and scientifically divides the saltwater intrusion risk level and early warning level from the perspectives of the influence degree and occurrence frequency of the long-sequence historical saltwater intrusion disasters, providing an effective decision-making basis for the water supply safety judgment in the estuary and coastal areas relying on the tidal river section as the water source.

Description

Technical Field

[0001] The present invention relates to the field of saltwater intrusion risk assessment, and particularly to a method, a system, a terminal device and a computer-readable storage medium for assessing the water supply risk under saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water. Background Art

[0002] The estuary and coastal areas are key regions for human survival and economic and social development. Saltwater intrusion is a common natural phenomenon in this region. Especially when the upstream runoff weakens, the tidal action will intensify the mixing of salt and fresh water, resulting in the upstream movement of the salt-fresh water interface along the river channel, and then polluting the fresh water resources in the tidal reach. Affected by the dual impacts of global climate change and human activities, sea level rise and extreme climate events occur frequently, which is expected to make the situation of global saltwater intrusion in estuaries increasingly severe, posing a greater challenge to the water supply security and social stability in the estuary and coastal areas.

[0003] From the research status at home and abroad, the current research mainly focuses on the mechanism and simulation of saltwater intrusion, while there is less discussion on the risk assessment and early warning methods for the threat of saltwater intrusion to water supply security. On the one hand, limited by the limitations of historical salinity observation data, the time periods involved in existing saltwater intrusion mechanism research and simulation research are generally short, which is not sufficient to scientifically evaluate the saltwater intrusion risk level and threshold from the perspectives of the impact degree and occurrence frequency of long-sequence historical saltwater intrusion disasters. On the other hand, when the main water intake cannot normally draw water due to excessive salinity, emergency scheduling measures are generally taken to increase the downstream river flow by using large upstream reservoirs, so as to alleviate the saltwater intrusion in the estuary, improve the probability of drawing fresh water at the water intake, and make full use of the regulation capacity of local reservoirs to seize and store fresh water resources to the greatest extent. Although existing research has judged the impact degree and early warning level of saltwater intrusion from the perspectives of the upstream distance of saltwater intrusion, duration, maximum salinity, and the number of hours of exceeding the standard at the water intake, the regulation role of local reservoirs has not been considered. The water stored in local reservoirs serves as supplementary fresh water resources in the initial stage of saltwater intrusion. When the upstream water inflow is severely depleted and the saltwater intrusion lasts for a long time, the limited reservoir capacity may not be able to ensure sufficient and continuous fresh water supply, and there is a risk of water cut-off. Summary of the Invention

[0004] In order to solve at least one of the above technical problems in the prior art, the present invention provides a method, a system, a terminal device and a computer-readable storage medium for assessing the water supply risk under saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water. By comprehensively evaluating the saltwater intrusion risk from the perspectives of the natural characteristics of saltwater intrusion, reservoir regulation, and the impact degree and occurrence frequency of historical disasters, different risk levels are clarified, providing a rapid and reliable decision-making basis for saltwater intrusion risk control, which is of great significance for improving the water safety guarantee ability in the estuary and coastal areas.

[0005] The first object of the present invention is to provide a method for assessing the water supply risk under the saltwater intrusion in the tidal estuary area based on the fresh water storage and diversion scheduling.

[0006] The second object of the present invention is to provide a system for assessing the water supply risk under the saltwater intrusion in the tidal estuary area based on the fresh water storage and diversion scheduling.

[0007] The third object of the present invention is to provide a terminal device.

[0008] The fourth object of the present invention is to provide a computer-readable storage medium.

[0009] The first object of the present invention can be achieved by adopting the following technical solutions:

[0010] A method for assessing the water supply risk under the saltwater intrusion in the tidal estuary area based on the fresh water storage and diversion scheduling, the method comprising:

[0011] Obtaining the target site data and the historical saltwater disaster simulation data on a daily basis within the same time series in the research area; wherein, the target site data includes the measured salinity data, flow data, tide level data and wind force data, and the historical saltwater disaster simulation data includes the flow data, tide level data and wind force data on a daily basis within a long time series;

[0012] Training the historical saltwater disaster simulation model by using the target site data, and inputting the historical saltwater disaster simulation data into the trained historical saltwater disaster simulation model to obtain the long-sequence historical salinity simulation values;

[0013] Generalizing the water supply system in the research area, and constructing a comprehensive evaluation index system for the water supply risk based on the fresh water storage and diversion scheduling; based on the comprehensive evaluation index system for the water supply risk, calculating the comprehensive water supply risk index according to the long-sequence historical salinity simulation values; wherein, the water supply system includes the main river channel water intake, the local water storage reservoir and the water supply object;

[0014] Performing frequency analysis on the comprehensive water supply risk index, and dividing the water supply risk level and the early warning level based on the set water supply risk threshold.

[0015] Further, calculating the comprehensive water supply risk index according to the long-sequence historical salinity simulation values based on the comprehensive evaluation index system for the water supply risk, includes:

[0016] Based on the comprehensive evaluation index system for the water supply risk, calculating the values of each quantitative evaluation index in the comprehensive water risk evaluation index system according to the long-sequence historical salinity simulation values, the maximum water intake capacity of the main river channel water intake, the adjustable storage capacity of the local water storage reservoir and the water demand of the water supply object;

[0017] The symbol conversion method is used to process the values of each quantitative evaluation index to obtain the positive index values after normalization of each quantitative evaluation index;

[0018] According to the values of each processed quantitative evaluation index, the analytic hierarchy process and the pairwise judgment method are used to construct a judgment matrix and calculate the corresponding weight matrix; if the consistency test index of the weight matrix is greater than or equal to the first set threshold, the judgment matrix is reconstructed;

[0019] According to the corresponding weight matrix of the determined judgment matrix and the values of each processed quantitative evaluation index, the comprehensive water supply risk index is calculated.

[0020] Further, the water supply system in the research area is generalized to construct a comprehensive water supply risk evaluation index system based on the fresh water storage and regulation, including:

[0021] Quantitative evaluation indexes are selected from the probability of taking fresh water from the main river water intake, the dependence on local water storage reservoirs, and the water shortage degree of water supply objects, specifically including:

[0022] From the aspect of the probability of taking fresh water from the main river water intake, the average probability of taking fresh water R1, the lowest probability of taking fresh water R2, and the number of days without taking fresh water R3 of the main river water intake in the next k days are selected; k is a positive integer greater than or equal to 1;

[0023] From the aspect of the dependence on local water storage reservoirs, the difference R4 between the average water demand and the average water intake in the next k days and the average full storage rate R5 in the next k days are selected;

[0024] From the aspect of water shortage degree, the average substantial water shortage amount R6 and the substantial water shortage days R7 in the next k days are selected;

[0025] The calculation formulas of each quantitative evaluation index are as follows:

[0026]

[0027] Where:

[0028]

[0029] In the formula, T hour is the number of hours of salinity exceeding the standard at the main river water intake, and f hour (S) is a function fitted according to the empirical relationship between the number of hours of daily exceeding the standard and the daily average salinity in the measured salinity data, and S is the long-sequence historical salinity simulation value; ave(T hour ) is the average daily number of hours of exceeding the standard at the main river water intake in the next k days; min(T hour ) represents the longest daily number of hours of exceeding the standard at the main river water intake in the next k days; sum(n) represents the number of days without taking fresh water at the main river water intake in the next k days. Considering that the actual water pump cannot be started and stopped quickly, the probability of taking fresh water 1 - Thour When the value of / 24 is less than 0.1, it is counted as unable to obtain fresh water within one day, that is, n = 1; D a is the daily water demand of the water supply object, ave(D a ) is the average daily water demand in the next k days; W a is the daily water intake at the main water intake of the river channel, ave(W a ) is the average daily water intake in the next k days; The full storage rate P t is the effective water storage V on the current day t and the dead storage capacity V s The ratio of the sum of and the normal storage capacity V of the reservoir total , ave(P t ) represents the average full storage rate of the local water storage reservoir in the next k days; W d is the substantial water shortage, ave(W D ) represents the average water shortage in the next k days; sum(m) represents the number of days of substantial water shortage in the next k days. If W d > 0 on the current day, it is counted that substantial water shortage occurs on the current day, that is, m = 1; W c is the maximum water intake capacity of the main water intake of the river channel; V a represents the amount of water used to supplement the local water storage reservoir with the excess water intake; W s is the amount of water that needs to be supplemented by the local water storage reservoir; V t and V t-1 represent the effective water storage of the local water storage reservoir on the current day and the previous day respectively; k is a positive integer greater than or equal to 1;

[0030] The rules for the dispatching of seizing and storing fresh water under the intrusion of salt tide are as follows: When the salt tide intrudes, each main water intake of the river channel takes water from the river channel to the maximum extent during the available water intake period. The water intake first meets the needs of the water supply object. If there is still a surplus, the excess water intake is used to supplement the local water storage reservoir. The effective water storage of the reservoir after replenishment should not exceed the regulating storage capacity of the reservoir; If the water intake cannot meet the needs of the water supply object, the local reservoir is used to release water for supplementary water supply. The effective water storage of the reservoir after supplementary water supply should not be less than 0; If the water intake and the supplementary water supply of the reservoir still cannot meet the needs of the water supply object, substantial water shortage occurs.

