Water-energy coupled water resource planning method, device, equipment and medium

Abstract

This invention discloses a water-energy coupled water resources planning method, apparatus, equipment, and medium. The method includes: collecting and processing historical index values ​​corresponding to water demand prediction for each water-using sector to obtain the upper and lower limits of each index for the planning year, and inputting them into a trained water demand prediction model to obtain the water demand interval prediction results for each water-using sector; collecting and preprocessing historical water resource availability and energy availability data; constructing a water-energy system coupling risk analysis model based on the Copula function to quantitatively analyze the coupling risk of the water-energy system, and setting four coupling risk scenarios: low, medium-low, medium-high, and high; constructing a multi-objective interval stochastic chance constrained programming model, which is decomposed into two deterministic sub-models; solving the two deterministic sub-models using a hybrid intelligent algorithm, and integrating the solution results to obtain and output the regional water resources optimization allocation results adapted to the four risk scenarios of low, medium-low, medium-high, and high.

Description

Technical Field

[0001] This invention belongs to the field of water resource management technology, specifically relating to a water-energy coupled water resource planning method, device, equipment, and medium. Background Technology

[0002] Driven by both intensifying global climate change and ongoing adjustments in regional development patterns, the uncertainty and complexity of water resource systems are becoming increasingly prominent. Traditional water resource management models are no longer adequate to meet the practical needs of multi-objective collaborative governance. Moreover, regional water resource allocation is closely linked to energy and environmental systems, requiring the consideration of multiple objectives such as energy consumption control, carbon emission constraints, and system resilience enhancement. This places higher demands on the comprehensiveness and adaptability of management models. Water resource allocation issues from the perspective of regional water-energy linkages, due to their involvement of multiple sectors, multiple time scales, nodal relationships, risk correlations, and parameter fluctuations, urgently require a comprehensive modeling framework that integrates prediction, risk assessment, and system uncertainty optimization.

[0003] In recent years, demand forecasting technology has been widely used in water resource management, especially in modeling water use behavior under the influence of urban expansion, industrial layout, and climate drivers. However, single forecasting models struggle to handle the uncertainties of future scenarios. Meanwhile, the operation and scheduling of water systems are often accompanied by significant energy consumption, particularly in pumping, transmission, and treatment, where energy-water coupling problems are becoming increasingly prominent. Therefore, it is necessary to introduce risk quantification mechanisms and establish weight adjustment mechanisms that dynamically link with risk scenarios to improve the robustness and adaptability of the models.

[0004] Therefore, constructing a regional water resource management model that combines uncertain demand forecasting, risk quantification and correlation, and explicitly considers energy consumption in the water system, not only helps improve resource allocation efficiency and system stability, but also provides technical support for achieving low-carbon and sustainable development goals. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a water-energy coupled water resource planning method. This method achieves scientific prediction and dynamic optimization of water demand from multiple water-using sectors by analyzing risk scenarios that couple water resources and energy systems. This improves water resource utilization efficiency under complex and uncertain conditions, while taking into account both energy consumption and system sustainability.

[0006] The second objective of this invention is to provide a water-energy coupled water resource planning device.

[0007] A third objective of this invention is to provide an electronic device.

[0008] A fourth objective of this invention is to provide a computer storage medium.

[0009] The technical solution of the present invention to solve the above-mentioned technical problems is:

[0010] A water-energy coupled water resource planning method includes the following steps:

[0011] Step 1: Collect historical index values ​​for water demand forecasting for each water-using sector within the water resource optimization allocation area. Process the historical index values ​​using the exponential smoothing forecasting method with fluctuation coefficient correction to obtain the upper and lower limit ranges of each index for the planning year. Input the upper and lower limit ranges of each index for the planning year into the water demand forecasting model. Use the water demand forecasting model to predict the water demand of each water-using sector to obtain the water demand range forecast results for each water-using sector.

[0012] Step 2: Collect and preprocess historical water availability data and historical energy availability data for the water resource optimization allocation area; construct a water-energy system coupling risk analysis model based on the Copula function to achieve quantitative analysis of the coupling risk of the water-energy system; combine the preset risk threshold and the quantitative analysis results to classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high, clarify the current risk level of the water-energy system, and map the risk parameters corresponding to the risk level to the default probability of the opportunity constraint;

[0013] Step 3: Based on the water demand range prediction results of each water-using sector and the quantitative analysis results output by the water-energy system coupled risk analysis model, with maximizing the comprehensive benefits of the system as the core optimization objective, and incorporating the joint risk constraints of the system, water supply and demand balance constraints, total water resource balance constraints, energy availability constraints and carbon emission control constraints as constraints, a multi-objective interval stochastic chance constrained programming model is constructed.

[0014] Step 4: Use an interactive algorithm to decompose the multi-objective interval stochastic chance constrained programming model into a deterministic sub-model representing the most optimistic scenario and a deterministic sub-model representing the most conservative scenario;

[0015] Step 5: Use a hybrid intelligent algorithm to solve the two deterministic sub-models respectively. Based on the solution results, integrate them to obtain the regional water resource optimization allocation results that are suitable for four types of coupled risk scenarios: low, medium-low, medium-high, and high.

[0016] In a preferred embodiment of the present invention, in step 1, the water-using sectors in the water resource optimization allocation area include the agricultural sector, the industrial sector, the residential sector, and the ecological environment sector. The historical indicators for the agricultural sector include the output value of the primary industry, water consumption per 10,000 yuan of agricultural output value, irrigated arable land area, and irrigation energy consumption. The historical indicators for the industrial sector include the output value of the secondary industry, water consumption per 10,000 yuan of industrial output value, the number of industrial enterprises above a certain scale, and industrial production energy consumption. The historical indicators for the residential sector include population density, per capita residential water consumption, GDP, and residential water energy consumption. The historical indicators for the ecological environment sector include urban green space area, rainfall, changes in groundwater reserves, and renewable energy utilization rate.