[0031] Furthermore, the frequency analysis of the comprehensive water supply risk index is carried out, and the water supply risk level and early warning level are divided based on the set water supply risk threshold, including:

[0032] First, remove the comprehensive water supply risk index less than the second set threshold from the comprehensive water supply risk index;

[0033] Then, conduct frequency analysis on the remaining long sequence of daily comprehensive water supply risk indexes, and count the cumulative frequencies corresponding to different comprehensive water supply risk indexes;

[0034] Finally, based on the set water supply risk threshold, the water supply risk level and warning level are divided according to the cumulative frequency.

[0035] Furthermore, the historical saltwater intrusion disaster simulation model is a multi-layer BP neural network model;

[0036] The training of the historical saltwater intrusion disaster simulation model using the target site data includes:

[0037] Input the measured flow data, tide level data and wind force data as the saltwater intrusion impact factors into the historical saltwater intrusion disaster simulation model, and output the salinity simulation values of the main water intake points of each river channel;

[0038] According to the measured salinity data of the main water intake points of each river channel and the long-term historical salinity simulation values, evaluate the performance of the historical saltwater intrusion disaster simulation model to obtain the optimal lag time of each saltwater intrusion impact factor. The corresponding BP neural network model is the trained historical saltwater intrusion disaster simulation model.

[0039] Furthermore, the optimal lag time of each saltwater intrusion impact factor is obtained through the following process:

[0040] Use the particle swarm optimization algorithm, and take the lag time of each saltwater intrusion impact factor together with the connection weights and thresholds of the multi-layer BP neural network as the position values of the particles;

[0041] Take the performance index value of the BP neural network model as the fitness value of the particles, and use the target site data for iterative calculation to determine the position value that maximizes the particle fitness value, that is, obtain the optimal lag time of each impact factor and the optimal neural network parameters. The corresponding BP neural network model is the BP neural network model with the optimal performance.

[0042] Furthermore, the long time series is not less than 20 years.

[0043] Furthermore, after the frequency analysis of the comprehensive water supply risk index, the method further includes determining the warning indicative indicators from the aspects of the flow of the upstream controlling hydrological station, the fresh water intake probability of the main water intake points of the river channel, and the average full storage rate of the local water storage reservoir; according to the warning indicative indicators, divide the risk level warning, specifically including:

[0044] Based on the set water supply risk threshold, divide the comprehensive water supply risk index RI into corresponding categories;

[0045] Use the Mann-Whitney U non-parametric test method to select three indicators: the average flow data in the next k days, the average fresh water intake probability of the main water intake points of the main river channels in the next k days, and the average full storage rate of the local water storage reservoir in the next k days. Analyze the differences between the three indicators in different categories, and take the indicators with significant differences as the warning indicative indicators;

[0046] If the warning indicative index M is greater than the determination threshold, it is classified into the corresponding warning level;

[0047] Among them, the determination threshold is:

[0048]

[0049] In the formula, M Aave and M Bave are respectively taken from the two categories of RIs that the index M is intended to distinguish; there are significant differences in the index M between the two categories of RIs, which are respectively marked as range A and range B, where range B corresponds to a higher warning level; M Aave is the average in range A; M Bave is the average in range B; M A-Bth is the determination threshold of the index M between range A and range B, and is used as the criterion for determining that the index M in range A enters the warning level corresponding to range B.

[0050] Further, after obtaining the daily target site data of the study area in the same time series, the method further includes:

[0051] Training the saltwater intrusion prediction model using the target site data, inputting the monitored target site data into the trained saltwater intrusion prediction model, and obtaining the predicted salinity values for the next f days; both the saltwater intrusion prediction model and the historical saltwater disaster simulation model adopt a multi-layer BP neural network model; where f is a positive integer greater than or equal to 1;

[0052] Based on the comprehensive evaluation index system of water supply risk, calculate the comprehensive water supply risk index according to the predicted salinity values for the future;

[0053] According to the comprehensive water supply risk index and the set threshold, divide the water supply risk level and grade warning.

[0054] Further, the training of the saltwater intrusion prediction model using the target site data includes:

[0055] Taking the daily average salinity in the salinity data as the cumulative effect of the historical salinity, and together with the flow data and wind data as the saltwater intrusion influencing factors, inputting them into the saltwater intrusion prediction model to obtain the predicted salinity values for the future;

[0056] Evaluating the performance of the saltwater intrusion prediction model according to the salinity data and the predicted salinity values for the future, obtaining the optimal lag time of each saltwater intrusion influencing factor, and the corresponding BP neural network model is the trained saltwater intrusion prediction model.

[0057] The second object of the present invention can be achieved by adopting the following technical solutions:

[0058] A water supply risk assessment system under saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water. The system includes the following acquisition steps:

[0059] An acquisition module, used to acquire the daily target site data and historical saltwater intrusion disaster simulation data in the same time series in the research area; among them, the target site data includes measured salinity data, flow data, tide level data, and wind force data, and the historical saltwater intrusion disaster simulation data includes the daily flow data, tide level data, and wind force data in a long time series;

[0060] A training module, used to train the historical saltwater intrusion disaster simulation model with the target site data, and input the historical saltwater intrusion disaster simulation data into the trained historical saltwater intrusion disaster simulation model to obtain long-sequence historical salinity simulation values;

[0061] A construction and calculation module, used to generalize the water supply system in the research area, and construct a comprehensive water supply risk evaluation index system based on the regulation of seizing and storing fresh water; based on the comprehensive water supply risk evaluation index system, calculate the comprehensive water supply risk index according to the long-sequence historical salinity simulation values; among them, the water supply system includes the main river channel water intake, local water storage reservoirs, and water supply objects;

[0062] A division module, used to perform frequency analysis on the comprehensive water supply risk index, and divide the water supply risk level and early warning level based on the set water supply risk threshold.

[0063] The third object of the present invention can be achieved by adopting the following technical solutions:

[0064] A terminal device includes a processor and a memory for storing the executable program of the processor. When the processor executes the program stored in the memory, it realizes the above-mentioned water supply risk assessment method under saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water.

[0065] The fourth object of the present invention can be achieved by adopting the following technical solutions:

[0066] A computer-readable storage medium stores a program. When the program is executed by a processor, it realizes the above-mentioned water supply risk assessment method under saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water.

[0067] The present invention has the following beneficial effects compared with the prior art:

[0068] The present invention is based on the dispatching rules for seizing and storing fresh water under the condition of saltwater intrusion, combined with the saltwater intrusion simulation, the dispatching of seizing and storing fresh water in local reservoirs, and the comprehensive evaluation index system for regional water supply risks. The comprehensive water supply risk index is used to quantify the regional water supply risk under saltwater intrusion. By analyzing the frequency distribution of the comprehensive water supply risk index under the long-sequence historical saltwater intrusion simulation, different degrees of water supply risk and early warning levels are clarified. The flow rate of the upstream controlled hydrological station, the probability of taking fresh water at the main water intake, the average full storage rate of local reservoirs (groups), etc. are used as early warning indicative indicators to divide the risk level early warning. The present invention provides an effective analysis idea and evaluation technology for the water supply safety judgment in the estuary and coastal areas that rely on the tidal reach as the water source, and plays a promoting role in improving the water safety guarantee ability in the estuary and coastal areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the structures shown in these drawings.