[0017] In a preferred embodiment of the present invention, the specific steps for processing the historical indicator values ​​using the exponential smoothing prediction method with volatility coefficient correction in step 1 are as follows:

[0018] A smoothing coefficient is determined, and an initial smoothing value is determined based on historical indicator values. The smoothing value is iteratively calculated, and the smoothing coefficient is adjusted according to the iteration results and mean square error to complete the point prediction of the indicator values ​​for the planning year. The point prediction is converted into an interval prediction using the fluctuation coefficient correction method. The historical coefficient of variation of each historical indicator value is calculated, which is the ratio of the standard deviation to the mean. The indicator values ​​obtained by the exponential smoothing prediction method are then corrected to the upper and lower limit interval values ​​of each indicator for the planning year based on the historical coefficient of variation.

[0019] In a preferred embodiment of the present invention, the construction and training steps of the water demand prediction model in step 1 are as follows:

[0020] Step 101: Collect historical indicator values ​​of the agricultural, industrial, residential, and ecological environmental sectors within the water resource optimization allocation area; standardize the collected historical indicator values, imputate missing values, and remove outliers to obtain a standardized dataset; divide the standardized dataset into a training set and a validation set according to a preset ratio.

[0021] Step 102: Construct a BP neural network model, wherein the input layer of the BP neural network model corresponds to the historical index values ​​of each water-using sector, and the output layer corresponds to the water demand of each water-using sector;

[0022] Step 103: Using the root mean square error as the evaluation criterion for the fitness function, a genetic algorithm is used to globally optimize the initial weights and thresholds of the BP neural network model, output the optimal initial weights and thresholds, and replace the random initial parameters in the BP neural network model with the optimal initial weights and thresholds.

[0023] Step 104: Iteratively train the optimized BP neural network model based on the training set, and use the validation set to monitor the loss function value of the model in real time to determine whether the loss function value is less than a preset threshold or whether the current iteration number has reached a preset upper limit; if either judgment condition is met, stop training and obtain the trained water demand prediction model.

[0024] In a preferred embodiment of the present invention, the construction and quantitative analysis process of the water-energy system coupling risk analysis model in step 2 is as follows:

[0025] Step 201: Collect historical water availability data and historical energy availability data within the water resource optimization allocation area, and perform outlier removal and standardization processing in sequence to obtain water availability dataset and energy availability dataset;

[0026] Step 202: For the preprocessed water resource availability dataset and energy availability dataset, the marginal distribution is fitted using normal distribution, log-normal distribution and Weibull distribution respectively. The root mean square error is used as the evaluation index of the fitting effect. The root mean square error corresponding to each probability distribution model is calculated. The probability distribution model with the smallest root mean square error is selected as the optimal marginal distribution for water resource availability and energy availability respectively.

[0027] Step 203: Based on the optimal marginal distribution of water resource availability and energy availability, four Copula functions, namely Clayton, Frank, Gumbel, and Student-t, are used to construct a joint probability distribution model of water resource availability and energy availability. The Copula function with the smallest root mean square error and mean square error, and the lowest Akaike information criterion and Bayesian information criterion, is selected as the core mathematical tool to characterize the coupling relationship between water resources and energy, so as to construct a water-energy system coupling risk analysis model.

[0028] Step 204: Pre-set the risk threshold for water-energy composite risk. Based on the water-energy system coupled risk analysis model, calculate the joint probability of water resource and energy availability. Combine the optimal marginal distribution of the water resource subsystem and the energy subsystem. Determine the feasible combination of water resource subsystem risk and energy subsystem risk through probability inverse reasoning. Then, combine the preset risk threshold with the feasible combination of water resource subsystem and energy subsystem risk to classify the water-energy system coupled risk into four categories: low, low-medium, medium-high, and high. Clarify the current risk level of the water-energy system. Map the core risk parameters corresponding to the risk level to the default probability of the opportunity constraint.

[0029] In a preferred embodiment of the present invention, in step 3, the objective function in the multi-objective interval stochastic chance constrained programming model includes a first sub-objective, a second sub-objective, and a third sub-objective, wherein,

[0030] The first sub-objective is to maximize the profits generated by water use in each water-using sector;

[0031] The second sub-objective is to minimize the total energy consumption of the entire water-energy coupling system.

[0032] The third sub-objective is to minimize the overall carbon emissions of the system operation;

[0033] The first, second, and third sub-objectives are assigned corresponding weight coefficients, and the weight coefficients of the first, second, and third sub-objectives are dynamically adjusted according to different water-energy coupling risk scenarios. The sum of the products of the first, second, and third sub-objectives and their corresponding weight coefficients is taken as the comprehensive benefit of the system.

[0034] In a preferred embodiment of the present invention, in step 5, the solution steps of the intelligent hybrid algorithm are as follows:

[0035] Using the weighted objective function corresponding to each deterministic sub-model as the fitness function, a global pre-search is performed on a single deterministic sub-model using the particle swarm optimization algorithm to obtain the initial optimal solution set of each deterministic sub-model. The initial optimal solution set is then merged with the random initial solution generated by the NSGA-II algorithm to form the initial population of the NSGA-II algorithm. The initial population is then subjected to non-dominated sorting, crowding calculation, selection, crossover, and mutation operations in sequence. After the iteration terminates, the solution of each deterministic sub-model is completed, and the final solution set of the corresponding deterministic sub-model is obtained.