[0070] Figure 1 It is a flow chart of the water supply risk assessment method under saltwater intrusion in the tidal estuary area based on the dispatching of seizing and storing fresh water in Embodiment 1 of the present invention;

[0071] Figure 2 It is a schematic diagram of the water supply risk assessment method under saltwater intrusion in the tidal estuary area based on the dispatching of seizing and storing fresh water in Embodiment 1 of the present invention;

[0072] Figure 3 It is the daily saltwater intrusion risk early warning result of the Zhuhai - Macau, China water supply system during the dry season from 1979 to 2019 in Embodiment 1 of the present invention;

[0073] Figure 4 It is the daily saltwater intrusion risk early warning result of the Zhuhai - Macau, China water supply system during the dry season in 2022 based on the early warning indicative index judgment method in Embodiment 1 of the present invention;

[0074] Figure 5 It is a structural block diagram of the water supply risk assessment system under saltwater intrusion in the tidal estuary area based on the dispatching of seizing and storing fresh water in Embodiment 2 of the present invention;

[0075] Figure 6 It is a structural block diagram of the terminal device in Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0076] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention. It should be understood that the specific embodiments described are only for explaining the present application and not for limiting the present application.

[0077] Embodiment 1:

[0078] Reference Figure 1 、 2 This embodiment provides a method for assessing the water supply risk under the saltwater intrusion in a tidal estuary area based on the regulation of seizing and storing fresh water, including the following steps:

[0079] S101. Obtain the daily target site data and historical saltwater disaster simulation data in the same time series for the research area; among them, the target site data includes measured salinity data, flow data, tide level data, and wind force data, and the historical saltwater disaster simulation data includes daily flow data, tide level data, and wind force data in a long time series;

[0080] S102. Use the target site data to train the historical saltwater disaster simulation model and the saltwater intrusion prediction model respectively; input the historical saltwater disaster simulation data into the trained historical saltwater disaster simulation model to obtain the long-sequence historical salinity simulation values; input the monitored target site data into the trained saltwater intrusion prediction model to obtain the future salinity prediction values;

[0081] S103. Generalize the water supply system in the research area and construct a comprehensive evaluation index system for water supply risk based on the regulation of seizing and storing fresh water; based on the comprehensive evaluation index system for water supply risk, calculate the comprehensive water supply risk index according to the long-sequence historical salinity simulation values or future salinity prediction values; among them, the water supply system includes at least the main river channel water intake, local water storage reservoirs, and water supply objects;

[0082] S104. Conduct a frequency analysis on the comprehensive water supply risk index calculated according to the long-sequence historical salinity simulation values, divide the water supply risk level and early warning level; and determine the early warning indicative indicators from the aspects of the flow of the upstream control hydrological station, the probability of taking fresh water at the main river channel water intake, and the average full storage rate of the local reservoir group; divide the corresponding level of early warning according to the comprehensive water supply risk index calculated from the future salinity prediction values or its early warning indicative indicators.

[0083] This embodiment takes the Modaomen Waterway in the Pearl River Estuary and the Zhuhai - China Macau Water Supply System as the research object. First, determine the target stations in the research area. The target stations include the main water intake points of the river channel, the upstream controlled hydrological stations, the estuary tide and meteorological stations, and collect data such as salinity at the main water intake points of the river channel, flow at the hydrological stations, tide level at the tide stations, and wind force at the meteorological stations. Then, based on machine learning methods, construct a historical saltwater intrusion disaster simulation model and a saltwater intrusion prediction model for the research area to obtain the long - sequence historical salinity simulation values and future salinity prediction values of the main water intake points of the Modaomen Waterway respectively. Combining the characteristics of the Zhuhai - China Macau Water Supply System, construct a water resources supply - demand balance model and a comprehensive water supply risk evaluation index system for the research area. Based on the comprehensive water supply risk evaluation index system, calculate the comprehensive water supply risk index of the area according to the long - sequence historical salinity simulation values or future salinity prediction values. Subsequently, conduct a frequency analysis on the comprehensive water supply risk index calculated based on the long - sequence historical salinity simulation values, determine the water supply risk threshold and warning level of the Zhuhai - China Macau Water Supply System under saltwater intrusion, and determine the warning indicative indicators from aspects such as the flow at the upstream controlled hydrological stations, the fresh - water intake probability at the main water intake points of the main river channel, and the average full - storage rate of the local reservoir group. Finally, issue warnings of corresponding levels according to the comprehensive water supply risk index or warning indicative indicator values calculated from the future salinity prediction values.

[0084] Further, step S101 specifically includes:

[0085] (1) Collect data of the target stations as the model training and test data sets.

[0086] Obtain the daily measured salinity data, flow data, tide level data, and wind force data of the target stations within the same time series, which are mainly used for the training and testing of the machine learning model. Select Zhuzhoutou, Pinggang, and Guangchang Pumping Stations in the Modaomen Waterway as the 3 main water intake points, and collect the hourly chloride content data of each water intake point from 2019 to 2023. After inspection and processing, obtain the daily number of hours exceeding the standard and the daily average salinity data. Collect the daily flow data of Wuzhou Hydrological Station, the upstream control station of the Xijiang River, and Shijiao Hydrological Station, the upstream control station of the Beijiang River, in the Modaomen Waterway from 2019 to 2023. Collect the daily average sea level and daily maximum tidal range data of Causeway Bay Station, Hong Kong, outside the estuary of the Modaomen Waterway from 2019 to 2023. Collect the daily average wind speed and prevailing wind direction data of Waglan Island Station, Hong Kong, outside the estuary of the Modaomen Waterway from 2019 to 2023. Conduct quality control on the above data to obtain the model training and test data sets.

[0087] (2) Collect historical saltwater intrusion disaster simulation data.

[0088] Obtain the daily measured flow data, tide level data, and wind force data of the target stations within a relatively long time series, which are mainly used for the long - sequence historical saltwater intrusion disaster simulation.

[0089] In this embodiment, the long-term measured flow data of hydrological stations are obtained to obtain the natural runoff process through runoff restoration calculation, so as to provide input for simulating historical saltwater intrusion disasters without human activity interference. Through the "Global 3-Hour River Flood Reanalysis Data GRFR V1.0" of the Qinghai-Tibet Plateau Scientific Data Center (the data range is from 1979 to 2019), the daily natural flow data of the river sections where Wuzhou Station and Shijiao Station are located from 1979 to 2019 are extracted; collect the daily average sea level and daily maximum tidal range data of Causeway Bay Station, Hong Kong, China outside the estuary of Modaomen Waterway from 1979 to 2019; collect the daily average wind speed and prevailing wind direction data of Waglan Island Station, Hong Kong, China outside the estuary of Modaomen Waterway from 1979 to 2019; conduct quality control on the above data to obtain historical saltwater intrusion disaster simulation data.

[0090] Further, step S102 specifically includes:

[0091] (1) Construct a historical saltwater intrusion disaster simulation model and a saltwater intrusion prediction model.

[0092] Select a multi-layer BP neural network model to construct a historical saltwater intrusion disaster simulation model. Taking the upstream flow, estuary tide level and wind force as saltwater intrusion influencing factors, the simulated salinity values of each main water intake are obtained. Take the flows of Wuzhou Station and Shijiao Station as 2 flow factors, the average sea level and maximum tidal range of the tidal station outside the estuary as 2 tidal factors, and the average wind speed and prevailing wind direction of the meteorological station outside the estuary as 2 wind force factors, a total of 6 model input factors; take the daily average salinity of Zhuzhoutou, Pinggang and Guangchang Pumping Stations as the model simulation objects to construct a historical saltwater intrusion disaster simulation model with a multi-layer neural network structure.

[0093] Select a multi-layer BP neural network model to construct a saltwater intrusion prediction model. Represent the cumulative effect of historical salinity with the past daily average salinity of each main water intake, and take it together with the upstream flow, estuary tide level and wind force as saltwater intrusion influencing factors to obtain the predicted salinity values of each main water intake. Considering the influence of the cumulative effect of past salinity, take the past daily average salinity of each water intake as the salinity factor, together with the flow factor, tidal factor and wind force factor as the model input factors, a total of 9; take the daily average salinity of Zhuzhoutou, Pinggang and Guangchang Pumping Stations as the model simulation objects to construct a saltwater intrusion prediction model with a multi-layer neural network structure. Compared with the historical saltwater intrusion disaster simulation model, the saltwater intrusion prediction model has higher simulation accuracy because it considers the cumulative effect of previous salinity; but also because of this (requiring the salinity factor as an input), it cannot be used for long-term historical saltwater intrusion disaster simulation lacking salinity observation data.