[0036] A water-energy coupled water resource planning device, comprising:

[0037] The water demand range prediction module is used to collect historical index values ​​corresponding to each water-using sector within the water resource optimization allocation area, process the historical index values ​​using the exponential smoothing prediction method with fluctuation coefficient correction, and obtain the upper and lower limit ranges of each index in the planning year; and input the upper and lower limit ranges of each index in the planning year into the water demand prediction model, and predict the water demand of each water-using sector through the water demand prediction model, and output the water demand range prediction results of each water-using sector.

[0038] The coupling risk scenario quantification module is used to collect historical water resource availability data and historical energy availability data of the water resource optimization allocation area, and preprocess the data; construct a water-energy system coupling risk analysis model based on the Copula function to realize quantitative analysis of the coupling risk of the water-energy system; combine the preset risk threshold and the quantitative analysis results to classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high, clarify the current risk level of the water-energy system, and map the risk parameters corresponding to the risk level to the default probability of the opportunity constraint;

[0039] The optimization model construction module is used to construct a multi-objective interval stochastic chance constrained programming model based on the water demand interval prediction results of each water-using sector and the quantitative analysis results output by the coupled risk scenario quantification module. The core optimization objective is to maximize the comprehensive benefits of the system. At the same time, the system joint risk constraints, water supply and demand balance constraints, total water resource balance constraints, energy availability constraints and carbon emission control constraints are incorporated as constraints.

[0040] The model decomposition module is used to decompose the multi-objective interval stochastic chance constrained programming model into a sub-model representing the most optimistic scenario and a sub-model representing the most conservative scenario using an interactive algorithm.

[0041] The model solving and result integration module is used to solve the two sub-models separately using a hybrid intelligent algorithm. Based on the solution results, it integrates and outputs the regional water resource optimization allocation results adapted to four types of coupled risk scenarios: low, medium-low, medium-high, and high.

[0042] An electronic device includes a central processing unit and a memory, the central processing unit being configured to invoke and run a computer program stored in the memory to perform the steps of the water-energy coupled water resources planning method.

[0043] A computer-readable storage medium stores, in the form of computer-readable instructions, a computer program implementing the water-energy coupled water resources planning method, which, when invoked by a computer, executes the steps included in the corresponding method.

[0044] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0045] 1. The water-energy coupled water resource planning method of this invention uses a water demand prediction model combined with an exponential smoothing interval prediction method with fluctuation coefficient correction to reasonably predict the water demand intervals of major water-using sectors within the water resource optimization allocation area. Compared with a single prediction model, it significantly improves prediction accuracy and interval reliability. At the same time, based on Copula analysis, a water-energy coupled risk analysis model is constructed. Through the water-energy coupled risk analysis model, the coupling risk of the water-energy system is divided into four major risk scenarios: low, medium-low, medium-high, and high. A dynamic weight association mechanism is established between the risk scenarios and the three core objectives of economic benefits, energy consumption control, and carbon emission constraints to achieve risk-oriented multi-objective collaborative optimization. On this basis, an interval stochastic chance constraint planning model considering multiple uncertainties is constructed. The energy consumption parameters of the entire water system chain are introduced and solved through a hybrid intelligent algorithm. Finally, a regional water resource optimization allocation scheme that adapts to different risk scenarios and takes into account energy conservation and low-carbon development is output.

[0046] 2. In terms of data support, the water-energy coupled water resources planning method of the present invention improves the robustness and scientificity of the prediction of core indicators by using an adaptively adjusted water demand prediction model through algorithm parameter iteration and combining it with the exponential smoothing method with fluctuation coefficient correction to carry out interval prediction, thus providing an accurate and reliable data foundation for water resources allocation.

[0047] 3. In terms of uncertainty and risk handling, the water-energy coupled water resources planning method of the present invention integrates Copula analysis, interval stochastic programming and chance-constrained programming into a multi-objective interval stochastic chance-constrained programming method. It can simultaneously handle multiple uncertainties and accurately quantify them, and cope with the coupling risks of water-energy systems, thereby enhancing its adaptability to complex scenarios.

[0048] 4. The water-energy coupled water resource planning method of the present invention designs multiple risk scenarios from the perspective of water-energy related risk linkage in resource allocation, and establishes a correlation mechanism between risk scenarios and dynamic weights of objective functions, thereby realizing risk-oriented water resource optimization allocation and making the allocation results more in line with the actual needs of different risk levels.

[0049] 5. The water-energy coupled water resource planning method of the present invention incorporates the energy intensity and carbon emission of the entire water production chain into the green development, thereby effectively identifying key nodes with high energy consumption and high carbon emissions, pointing the way for energy conservation and carbon reduction optimization in water resource planning, and contributing to the sustainable development of water resource allocation. Attached Figure Description

[0050] Figure 1 This is a schematic flowchart of the water-energy coupled water resource planning method of the present invention.

[0051] Figure 2 This is a flowchart of the water-energy coupled water resource planning method of the present invention.

[0052] Figure 3 This is a graph showing the training results of the GA-BP neural network model (water demand prediction model).

[0053] Figure 4 The figure shows the fitting results of the joint distribution of water resource availability and energy availability.

[0054] Figure 5 This is a graph showing the water allocation results for each water-using department.

[0055] Figure 6 A diagram showing the water supply results from different water sources. Detailed Implementation

[0056] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0057] Example 1

[0058] See Figures 1-6 The water-energy coupled water resource planning method of the present invention includes the following steps:

[0059] Step 1: Collect historical index values ​​for each water-using sector within the water resource optimization allocation area, and process the historical index values ​​using the exponential smoothing prediction method with fluctuation coefficient correction to obtain the upper and lower limit ranges of each index for the planning year; input the upper and lower limit ranges of each index for the planning year into the water demand prediction model, and predict the water demand of each water-using sector through the water demand prediction model to obtain the water demand range prediction results for each water-using sector.

[0060] The water-using sectors in the water resource optimization allocation area include agriculture, industry, residential use, and the ecological environment sector. Historical indicators for the agriculture sector include primary industry output value, water consumption per 10,000 yuan of agricultural output value, irrigated arable land area, and irrigation energy consumption. Historical indicators for the industry sector include secondary industry output value, water consumption per 10,000 yuan of industrial output value, number of large-scale industrial enterprises, and industrial production energy consumption. Historical indicators for the residential sector include population density, per capita residential water consumption, GDP, and residential water energy consumption. Historical indicators for the ecological environment sector include urban green space area, rainfall, changes in groundwater reserves, and renewable energy utilization rate.

[0061] In this embodiment, historical data from 2005 to 2025 (20 years) of data on indicators and water demand (annual data for each indicator in the agricultural, industrial, domestic, and ecological sectors) were collected for the input layer of the BP neural network model in the study area. The data sources included the *Regional Statistical Yearbook*, *Water Resources Bulletin*, and *Energy Statistical Yearbook*. The historical input indicator values ​​were preprocessed (including outlier removal, missing value imputation, and normalization) to obtain a standardized dataset. This standardized dataset was then divided into a training set and a validation set in a 7:3 ratio: 14 years of data from 2005 to 2018 were used for model training, and 6 years of data from 2019 to 2025 were used for model validation. The training set was used to optimize model parameters, and the validation set was used to verify prediction accuracy. This provides high-quality data support for subsequent iterative optimization and fine-tuning of the water demand prediction model, ensuring its stability and predictive reliability.

[0062] The water demand prediction module is used to respond to the impact of future demand and system fluctuation uncertainties on regional water resource demand, so as to achieve the goal of effectively predicting the annual water demand of various water-using sectors. The specific construction and training steps are as follows:

[0063] A backpropagation (BP) neural network model is constructed. The input layer of the BP neural network model corresponds to the historical index values ​​of each water-using sector, and the output layer corresponds to the water demand of each water-using sector, thus forming an input layer with 4 nodes and an output layer with 1 node. The number of hidden layers and nodes are set to 2 and 5 respectively to balance prediction accuracy and computational efficiency. The activation function adopted is the Leaky ReLU function. The Leaky ReLU function avoids gradient vanishing by preserving small gradients in the negative interval, thus being compatible with both stable data from industrial and residential sectors and fluctuating data from agricultural and ecological sectors. The expression of the Leaky ReLU function is as follows:

[0064] ;

[0065] A parameter-adaptive genetic algorithm (GA) is employed to optimize the structure and internal parameters of a backpropagation (BP) neural network model, thereby constructing a high-precision water demand prediction model. While the BP neural network model can achieve eventual convergence, its drawback is the long learning and training time, which may lead to convergence to a local optimum. To address this issue, a genetic algorithm (GA) is incorporated into the BP neural network model; hence, the water demand prediction model can also be called a GA-BP neural network model. Since the genetic algorithm (GA) is an optimization method used to find exact or approximate solutions to optimization problems, it is applied in many research fields, based on the mechanism of natural selection and Darwin's main principle: survival of the fittest. The genetic algorithm (GA) is efficient, practical, and robust. It starts with an initial random solution set called a population, where each individual is called a chromosome. By encoding data into chromosomes, the fitness of each chromosome is evaluated according to the objective function. Through a closed-loop process of "initialization-evaluation-selection-crossover-mutation-iteration," individuals with higher fitness are gradually selected, while those with lower fitness are excluded, resulting in a new population. This process is repeated until the chromosomes meet the preset criteria.

[0066] In this embodiment, the genetic algorithm (GA) uses the root mean square error (RMSE) to update the fitness. The parameters of the GA are adaptively adjusted iteratively. The mutation rate is set to 0.12 in the first half of the iteration to expand the search range, and then automatically reduced to 0.08 in the second half to stabilize the optimal solution. The population size is set to 10. If the root mean square error (RMSE) drops to less than 0.01 for five consecutive generations, the population size is automatically increased from 10 to 12 temporarily to avoid getting trapped in local optima. The crossover rate is set to 0.5, the elite retention rate is set to 0.1, and the maximum number of iterations is set to 30.

[0067] The entire training process begins with a genetic algorithm (GA). The GA performs a global search on the weight range of the BP neural network model and finds the optimal initial weights and thresholds. Then, the BP neural network model begins the training process with the optimal initial weights and thresholds provided by the GA and approaches the optimal solution. By optimizing the weights and thresholds of the BP neural network model, it avoids getting stuck in local optima, thereby effectively improving the training speed. At the same time, the GA-BP neural network model outperforms the BP neural network model in terms of prediction accuracy.

[0068] The fitness of each individual in the new generation of weighted and thresholded populations is calculated. Iteration stops if the fitness threshold is met (error below the threshold) or the number of generations reaches its maximum. After iteration terminates, the individual with the highest fitness in the new generation of weighted and thresholded populations becomes the optimal initial weights and thresholds for the BP neural network model. The optimal weights and thresholds output by the genetic algorithm (GA) are substituted into the BP neural network model to replace the random initial parameters of the traditional BP neural network model. The training and validation sets are then input into the optimized BP neural network model for training. During the training phase, the learning rate and learning target error of the BP neural network model are set to 0.01 and 0.001, respectively. Mean squared error (MSE) and root mean square error (RMSE) are used as evaluation metrics. The training results can be found in [reference needed]. Figure 3 The optimized BP neural network model is iteratively trained using the training set, and the loss function value of the model is monitored in real time using the validation set to determine whether the loss function value is less than a preset threshold or whether the current iteration number has reached a preset upper limit. If either condition is met, training is stopped, and the trained water demand prediction model is obtained.