[0094] (2) Use the data of the target sites to train the historical saltwater intrusion disaster simulation model and the saltwater intrusion prediction model respectively.

[0095] Using the model training and test datasets, the two models are trained through repeated iterations. Based on the measured salinity data at the main water intake and the output results of the neural network model, the performance of the models is evaluated to obtain the optimal lag time for each influencing factor and determine the neural network model with the best performance.

[0096] Further, step (2) specifically includes:

[0097] (2-1) Divide the model training set and test dataset into a training set and a test set.

[0098] In this embodiment, the daily measured data from 2019 to 2023 is divided into two subsets, namely 2019 - 2021 and 2022 - 2023, which are used as the training set and the test set respectively.

[0099] (2-2) Use the training set and the test set to train and test the two models respectively.

[0100] Using the particle swarm optimization algorithm, the lag time of each influencing factor, the connection weights and thresholds of the multi-layer neural network are taken together as the position values of the particles, and the performance index value of the neural network model is taken as the fitness value of the particles. Use the training set to perform iterative calculations to determine the position value that maximizes the particle fitness value, that is, obtain the optimal lag time for each influencing factor and the optimal neural network parameters. The corresponding neural network model is the neural network model with the best performance; use the test set to test the optimal neural network model. The salinity changes at each water intake are closely related to external factors such as upstream flow, estuary tides, and wind waves, and generally there is a lag effect for a certain period (in days). Therefore, it is necessary to clarify the optimal lag time for each influencing factor.

[0101] In this embodiment, the lag time of 4 types of influencing factors (salinity factor, flow factor, tide factor, wind force factor), the neural network connection weights and thresholds are taken together as model optimization parameters, and the particle swarm algorithm is used to optimize the parameters. Taking the minimum root mean square difference between the model simulation sequence and the observed data sequence as the optimization goal of the particle swarm algorithm, iterate until 1000 times. Finally, obtain the optimal lag time for each influencing factor and the neural network model with the best performance. Accordingly, the expressions for the long sequence simulation values and future salinity prediction values at the main water intake are:

[0102] Y f =f(D(X))=f(D(X q ,X z ,X w ))

[0103] Y h =h(D(X))=h(D(X s ,X q ,X z ,X w))

[0104] In the above formula, f(·) represents the relationship function between the salinity simulation and influencing factors generated by the historical saltwater intrusion disaster simulation model with optimal performance, and Y f represents the simulated salinity value at the water intake; h(·) represents the relationship function between the salinity prediction and influencing factors generated by the optimal saltwater intrusion prediction model, and Y h represents the predicted salinity value at the water intake; D(X) represents the influencing factor data at the optimal lag time; X s represents the characteristic set of salinity influencing factors, and X q represents the characteristic set of flow influencing factors, and X z represents the characteristic set of tidal influencing factors, and X w represents the characteristic set of wind influencing factors.

[0105] (3) Input the historical saltwater intrusion disaster simulation data into the trained historical saltwater intrusion disaster simulation model to obtain long-sequence historical salinity simulation values; input the monitored data of the target sites into the trained saltwater intrusion prediction model to obtain future salinity prediction values.

[0106] Input the historical saltwater intrusion disaster simulation data collected in step S101 into the trained historical saltwater intrusion disaster simulation model to obtain the daily salinity simulation values of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations in the Modaomen Waterway from 1979 to 2019.

[0107] Input the hydrological, salinity, tidal, meteorological and other monitoring data of each target site in the Modaomen Waterway into the trained saltwater intrusion prediction model to obtain the salinity prediction values of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations in the next k days; k can be any positive integer greater than or equal to 1. In this embodiment, k is taken as 3.

[0108] Further, step S103 specifically includes:

[0109] (1) Construct a comprehensive evaluation index system for water supply risk.

[0110] (1-1) Define the scope of the study area, determine the main water intakes affected by saltwater intrusion, local reservoirs (groups) available for storing fresh water, and water supply objects within the scope; obtain the maximum water intake capacity of the main water intakes, the adjustable storage capacity of the local reservoirs (groups) (the difference between the normal storage capacity and the dead storage capacity), and the water demand of the water supply objects.

[0111] In this embodiment, the Zhuhai-Macao water supply system in China is taken as the object. This system takes 3 pumping stations (Zhuzhoutou, Pinggang, and Guangchang Pumping Stations) in the Modaomen Waterway as the main water intake pumping stations, forming an interconnected system composed of pumping station projects, pipeline network projects, and water supply reservoir groups. The current total water demand of its water supply objects in the dry season is about 1.23 million m 3 / d. The current water intake capacities of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations are 2.7 million m 3 / d, 1.4 million m 3 / d, and 0.8 million m 3 / d respectively. The local reservoirs in the Zhuhai - Macau, China water supply system that undertake the water supply function include Zhuyin Reservoir, the North Reservoir Group, and the South Reservoir Group. Among them, the total storage capacity of Zhuyin Reservoir is 40.18 million m 3 , and the regulating storage capacity is 38.11 million m 3 . The supporting project, Yuekeng Reservoir, has a total storage capacity of 3.15 million m 3 , and the regulating storage capacity is 2.8 million m 3 ; the North Reservoir Group includes three reservoirs, Dajingshan, Phoenix Mountain, and Meixi, with a total storage capacity of 25.56 million m 3 , and the regulating storage capacity is 18.19 million m 3 ; the South Reservoir Group includes four reservoirs, Nanping, Zhuxian Cave, Shedikeng, and Yinkeng, with a total storage capacity of 8.27 million m 3 , and the regulating storage capacity is 5.45 million m 3 . Through the pipeline network project and the pumping stations, the reservoirs (groups) can replenish the storage through Zhuzhoutou, Pinggang, and Guangchang Pumping Stations, and there is basically no pressure for replenishing the storage; theoretically, only when the Zhuzhoutou Pumping Station, which is located at the uppermost reaches of the Modaomen Waterway, is unable to draw water due to excessive salinity, will the water use in Zhuhai and Macau, China completely rely on the local reservoir group for water supply. The data in this embodiment is referenced from the "Implementation Plan for Water Quantity Regulation during the Dry Season of the Pearl River in 2022 - 2023" of the Pearl River Water Resources Commission of the Ministry of Water Resources.

[0112] (1 - 2) Based on the main water intake points, local reservoirs available for storing fresh water, and water supply targets, a regional water resources supply - demand balance model based on the regulation of storing fresh water is constructed.

[0113] The specific expression of the regional water resources supply - demand balance model is:

[0114]

[0115] In the above formula:

[0116] T hour represents the number of hours of excessive salinity at the water intake point, calculated as a function f hour (S) of the daily average salinity S of the water intake point; according to the empirical relationship between the number of hours of daily exceedance and the daily average salinity in the measured salinity data of the water intake point, the function f hour (S) is obtained by fitting; in this embodiment, based on the measured salinity data of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations in the Modaomen Waterway from 2019 to 2023, the piecewise fitting functions of the number of hours of daily exceedance and the daily average salinity of each water intake point are obtained as follows:

[0117] For Zhuzhoutou Pumping Station:

[0118]

[0119] The Pinggang Pumping Station is as follows:

[0120]

[0121] The Guangchang Pumping Station is as follows:

[0122]

[0123] W a represents the daily water intake of the water intake point. When the salinity of the water intake point exceeds the standard, water cannot be taken from the water intake point; W a can be calculated according to the maximum water intake capacity W of the water intake point c and the daily average number of hours T when the salinity exceeds the standard hour ;

[0124] Clarify the dispatching rules for seizing and storing fresh water under the intrusion of salt water: (1) During the intrusion of salt water, each water intake point takes water from the river channel to the greatest extent during the available water intake period. The water intake volume first meets the needs of the water supply objects. If there is still a surplus, the extra water intake volume is used to supplement the local reservoir, and the effective storage volume of the reservoir after replenishment should not exceed the regulating storage capacity of the reservoir (that is, the water level of the reservoir should not exceed the normal storage level); (2) If the water intake volume cannot meet the needs of the water supply objects, supplementary water supply is carried out by discharging water from the local reservoir, and the effective storage volume of the reservoir after supplementary water supply should not be less than 0 (that is, the water level of the reservoir should not be lower than the dead water level); (3) If the water intake volume and the supplementary water supply volume of the reservoir still cannot meet the needs of the water supply objects, then there is a substantial water shortage;