[0069] Subsequently, an exponential smoothing forecasting algorithm based on fluctuation coefficient correction was used to predict the upper and lower limits of the input indicators for the annual water demand forecasting model. The exponential smoothing forecasting algorithm is a commonly used time series forecasting method that predicts future values ​​by assigning progressively decreasing weights to historical data, emphasizing the greater influence of recent data and the gradually weakening influence of long-term data. The specific steps of the exponential smoothing forecasting algorithm are: determining the smoothing coefficient... The initial smoothing value is determined based on historical time series data. Its core formula is: The process involves iteratively calculating smoothed values ​​and adjusting the smoothing coefficient based on the iteration results and mean squared error (MSE) for prediction. Then, the fluctuation coefficient correction method is used to convert point predictions into interval predictions. This involves first calculating the historical coefficient of variation (CV) for each historical indicator value, where CV = standard deviation / mean. Finally, the historical CV is combined with the original exponential smoothing prediction method to convert the indicator values... Revised to upper and lower limit range values ,in, Finally, the upper and lower limit intervals of the predicted historical index values ​​for the planning year are input into the trained water demand prediction model to obtain the water demand interval input values ​​for the multi-objective interval stochastic chance constrained planning model within the planning year. The error parameters and prediction results of the water demand prediction model are shown in Table 1.

[0070] Table 1: Evaluation Parameters and Prediction Results of Water Demand Prediction Model

[0071]

[0072] Step 2: Collect and preprocess historical water availability data and historical energy availability data for the water resource optimization allocation area; construct a water-energy system coupling risk analysis model based on the Copula function to achieve quantitative analysis of the coupling risk of the water-energy system; combine the preset risk threshold and the quantitative analysis results to classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high, clarify the current risk level of the water-energy system, and map the risk parameters corresponding to the risk level to the default probability of the opportunity constraint;

[0073] In this embodiment, the water-energy system coupled risk analysis model is used to quantify the joint risk level of water and energy for different water-using sectors under various scenarios, providing risk constraints for subsequent optimal allocation. The specific construction and quantitative analysis process of the water-energy system coupled risk analysis model is as follows:

[0074] Step 201: Collect historical water availability data and historical energy availability data within the water resource optimization allocation area, and perform outlier removal and standardization processing in sequence to obtain water availability dataset and energy availability dataset;

[0075] Step 202: For the preprocessed water resource availability dataset and energy availability dataset, the marginal distributions are fitted using normal distribution, log-normal distribution, and Weibull distribution, respectively. The root mean square error is used as the evaluation index for the fitting effect. The root mean square error corresponding to each probability distribution model is calculated, and the probability distribution model with the smallest root mean square error is selected as the optimal marginal distribution for water resource availability and energy availability, respectively. The specific results are shown in Table 2.

[0076] Table 2: Marginal Distribution Parameters and Assessment Parameters for Water and Energy Availability

[0077]

[0078] Step 203: Based on the optimal marginal distribution of water and energy availability, four Copula functions—Clayton, Frank, Gumbel, and Student-t—are used to construct a joint probability distribution model of water and energy availability. The Copula function with the lowest root mean square error (RMSE) and the lowest Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) is selected as the core mathematical tool to characterize the coupling relationship between water and energy, and a water-energy system coupling risk analysis model is constructed. The fitting parameters are shown in Table 3, and the fitting results are shown in [Table 3]. Figure 4 .

[0079] Table 3: Copula Fitting Parameters Table

[0080]

[0081] Step 204: Pre-set the risk threshold for water-energy composite risk. Based on the joint probability of water resource and energy availability calculated by the water-energy system coupling risk analysis model, and combined with the optimal marginal distribution of the water resource subsystem and energy subsystem, determine feasible combinations of water resource subsystem risk and energy subsystem risk through a probability inverse method. Subsequently, combining the preset risk threshold with the feasible combinations of water resource subsystem risk and energy subsystem risk, classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high. Clarify the current risk level of the water-energy system, and map the core risk parameters corresponding to the risk level to the default probability of the opportunity constraint. The specific implementation process is as follows:

[0082] A risk threshold for water-energy composite risks is pre-set. This risk threshold needs to be determined in combination with the actual carrying capacity of regional water resources and energy supply, the frequency of historical risk occurrences, and regional development planning goals, and is used to define the critical state of risk occurrence.

[0083] Based on the joint probability of water and energy availability calculated by the optimal water-energy system coupling risk analysis model, and combined with the optimal marginal probability distribution of each water and energy subsystem, the feasible combination of subsystem risks is solved by the probability inverse method. That is, the corresponding feasible range of energy supply risk under different water availability risk levels, and the range of water availability risk under different energy supply risk levels are determined, so as to ensure the rationality and actual adaptability of the risk combination.

[0084] The core risk parameters obtained by quantification (including subsystem risk values, joint risk probabilities, etc.) are accurately mapped to the default probabilities of opportunity constraints. By setting different default probability ranges, four differentiated water-energy coupling risk scenarios of high, medium-high, medium-low, and low are divided and set. Each water-energy coupling risk scenario clearly corresponds to different degrees of water resource and energy supply and demand tension, providing multi-dimensional and quantifiable risk scenario input support for subsequent water demand prediction models, and ensuring the applicability and prediction accuracy of the prediction model under different risk scenarios.

[0085] In this embodiment, the low-risk threshold is first set to 0.02 and the high-risk threshold is set to 0.04. Then, based on the joint probability and subsystem probability distribution constructed by the Copula function, feasible combinations of subsystem risks are derived. Finally, the risk parameters are mapped to the default probability of the opportunity constraint. The parameter combinations (p, p1, p2) that characterize the level of coupled risk are as follows: low-coupling risk scenario S1 (0.02, 0.06, 0.044), medium-low coupling risk scenario S2 (0.02, 0.1, 0.032), medium-high coupling risk scenario S3 (0.04, 0.1, 0.09), and high-coupling risk scenario S4 (0.04, 0.12, 0.06).

[0086] Step 3: Based on the water demand range prediction results of each water user sector and the quantitative analysis results output by the water-energy system coupled risk analysis model, with the goal of maximizing the comprehensive benefits of the system as the core optimization objective, and incorporating the system joint risk constraints, water supply and demand balance constraints, total water supply balance constraints, energy availability constraints and carbon emission control constraints as constraints, a multi-objective interval stochastic chance constrained programming (CISP) model based on Copula is constructed.

[0087] The multi-objective interval stochastic chance constrained programming (CISP) model based on Copula can be formulated as follows:

[0088] ;

[0089] Constrained by:

[0090] ;

[0091] ;

[0092] ;

[0093] ;

[0094] in, Let be the objective function. and For interval parameters, It is a constrained threshold parameter that combines interval uncertainty and random characteristics. For a predetermined optimal Copula function, To constrain the probability of violation; N=1,2,…,n represents different sub-objectives (such as economic, energy, and environmental objectives), and i=1,2,…,k represents the number of the random constraint;

[0095] The above-mentioned formulaic model is an abstract and concise framework for the optimization problem of water-energy systems. The multi-objective interval stochastic chance constrained programming model, which is elaborated in the subsequent specification, is the practical implementation of this abstract framework in a specific technical scenario.

[0096] In this embodiment, a multi-objective optimization framework is established, incorporating energy consumption parameters during the water system operation process. This framework constructs a water-energy dual-dimensional optimization objective, comprehensively considering multiple objective functions such as water use revenue, water intake costs, wastewater treatment costs, energy consumption, and carbon emissions. A dynamic weight adjustment mechanism is introduced, defining a normalized dynamic weight vector. ,in, For economic benefit weighting, For energy consumption control weight, As the carbon emission constraint weight, and satisfying The normalized constraints enable precise linkage between risk scenarios and target priorities. Water-energy coupled risk scenarios (low, low-medium, medium-high, and high risk) serve as the core triggering conditions. In the low-risk scenario, economic benefits are emphasized, with weights set as follows: In low-to-medium risk scenarios, the focus is on balancing economic and security considerations, with weights set as follows: Under medium- to high-risk scenarios, the weighting of energy consumption control is increased while also taking carbon emissions into account, with the weighting set as ( In high-risk scenarios, priority is given to energy consumption and carbon emission control, with weights set as follows: ).

[0097] This paper constructs a multi-objective interval stochastic chance-constrained programming model based on Copula analysis, which considers economic benefits, energy consumption control, and carbon emission limits. It uses interval and stochastic parameters to express the uncertainties of water supply, demand, and related objectives and constraints. A Copula-based chance-constrained programming model is introduced to control system feasibility risks. Specifically, based on the system analysis of the multi-objective collaborative decision-making logic of the water resource management system, covering the energy consumption characteristics of the entire chain from water intake to water production, distribution, and wastewater treatment, and clarifying key uncertainty parameters such as water demand, water availability, and energy availability, a Copula-based multi-objective interval stochastic chance-constrained programming model is constructed, taking into account economic benefits, energy consumption control, and carbon emission constraints.

[0098] Objective function (maximizing overall system benefits):

[0099] ;

[0100] (1-1) Revenue from water use by the department:

[0101] ;

[0102] (1-2) Cost of water intake from water source

[0103] ;

[0104] (1-3) Wastewater treatment costs

[0105] ;

[0106] (1-4) Energy consumption of the entire water-energy coupling system:

[0107] ;

[0108] (1-5) Overall carbon emissions during system operation:

[0109] ;

[0110] The constraints are:

[0111] Joint System Risks:

[0112] ;

[0113] Water resource balance constraints (opportunity constraints):

[0114] ;

[0115] ;

[0116] ;

[0117] ;

[0118] Water resource supply and demand constraints:

[0119] ;

[0120] Energy availability constraints (opportunity constraints):

[0121] ;

[0122] Carbon emission constraints:

[0123] ;

[0124] In the formula:

[0125] j represents the final water user sector, where j = 1, 2, 3, 4 represent the agricultural, industrial, residential, and ecological sectors, respectively.

[0126] s represents the risk scenario, where s=1,2,3,4 represent scenarios S1, S2, S3, and S4, respectively.

[0127] t represents the planning period, which is generally five years.

[0128] Revenue for water users (unit: 10) 6 RMB / 10 6 m 3 );

[0129] Cost of water extraction (unit: 10) 6 RMB / 10 6 m 3 );

[0130] Wastewater rate;

[0131] Cost of wastewater treatment (unit: 10) 6 RMB / 10 6 m 3 );

[0132] Water intake from water source (unit: 10) 6 m 3 );

[0133] Water distribution from water source to water user (unit: 10) 6 m 3 )

[0134] Energy consumption for water extraction (unit: 10) 6 kWh / 10 6 m 3 );

[0135] Energy consumption for water production (unit: 10) 6 kWh / 10 6 m 3 );

[0136] Energy consumption for water distribution (unit: 10) 6 kWh / 10 6 m 3 );

[0137] Energy consumption for wastewater treatment (10 6 kWh / 10 6 m 3 );

[0138] The proportion of new energy types;

[0139] Traditional energy prices (unit: 10) 6 RMB / 10 6 m 3 );

[0140] Price of new energy (unit: 10) 6 RMB / 10 6 m 3 );

[0141] Carbon emission factor (unit: 10) 6 kgCO2 / 10 6 m 3 );

[0142] Carbon emission price for traditional energy-driven systems (in units of 10) 6 RMB / 10 6 kgCO2);

[0143] Carbon emission price of new energy drive systems (unit: 10) 6 RMB / 10 6 kgCO2);

[0144] Water resource availability (unit: 10) 6 m 3 );

[0145] Water loss during the water intake process;

[0146] Losses during the water production process;

[0147] This is due to losses during the water distribution process;

[0148] Departmental water demand (unit: 10) 6 m 3 );

[0149] Energy availability (in units of 10) 6 kWh);

[0150] Maximum carbon emissions (unit: 10) 6 kgCO2);

[0151] Water allocation for the department (unit: 10)6 m 3 );

[0152] Water intake from water source (unit: 10) 6 m 3 ).