[0125] D a represents the daily water demand of the water supply object; V c represents the regulating storage capacity of the local reservoir, that is, the storage capacity between the normal storage level and the dead water level; V t and V t-1 respectively represent the effective storage volumes of the local reservoir on the current day and the previous day; V a represents the amount of water used to supplement the local reservoir with the excess water intake; W s represents the amount of water that needs to be supplemented by the local reservoir; W d represents the substantial water shortage amount; In this embodiment, the regional water resources supply-demand balance model based on the dispatching of seizing and storing fresh water is used for the dry season of the Modaomen Waterway in the Pearl River Estuary, calculated from September of the current year to April of the following year; Referring to the Zhuhai Water Resources Bulletin, the initial effective storage volume of the local reservoir (group) in the dry season is set to 1 / 3 of the regulating storage capacity. In this embodiment, the available water intake volumes of the 3 water intake points are calculated respectively according to their respective water intake capacities and the daily number of hours of exceeding the standard, and then the daily water intake volume W a is obtained; The Zhuyin Reservoir (including its supporting projects), the North Reservoir Group, and the South Reservoir Group are combined into a whole, and the supply-demand balance calculation is carried out as a whole according to the combined regulating storage capacity.

[0126] (1-3) Select and calculate the quantitative evaluation indicators in the comprehensive evaluation index system of water supply risk.

[0127] Seven quantitative evaluation indicators were selected from three aspects: the probability of fresh water intake, the degree of dependence on local reservoirs and the degree of water shortage, including:

[0128] First, in terms of the probability of fresh water intake, the average fresh water intake probability R1 of the main water intake in the next three days, the minimum fresh water intake probability R2, and the number of days where fresh water cannot be taken R3 are selected, totaling 3 quantitative evaluation indicators, representing the degree of influence of saltwater intrusion on the water intake capacity of the river. In this embodiment, the average values ​​of Zhuzhoutou, Pinggang, and Guangchang pumping stations are taken to calculate R1, R2, and R2.

[0129] Secondly, in terms of the degree of dependence on local reservoirs, two quantitative evaluation indicators are selected: the difference between the average water demand and the average water intake in the next three days, R4, and the average storage rate of the local reservoir (group) in the next three days, R5, representing the dependence and vulnerability of the water supply system on the water supply of local reservoirs. In this embodiment, R4 is calculated based on the total water intake of Zhuzhoutou, Pinggang, and Guangchang pumping stations; and R5 is calculated by combining Zhuyin Reservoir (including its supporting projects), North Reservoir Group, and South Reservoir Group as a whole.

[0130] Then, in terms of the degree of water shortage, two quantitative evaluation indicators were selected, namely the average substantial water shortage in the next three days R6 and the number of substantial water shortage days R7, which represent the severity of substantial water shortage caused by saltwater intrusion.

[0131] Calculate the values ​​of each quantitative evaluation index:

[0132]

[0133] In the above formula, ave(T hour ) represents the average daily exceeding-limit hours of the water intake in the next three days; min(T hour ) represents the longest daily exceeding hours of the water intake in the next 3 days; sum(n) represents the number of days when fresh water cannot be taken from the water intake in the next 3 days. Considering that the actual water pump cannot be started and stopped quickly, the case where the probability of taking fresh water is less than 0.1 ((1-T hour / 24)<0.1) is considered as being unable to be depreciated within one day (n=1); ave(D a ) represents the average daily water demand for the next three days; ave(W a ) represents the average daily water intake in the next three days; ave(P t ) represents the average storage rate of the local reservoir in the next three days, where the storage rate P t Calculated as the total amount of water stored on that day (effective water storage V t and dead storage capacity V s The sum of the water volume and the normal storage capacity of the reservoir V total(Ratio of the corresponding storage capacity at normal pool level); ave(W d ) represents the average water shortage in the next 3 days; sum(m) represents the number of days with substantial water shortage in the next 3 days. If W d >0 on a certain day, it is counted as a day with substantial water shortage (m = 1).

[0134] (2) Use the symbol conversion method to process each quantitative evaluation index in the comprehensive water supply risk evaluation index system to make them all positive indexes.

[0135] Specifically, the larger the values of the indexes R1, R2, and R5, the smaller the water supply risk. They are negative indexes and should be converted from the original positive values to negative values, thus becoming positive indexes; the larger the positive index value, the greater the water supply risk. In this embodiment, the average fresh water intake probability R1, the lowest fresh water intake probability R2, and the average reservoir fullness rate R2 are converted from the original positive values to negative values, thus changing from negative indexes to positive indexes; then, together with other indexes, they are respectively normalized to obtain the normalized positive evaluation index values.

[0136] (3) Calculate the comprehensive water supply risk index according to the values of the processed quantitative evaluation indexes.

[0137] Use the analytic hierarchy process and pairwise judgment method to construct the judgment matrix of 7 quantitative evaluation indexes and calculate the corresponding weight matrix, and calculate the consistency test index CR of the weight matrix; when CR is less than 0.1, the matrix consistency is acceptable, otherwise reconstruct the judgment matrix; according to the determined weight matrix and the values of each evaluation index, calculate the comprehensive water supply risk index RI. In this embodiment, the pairwise judgment method is used to construct the judgment matrix of each index, as shown in Table 1; after calculation, the consistency index CR value of the judgment matrix is 0.095, which is less than 0.1 and meets the consistency requirement.

[0138] Table 1 Judgment matrix of comprehensive water supply risk evaluation indexes

[0139]

[0140] Further, step S104 specifically includes:[[]]

[0141] (1) Conduct frequency analysis on the comprehensive water supply risk index calculated according to the long-sequence historical salinity simulation values, and divide the water supply risk level and early warning level.

[0142] Based on the historical saltwater intrusion disaster simulation dataset from 1979 to 2019 collected in step S101, this embodiment uses the machine learning-based historical saltwater intrusion disaster simulation model trained in step S102 to obtain the daily salinity simulation values of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations from 1979 to 2019. On this basis, using the regional water resources supply-demand balance model based on the freshwater storage and diversion scheduling in step S103, the water resources supply-demand balance state of the Zhuhai-Macao water supply system during the dry seasons from 1979 to 2019 is obtained, and the comprehensive evaluation index values of various water supply risks are calculated daily. Finally, the daily comprehensive water supply risk index of the Zhuhai-Macao water supply system during the dry seasons from 1979 to 2019 is obtained.

[0143] Perform frequency analysis on the daily comprehensive water supply risk index during the long-term dry season: First, exclude the cases where RI < 0.1 from the daily comprehensive water supply risk index during the long-term dry season, because RI < 0.1 represents that the water supply risk is very small, and such cases are expected to account for a large proportion of the long-term daily RI. After exclusion, it will be beneficial to further subdivide the distribution of the long-term daily RI. Then, perform frequency analysis on the remaining long-term daily RI, and count the cumulative frequency P cum (n); if the cumulative frequency of a certain RI value is P cum (n), it means that the probability that the long-term daily RI is less than or equal to this value is n; finally, according to the percentile method, take the RI values corresponding to P cum (30%), P cum (60%), P cum (80%), P cum (90%) as the water supply risk thresholds, corresponding to general, relatively large, major, and extremely major water supply risks respectively, corresponding to 4 warning levels of blue (Ⅳ), yellow (Ⅲ), orange (Ⅱ), and red (Ⅰ) respectively. In this embodiment, through statistical analysis of the daily comprehensive water supply risk index during the dry seasons from 1979 to 2019, it is found that the proportion of days with RI < 0.1 is as high as 72.9%, indicating that the water supply risk of the Zhuhai-Macao water supply system is very small in most cases; further, analyzing the cases where RI ≥ 0.1, it is found that the frequency of RI values not exceeding 0.2 is about 30%, not exceeding 0.3 is about 60%, not exceeding 0.5 is about 80%, and not exceeding 0.8 is about 90%. Thus, determine P cum (30%), P cum (60%), P cum (80%), P cum(90%) The corresponding RI values are 0.2, 0.3, 0.5, and 0.8 respectively. Accordingly, the warning levels for the regional water supply risk under saltwater intrusion are divided as follows: ① When 0.2 ≤ RI < 0.3, the saltwater intrusion is weak and the water supply risk is average, corresponding to the blue (IV) warning; ② When 0.3 ≤ RI < 0.5, the saltwater intrusion is obvious and the water supply risk is relatively large, corresponding to the yellow (III) warning; ③ When 0.5 ≤ RI < 0.8, the saltwater intrusion is relatively serious and the water supply risk is significant, corresponding to the orange (II) warning; ④ When RI ≥ 0.8, the saltwater intrusion is serious and the water supply risk is extremely significant, corresponding to the red (I) warning. In summary, the water supply risk warning and warning levels for the Zhuhai - Macau, China water supply system in the Modaomen Waterway in this embodiment are shown in Table 2.