[0153] Step 4: Use an interactive algorithm to decompose the multi-objective interval stochastic chance constrained programming model into a deterministic sub-model representing the most optimistic scenario and a deterministic sub-model representing the most conservative scenario;

[0154] In this embodiment, the predicted water demand interval parameters of the departments and the interval random parameters of the objective function and constraints of the collected regional coverage model are substituted into the multi-objective interval stochastic chance constrained programming model to solve for the final water resource optimization allocation result.

[0155] The CISP model is decomposed into deterministic upper and lower bound sub-models using an interactive algorithm. This interactive algorithm is a solution framework for linear programming problems with interval parameters. Its core objective is to transform the interval optimization model, which cannot be directly solved, into two sub-models—an "upper bound sub-model" and a "lower bound sub-model"—that can be solved using conventional linear programming methods. Ultimately, this yields the interval solution for the objective function and decision variables. Specifically, this is achieved by adjusting the coefficients of the objective function... and constraint function coefficients and Adjustments are made; the goal of the upper bound sub-model is to find the maximum possible value of the objective function, which needs to be achieved by maximizing the contribution of the objective function coefficients and the relaxed constraints; while the goal of the lower bound sub-model is to find the minimum possible value of the objective function, which needs to be achieved by minimizing the contribution of the objective function coefficients and the strict constraints. The specific upper and lower bound sub-models are as follows:

[0156] Upper bound sub-model (corresponding to the most optimistic scenario):

[0157] ;

[0158] Constrained by:

[0159] ;

[0160] ;

[0161] ;

[0162] Lower bound sub-model (corresponding to the most conservative scenario):

[0163] ;

[0164] Constrained by:

[0165] ;

[0166] ;

[0167] ;

[0168] ;

[0169] ;

[0170] ;

[0171] Step 5: Use a hybrid intelligent algorithm to solve the two deterministic sub-models respectively. Based on the solution results, integrate them to obtain the regional water resource optimization allocation results that are suitable for four types of coupled risk scenarios: low, medium-low, medium-high, and high.

[0172] In this embodiment, a hybrid intelligent algorithm is used to solve the two deterministic sub-models. The hybrid intelligent algorithm is a fusion of particle swarm optimization (PSO) and the NSGA-II algorithm. The solution process is as follows: using the weighted objective function of the deterministic sub-model as the fitness function, a global pre-search of the solution space is performed using PSO to obtain an initial solution set. This initial solution set is then merged with the random initial solutions generated by the NSGA-II algorithm to form a fused initial population. Non-dominated sorting, crowding calculation, selection, crossover, and mutation operations are sequentially performed on this initial population. After iteration, the final solution sets of the two deterministic sub-models are obtained. Based on the final solution sets of the two deterministic sub-models, and combining the default probability intervals corresponding to the four types of coupled risk scenarios (low, low-medium, medium-high, and high), interpolation methods are used to generate regional water resource optimization allocation results that are adapted to the four types of coupled risk scenarios.

[0173] The solution steps of the intelligent hybrid algorithm are as follows:

[0174] First, the PSO (Particle Swarm Optimization) algorithm is started to perform a global pre-search on a single deterministic sub-model. The algorithm parameters are set as follows: the particle swarm size is 100 particles, the inertia weight adopts a dynamic linear decreasing strategy (initial value 0.9, reduced to 0.4 after 50 iterations), the individual learning factor is 2.0, the social learning factor is 2.0, the velocity boundary is set to [-0.2X_max, 0.2X_max] (where X_max is the upper limit of the decision variable), the position boundary is consistent with the constraint interval of the sub-model's decision variables, the maximum number of iterations is set to 50 generations, and the fitness function adopts the weighted objective function under each risk level of the sub-model. After each iteration, the individual fitness value is calculated, and the optimal fitness of the population and the corresponding decision variable are recorded. After the iteration terminates, the top 50% of the individuals in terms of fitness are selected from the population to form a PSO solution set containing 50 relatively good initial solutions.

[0175] Subsequently, based on the constraints of the decision variables of the sub-model, 50 valid initial solutions are randomly generated to form a random solution set of the Non-Dominated Sorting Genetic Algorithm (NSGA-II); the above 50 better initial solutions of PSO are merged with 50 random initial solutions of NSGA-II to form a final initial population of NSGA-II with a size of 100 individuals.

[0176] Finally, the NSGA-II algorithm is used to solve the upper and lower bound deterministic sub-models respectively. The algorithm parameters are set as follows: maximum number of iterations 200, crossover rate 0.8, mutation rate 0.05, and Euclidean distance is used to calculate the crowding degree. When the maximum number of iterations 200 is reached, or the crowding degree change rate of the solution set is <1% for 20 consecutive generations, the algorithm terminates and outputs the Pareto optimal solution sets corresponding to the upper bound sub-model and the lower bound sub-model respectively.

[0177] Finally, by integrating the Pareto optimal solution sets of the upper and lower bound sub-models, a diversified water resource allocation scheme set is output that simultaneously satisfies energy availability constraints, carbon emission control requirements, and tolerance for different coupled risk scenarios. This scheme set can provide solid decision-making support for the scientific scheduling and refined planning of regional water resources in the future. For detailed allocation results, please refer to [link to relevant documentation]. Figure 5 and Figure 6 .