[0144] Table 2 Risk thresholds and warning levels for the Zhuhai - Macau, China water supply system in the Modaomen Waterway

[0145]

[0146] (2) Determine the warning indicative indicators from aspects such as the flow of the upstream control hydrological station, the probability of taking fresh water at the main water intake, and the average full - storage rate of local reservoirs (groups); divide the risk level warnings according to the warning indicative indicators.

[0147] The long - sequence daily RI is divided into 5 categories (0 ≤ RI < RI 30％ , RI 30％ ≤ RI < RI 60％ , RI 60％ ≤ RI < RI 80％ , RI 80％ ≤ RI < RI 90％ , RI 90％ ≤ RI ≤ 1) according to the risk thresholds. Using the Mann - Whitney U non - parametric test method, select indicators such as the average flow of the upstream control hydrological station, the average probability of taking fresh water at the main water intake, and the average full - storage rate of local reservoirs (groups) in the next 3 days, and analyze whether there are significant differences between different RI ranges. Then, take the indicators with significant differences as the warning indicative indicators. In this embodiment, the daily RI in the dry season from 1979 to 2019 is divided into 5 categories: 0.1 ≤ RI < 0.2, 0.2 ≤ RI < 0.3, 0.3 ≤ RI < 0.5, 0.5 ≤ RI < 0.8, 0.8 ≤ RI ≤ 1; select the combined flow of the upstream Wuzhou Station and Shijiao Station, the average probability of taking fresh water at the main water intake, and the average full - storage rate of local reservoirs (groups) as indicators, and conduct the Mann - Whitney U non - parametric test between different RI ranges. The significant test results are shown in Table 3.

[0148] Table 3 Characteristic comparison of warning indicative indicators between different warning levels

[0149]

[0150] Note: In Table 3, categories O, Ⅳ, Ⅲ, Ⅱ, and Ⅰ represent 0.1 ≤ RI < 0.2, 0.2 ≤ RI < 0.3, 0.3 ≤ RI < 0.5, 0.5 ≤ RI < 0.8, and 0.8 ≤ RI ≤ 1 respectively. The letters A / B / C / D are used to indicate the differences between groups. There are significant differences between two categories with different letters. The numbers in the table represent the means of each category, and the flow unit is m 3 / s.

[0151] In this embodiment, according to the results of Table 3, the total flow of Wuzhou Station and Shijiao Station and the average full storage rate of local reservoirs (groups) are selected as the warning indicative indicators for blue warning (Level Ⅳ) (as No. ④ in Table 3); further, the average fresh water intake probability of the main water intake points and the average full storage rate of local reservoirs (groups) are selected as the warning indicative indicators for yellow warning (Level Ⅲ) (as No. ③ in Table 3); further, the average fresh water intake probability of the main water intake points is selected as the warning indicative indicator for orange warning (Level Ⅱ) (as No. ② in Table 3); further, the total flow of Wuzhou Station and Shijiao Station, the average fresh water intake probability of the main water intake points, and the average full storage rate of local reservoirs (groups) are used as the warning indicative indicators for red warning (Level Ⅰ) (as No. ① in Table 3).

[0152] Calculate the determination thresholds corresponding to the warning indicative indicators, as follows:

[0153]

[0154] In the formula, M Aave and M Bave are taken from the 2 RI ranges to be distinguished by the warning indicative indicator M; the indicator M between the 2 RI ranges has significant differences and is respectively marked as range A and range B, where range B corresponds to a higher warning level; M Aave represents the average in range A; M Bave represents the average in range B; M A-Bth represents the determination threshold of the indicator M between range A and range B, which is used as the criterion for determining the warning level corresponding to entering range B. In this embodiment, calculations are performed according to the means shown in Table 3, and some results are corrected to make them easy to distinguish, obtaining the warning indicative indicators and their thresholds for each level, which are further as follows:

[0155] Blue warning (Level Ⅳ): The total flow of upstream Wuzhou Station and Shijiao Station is lower than 3000 m3 / s, and the average full storage rate of local reservoirs (groups) is lower than 0.8;

[0156] Orange warning (Level Ⅲ): The average fresh water intake probability of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations is lower than 0.6, and the average full storage rate of local reservoirs (groups) is lower than 0.5;

[0157] Yellow Alert (Level II): The average fresh - water intake probability of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations is lower than 0.3, and the average full - storage rate of local reservoirs (groups) is lower than 0.5;

[0158] Red Alert (Level I): The combined flow rate of the upstream Wuzhou Station and Shijiao Station is lower than 1700 m³ / s, or the average fresh - water intake probability of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations is lower than 0.1, and the average full - storage rate of local reservoirs (groups) is lower than 0.2.

[0159] (3) If the comprehensive water supply risk index calculated based on the predicted salinity value in the future or its early - warning indicative index exceeds the set threshold, then issue an early - warning of the corresponding level.

[0160] First, according to the hydrological, salinity, tide, meteorological and other monitoring data of each target station in the Modaomen Waterway, using the salt - water intrusion prediction model trained in step S102, obtain the predicted salinity values of the water intake points of Zhuzhoutou, Pinggang, and Guangchang Pumping Stations in the next 3 days; then, use the regional water resources supply - demand balance model and the comprehensive water supply risk assessment index system established in step S103 for calculation to obtain the comprehensive water supply risk index value and the corresponding early - warning indicative index value; finally, the comprehensive water supply risk index determination method can be used. According to the calculated comprehensive water supply risk index and relevant thresholds, determine the current water supply risk level and early - warning level; the early - warning indicative index determination method can also be used. According to the relevant early - warning indicative index values and relevant thresholds, determine the current early - warning level; the comprehensive water supply risk index determination method and the early - warning indicative index determination method can be used simultaneously and mutually verified.

[0161] Specifically, in this embodiment, based on the comprehensive water supply risk index determination method and the early - warning indicative index determination method, the early - warning levels of the long - sequence historical salt - water intrusion disasters from 1979 to 2019 are divided, and the results are shown in Figure 3 ; based on the observed data in the dry season of 2022, use the salt - water intrusion prediction model to obtain the predicted salinity values of each water intake point, and conduct salt - water intrusion risk early - warning based on the early - warning indicative index determination method. The results are shown in Figure 4 . As Figure 4 shown, compared with the comprehensive water supply risk index determination method, dividing the salt - water intrusion risk level according to the early - warning indicative indexes such as the average flow rate of the hydrological station, the average fresh - water intake probability of the main water intake points, and the average full - storage rate of local reservoirs (groups) is not only more intuitive, but also more conducive to achieving goal - oriented human regulation, and can provide a scientific basis for quickly responding to salt - water intrusion in multiple decision - making stages such as monitoring, evaluation, and adjustment.

[0162] This embodiment should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed within the present invention. Those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiment can be completed by instructing relevant hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

[0163] It should be noted that although the method operations of the above embodiment are described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the depicted steps can be executed in a changed order. Additionally or alternatively, certain steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution.

[0164] Embodiment 2:

[0165] As Figure 5 shown, this embodiment provides a water supply risk assessment system under saltwater intrusion in a tidal estuary area based on the rush storage of fresh water scheduling. The system includes an acquisition module 501, a training module 502, a construction and calculation module 503, and a division module 504, where:

[0166] The acquisition module 501 is used to acquire the daily target site data and historical saltwater disaster simulation data in the same time series in the research area; among them, the target site data includes measured salinity data, flow data, tide level data, and wind force data, and the historical saltwater disaster simulation data includes the daily flow data, tide level data, and wind force data in a long time series;

[0167] The training module 502 is used to train the historical saltwater disaster simulation model with the target site data, and input the historical saltwater disaster simulation data into the trained historical saltwater disaster simulation model to obtain the long-sequence historical salinity simulation values;

[0168] The construction and calculation module 503 is used to generalize the water supply system in the research area, construct a comprehensive water supply risk evaluation index system based on the rush storage of fresh water scheduling; based on the comprehensive water supply risk evaluation index system, calculate the comprehensive water supply risk index according to the long-sequence historical salinity simulation values; among them, the water supply system includes the main river channel water intake, local water storage reservoirs, and water supply objects;

[0169] The division module 504 is used to perform frequency analysis on the comprehensive water supply risk index, and divide the water supply risk level and early warning level based on the set water supply risk threshold.

[0170] For the specific implementation of each module in this embodiment, reference can be made to Embodiment 1 above, which will not be elaborated here one by one. It should be noted that the system provided in this embodiment is only illustrated by the above division of each functional module. In actual applications, the above functions can be allocated to different functional modules as needed, that is, the internal structure is divided into different functional modules to complete all or part of the functions described above.

[0171] Embodiment 3:

[0172] This embodiment provides a terminal device, which can be a computer. As Figure 6 shown, it includes a processor 602, a memory, an input device 603, a display 604, and a network interface 605 connected through a system bus 601. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium 606 and an internal memory 607. The non-volatile storage medium 606 stores an operating system, a computer program, and a database. The internal memory 607 provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the processor 602 executes the computer program stored in the memory, the method for evaluating the water supply risk under the saltwater intrusion in the tidal estuary area based on the grab and storage of fresh water scheduling in Embodiment 1 above is implemented as follows:

[0173] Obtain the daily target site data and historical saltwater disaster simulation data in the same time series in the study area; among them, the target site data includes measured salinity data, flow data, tide level data, and wind force data, and the historical saltwater disaster simulation data includes daily flow data, tide level data, and wind force data in a long time series;

[0174] Use the target site data to train the historical saltwater disaster simulation model, and input the historical saltwater disaster simulation data into the trained historical saltwater disaster simulation model to obtain long-sequence historical salinity simulation values;

[0175] Generalize the water supply system in the study area, and construct a comprehensive evaluation index system for water supply risk based on the grab and storage of fresh water scheduling; based on the comprehensive evaluation index system for water supply risk, calculate the comprehensive water supply risk index according to the long-sequence historical salinity simulation values; among them, the water supply system includes the main river intake, local storage reservoirs, and water supply objects;

[0176] Conduct frequency analysis on the comprehensive water supply risk index, and divide the water supply risk level and early warning level based on the set water supply risk threshold.

[0177] Embodiment 4:

[0178] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for assessing the water supply risk under saltwater intrusion in a tidal estuary area based on the grab and storage of fresh water scheduling in the above-mentioned Embodiment 1, as follows:

[0179] Obtain the daily target site data and historical saltwater intrusion disaster simulation data in the same time series in the study area; among them, the target site data includes measured salinity data, flow data, tide level data, and wind force data, and the historical saltwater intrusion disaster simulation data includes daily flow data, tide level data, and wind force data in a long time series;

[0180] Use the target site data to train the historical saltwater intrusion disaster simulation model, and input the historical saltwater intrusion disaster simulation data into the trained historical saltwater intrusion disaster simulation model to obtain long-sequence historical salinity simulation values;

[0181] Generalize the water supply system in the study area and construct a comprehensive evaluation index system for water supply risk based on the grab and storage of fresh water scheduling; based on the comprehensive evaluation index system for water supply risk, calculate the comprehensive water supply risk index according to the long-sequence historical salinity simulation values; among them, the water supply system includes the main river channel water intake, local water storage reservoir, and water supply objects;

[0182] Conduct frequency analysis on the comprehensive water supply risk index, and divide the water supply risk level and warning level based on the set water supply risk threshold.

[0183] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0184] In summary, the method, system, device and storage medium for evaluating the water supply risk under the saltwater intrusion in the tidal estuary area based on the fresh water storage and diversion scheduling provided by the present invention include: obtaining data such as the salinity of the water intake, the flow rate of the hydrological station, the tidal level of the tidal station, and the wind force of the meteorological station at the target site in the study area; training a historical saltwater intrusion disaster simulation model based on machine learning using the obtained data, and obtaining the long-sequence historical salinity simulation values of the main water intakes based on the historical saltwater intrusion disaster simulation model; generalizing the water supply system in the study area and constructing a comprehensive evaluation index system for water supply risk based on the fresh water storage and diversion scheduling; calculating the comprehensive water supply risk index of the study area according to the salinity simulation values based on the comprehensive evaluation index system for water supply risk based on the fresh water storage and diversion scheduling; performing frequency analysis on the comprehensive water supply risk index under the long-sequence historical saltwater intrusion simulation, and dividing the water supply risk threshold and early warning level. The present invention quantifies the regional water supply risk based on the fresh water storage and diversion scheduling rules under the saltwater intrusion, and scientifically divides the saltwater intrusion risk level from the perspectives of the influence degree and occurrence frequency of the long-sequence historical saltwater intrusion disasters. In addition, the early warning indicative indicators are determined from aspects such as the flow rate of the upstream controlled hydrological station, the fresh water intake probability of the main water intakes, and the average full storage rate of the local reservoir group; according to the early warning indicative indicators, corresponding-level early warnings are issued. The results of this technical solution are not only more intuitive, but also more conducive to achieving goal-oriented artificial regulation, and can provide a scientific basis for quickly responding to saltwater intrusion in multiple decision-making stages such as monitoring, evaluation, and adjustment. In addition, the present invention also constructs a saltwater intrusion prediction model for predicting the future salinity prediction values of the main water intakes, and further predicting the water supply risk level under the future saltwater intrusion. The present invention provides an effective technical method for judging the water supply safety in the estuary and coastal areas relying on the tidal river section as the water source, and can be widely applied to the technical field of saltwater intrusion risk assessment.

[0185] As mentioned above, only the preferred embodiments of the present invention for patents are described, but the protection scope of the present invention for patents is not limited thereto. Any person skilled in the art within the scope disclosed by the present invention for patents, according to the technical solution and inventive concept of the present invention for patents, makes equivalent substitutions or changes, and all belong to the protection scope of the present invention for patents.

Claims

1. A water supply risk assessment method for saltwater intrusion in tidal estuaries based on emergency freshwater storage scheduling, characterized in that: The method comprises: Obtain daily target site data and historical saltwater disaster simulation data in the same time series in the study area; the target site data include measured salinity data, flow data, tide data and wind data, and the historical saltwater disaster simulation data include daily flow data, tide data and wind data in a long time series; The historical saltwater disaster simulation model is trained using the target site data, and the historical saltwater disaster simulation data is input into the trained historical saltwater disaster simulation model to obtain a long sequence of historical salinity simulation values; the historical saltwater disaster simulation model is a multi-layer BP neural network model; The water supply system in the study area is generalized, and a comprehensive water supply risk evaluation index system based on emergency freshwater scheduling is constructed; based on the comprehensive water supply risk evaluation index system, the comprehensive water supply risk index is calculated according to the long-sequence historical salinity simulation value; the water supply system includes the main water intake of the river, local water storage reservoirs and water supply objects; Conduct frequency analysis on the comprehensive water supply risk index and divide the water supply risk degree and warning level based on the set water supply risk threshold; The method of using the target site data to train the historical saltwater disaster simulation model includes: The measured flow data, tide data and wind data are used as saltwater influencing factors to input into the historical saltwater disaster simulation model, and the salinity simulation values ​​of the main water intakes of each river are output; According to the measured salinity data of the main water intakes of each river and the long-sequence historical salinity simulation values, the performance of the historical saltwater disaster simulation model was evaluated, and the optimal lag time of each saltwater influencing factor was obtained. The corresponding BP neural network model is the trained historical saltwater disaster simulation model. The optimal lag time of each saltwater tide influencing factor is obtained by the following process: Using the particle swarm optimization algorithm, the lag time of each saltwater influencing factor, the connection weight and threshold of the multi-layer BP neural network are taken as the position value of the particle. The performance index value of the BP neural network model is used as the fitness value of the particle. The target site data is used for iterative calculation to determine the position value that maximizes the particle fitness value, that is, the optimal lag time and optimal neural network parameters of each influencing factor are obtained. The corresponding BP neural network model is the BP neural network model with the best performance.

2. The method for assessing water supply risk under saltwater intrusion in tidal estuary areas according to claim 1 is characterized in that: The comprehensive water supply risk evaluation index system is based on which the comprehensive water supply risk index is calculated according to the long-sequence historical salinity simulation value, including: Based on the comprehensive evaluation index system of water supply risk, the values ​​of each quantitative evaluation index in the comprehensive evaluation index system of water risk are calculated according to the long-sequence historical salinity simulation values, the maximum water intake capacity of the main water intakes of the river, the adjustable storage capacity of the local water storage reservoirs and the water demand of the water supply objects; The symbol conversion method is used to process the values ​​of each quantitative evaluation index to obtain the normalized positive index value of each quantitative evaluation index; According to the processed values ​​of each quantitative evaluation index, the judgment matrix is ​​constructed by using the hierarchical analysis method and the pairwise judgment method and the corresponding weight matrix is ​​calculated; if the consistency test index of the weight matrix is ​​greater than or equal to the first set threshold, the judgment matrix is ​​reconstructed; The comprehensive water supply risk index is calculated based on the corresponding weight matrix of the determined judgment matrix and the values ​​of each quantitative evaluation index after processing.

3. The method for assessing water supply risk under saltwater intrusion in tidal estuaries according to any one of claims 1 and 2, characterized in that: The water supply system in the study area is generalized, and a comprehensive evaluation index system for water supply risk based on emergency storage of fresh water is constructed, including: Quantitative evaluation indicators are selected from the probability of fresh water intake at the main river intake, the degree of dependence on local water storage reservoirs and the degree of water shortage of water supply objects, including: From the perspective of the probability of taking fresh water from the main water intake of the river, the average probability of taking fresh water R1, the minimum probability of taking fresh water R2 and the number of days where fresh water cannot be taken R3 of the main water intake of the river in the next k days are selected; k is a positive integer greater than or equal to 1; From the perspective of local water storage reservoir dependence, the difference between the average water demand and average water withdrawal in the next k days, R4, and the average storage rate in the next k days, R5, are selected; In terms of water shortage degree, the average substantial water shortage amount R6 and the number of substantial water shortage days R7 in the next k days are selected; The calculation formulas for each quantitative evaluation index are as follows: in: Where, T hour is the number of hours when the salinity of the main water intake of the river exceeds the standard, f hour (S) is the function obtained by fitting the empirical relationship between the daily salinity exceeding the standard and the daily average salinity in the measured salinity data, and S is the long-series historical salinity simulation value; ave(T hour ) is the average daily exceeding-limit hours of the main water intake of the river in the next k days; min(T hour ) represents the longest daily exceeding-limit hours at the main water intake of the river in the next k days; sum(n) represents the number of days when fresh water cannot be taken at the main water intake of the river in the next k days. Considering that the actual water pump cannot be started and stopped quickly, the fresh water taking probability is 1-T hour / 24 is less than 0.1, which means that it is impossible to take a short position within one day, that is, n = 1; D a is the daily water demand of the water supply object, ave(D a ) is the average daily water demand in the next k days; W a is the daily water intake of the main water intake of the river, ave(W a ) is the average daily water intake in the next k days; the storage rate P t is the effective water storage capacity V on that day t and dead storage capacity V s The sum of the normal storage capacity of the reservoir V total The ratio of ave(P t ) represents the average filling rate of the local water reservoir in the future k days; W d is the substantial water deficit, ave(W d ) represents the average water shortage in the next k days; sum(m) represents the number of days with substantial water shortage in the next k days. d >0, it is considered that substantial water shortage occurs on that day, that is, m=1; W c V is the maximum water intake capacity of the main water intake of the river; a Indicates the amount of excess water used to supplement local water storage reservoirs; W s The amount of water required to supplement the water supply from local water storage reservoirs; V t and V t-1 They represent the effective water storage capacity of the local water storage reservoir on the current day and the previous day respectively; k is a positive integer greater than or equal to 1; The rules for emergency freshwater storage and scheduling under saltwater intrusion are as follows: when saltwater intrusion occurs, the main water intakes of each river channel will draw water from the river channel to the maximum extent during the water-drawing period, and the water intake will give priority to meeting the needs of the water supply objects. If there is still surplus water, the excess water intake will be used to supplement the local water storage reservoir. The effective water storage capacity of the reservoir after replenishment should not exceed the regulating capacity of the reservoir; if the water intake cannot meet the needs of the water supply objects, water will be released from the local reservoir for supplementary water supply, and the effective water storage capacity of the reservoir after supplementary water supply should not be less than 0; if the water intake and the reservoir supplementary water supply still cannot meet the needs of the water supply objects, a substantial water shortage will occur.

4. The method for assessing water supply risk under saltwater intrusion in tidal estuary areas according to claim 1, characterized in that: The frequency analysis of the comprehensive water supply risk index is carried out to divide the water supply risk degree and warning level based on the set water supply risk threshold, including: Firstly, the comprehensive water supply risk index less than the second set threshold is eliminated from the comprehensive water supply risk index; Then, the frequency analysis of the remaining long-sequence daily comprehensive water supply risk index was carried out to count the cumulative frequencies corresponding to different comprehensive water supply risk indexes; Finally, based on the set water supply risk threshold, the water supply risk degree and warning level are divided according to the cumulative frequency.

5. The method for assessing water supply risk under saltwater intrusion in tidal estuary areas according to claim 1, characterized in that: The long time series is no less than 20 years.

6. The method for assessing water supply risk under saltwater intrusion in tidal estuary areas according to any one of claims 1, 2, 4 and 5, characterized in that: After conducting a frequency analysis of the comprehensive water supply risk index, the method also includes determining early warning indicative indicators from the flow of upstream control hydrological stations, the probability of fresh water intake at the main river intake, and the average filling rate of local water storage reservoirs; According to the early warning indicative indicators, risk level warnings are divided, including: Based on the set water supply risk threshold, the comprehensive water supply risk index RI is divided into corresponding categories; Using the Mann-Whitney U nonparametric test method, we selected three indicators: the average flow data for the next k days, the average probability of fresh water intake at the main water intake of the main river in the next k days, and the average fullness rate of local water storage reservoirs in the next k days. We analyzed the differences between the three indicators in different categories and used the indicators with significant differences as early warning indicators. If the warning indicator M is greater than the judgment threshold, it is classified into the corresponding warning level; Wherein, the determination threshold is: Where M Aave and M Bave They are taken from the two types of RI that the indicator M is intended to distinguish; the indicator M between the two types of RI has significant differences and is marked as range A and range B, respectively, where range B corresponds to a higher level of warning level; M Aave is the average in range A; M Bave is the average in range B; M A-Bth It is the judgment threshold of indicator M between range A and range B, and serves as the judgment criterion for the warning level corresponding to the indicator M in range A entering range B.

7. The method for assessing water supply risk under saltwater intrusion in tidal estuary areas according to any one of claims 1, 2, 4, and 5, characterized in that: After obtaining the daily target station data in the same time series in the study area, the method further includes: The saltwater intrusion prediction model is trained using the target site data, and the monitored target site data is input into the trained saltwater intrusion prediction model to obtain the salinity forecast value for the next f days; the saltwater intrusion prediction model and the historical saltwater disaster simulation model both adopt a multi-layer BP neural network model; wherein f is a positive integer greater than or equal to 1; Based on the comprehensive water supply risk evaluation index system, the comprehensive water supply risk index is calculated according to the future salinity forecast value; Water supply risk levels and warning levels are divided according to the comprehensive water supply risk index and set thresholds.

8. The method for assessing water supply risk under saltwater intrusion in tidal estuaries according to claim 7, characterized in that: The method of training the saltwater intrusion prediction model using the target site data includes: The daily average salinity in the salinity data is used as the cumulative effect of historical salinity, and together with the flow data and wind data, is input into the saltwater intrusion forecast model as a factor affecting saltwater intrusion to obtain the future salinity forecast value. According to the salinity data and future salinity forecast values, the performance of the saltwater intrusion forecast model is evaluated, and the optimal lag time of each saltwater influencing factor is obtained. The corresponding BP neural network model is the trained saltwater intrusion forecast model.

Images