[0178] Ultimately, the water-energy coupled water resource planning method of this invention, in terms of data support, utilizes a GA-BP neural network model with iteratively adaptive adjustment of algorithm parameters combined with an exponential smoothing method for fluctuation coefficient correction to conduct interval predictions, thereby improving the robustness and scientific rigor of core indicator predictions and providing a precise and reliable data foundation for water resource allocation. Regarding uncertainty and risk handling, a multi-objective interval stochastic chance-constrained programming model integrating Copula analysis, interval stochastic programming, and chance-constrained programming can simultaneously handle multiple uncertainties and accurately quantify them, addressing the risks of water-energy system coupling and enhancing adaptability to complex scenarios. In terms of resource allocation, multiple risk scenarios are designed from the perspective of water-energy related risk linkage, and a correlation mechanism is established between risk scenarios and the dynamic weights of the objective function, achieving risk-oriented water resource optimization allocation and making the allocation results more aligned with the actual needs of different risk levels. Regarding green development, the model incorporates energy consumption intensity and carbon emissions throughout the entire water production chain, effectively identifying key nodes with high energy consumption and high carbon emissions, pointing the way for energy-saving and carbon-reduction optimization in water resource planning, and contributing to the sustainable development of water resource allocation.

[0179] Example 2

[0180] The water-energy coupled water resource planning device of the present invention includes:

[0181] The water demand range prediction module is used to collect historical index values ​​corresponding to each water-using sector within the water resource optimization allocation area, process the historical index values ​​using the exponential smoothing prediction method with fluctuation coefficient correction, and obtain the upper and lower limit ranges of each index in the planning year; and input the upper and lower limit ranges of each index in the planning year into the water demand prediction model, and predict the water demand of each water-using sector through the water demand prediction model, and output the water demand range prediction results of each water-using sector.

[0182] The coupling risk scenario quantification module is used to collect historical water resource availability data and historical energy availability data of the water resource optimization allocation area, and preprocess the data; construct a water-energy system coupling risk analysis model based on the Copula function to realize quantitative analysis of the coupling risk of the water-energy system; combine the preset risk threshold and the quantitative analysis results to classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high, clarify the current risk level of the water-energy system, and map the risk parameters corresponding to the risk level to the default probability of the opportunity constraint;

[0183] The optimization model construction module is used to construct a multi-objective interval stochastic chance constrained programming model based on the water demand interval prediction results of each water-using sector and the quantitative analysis results output by the coupled risk scenario quantification module. The core optimization objective is to maximize the comprehensive benefits of the system. At the same time, the system joint risk constraints, water supply and demand balance constraints, total water resource balance constraints, energy availability constraints and carbon emission control constraints are incorporated as constraints.

[0184] The model decomposition module is used to decompose the multi-objective interval stochastic chance constrained programming model into a sub-model representing the most optimistic scenario and a sub-model representing the most conservative scenario using an interactive algorithm.

[0185] The model solving and result integration module is used to solve the two sub-models separately using a hybrid intelligent algorithm. Based on the solution results, it integrates and outputs the regional water resource optimization allocation results adapted to four types of coupled risk scenarios: low, medium-low, medium-high, and high.

[0186] Example 3

[0187] The electronic device of the present invention includes a central processing unit and a memory, wherein the central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the water-energy coupled water resource planning method as in Embodiment 1.

[0188] Example 4

[0189] The computer-readable storage medium of the present invention stores, in the form of computer-readable instructions, a computer program implemented according to the water-energy coupling water resource planning method in Embodiment 1, which, when called by a computer, executes the steps included in the corresponding method.

[0190] The above are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above content. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A water-energy coupled water resource planning method, characterized in that, Includes the following steps: Step 1: Collect historical index values ​​for water demand forecasting for each water-using sector within the water resource optimization allocation area. Process the historical index values ​​using the exponential smoothing forecasting method with fluctuation coefficient correction to obtain the upper and lower limit ranges of each index for the planning year. Input the upper and lower limit ranges of each index for the planning year into the water demand forecasting model. Use the water demand forecasting model to predict the water demand of each water-using sector to obtain the water demand range forecast results for each water-using sector. Step 2: Collect and preprocess historical water availability data and historical energy availability data for the water resource optimization allocation area; construct a water-energy system coupling risk analysis model based on the Copula function to achieve quantitative analysis of the coupling risk of the water-energy system; combine the preset risk threshold and the quantitative analysis results to classify the coupling risk of the water-energy system into four categories of coupling risk scenarios: low, low-medium, medium-high, and high, clarify the current risk level of the water-energy system, and map the risk parameters corresponding to the risk level to the default probability of the opportunity constraint; Step 3: Based on the water demand range prediction results of each water-using sector and the quantitative analysis results output by the water-energy system coupled risk analysis model, with maximizing the comprehensive benefits of the system as the core optimization objective, and incorporating the joint risk constraints of the system, water supply and demand balance constraints, total water resource balance constraints, energy availability constraints and carbon emission control constraints as constraints, a multi-objective interval stochastic chance constrained programming model is constructed. Step 4: Use an interactive algorithm to decompose the multi-objective interval stochastic chance constrained programming model into a deterministic sub-model representing the most optimistic scenario and a deterministic sub-model representing the most conservative scenario; Step 5: Use a hybrid intelligent algorithm to solve the two deterministic sub-models respectively. Based on the solution results, integrate them to obtain the regional water resource optimization allocation results that are suitable for four types of coupled risk scenarios: low, medium-low, medium-high, and high.

2. The water-energy coupled water resource planning method according to claim 1, characterized in that, In step 1, the water-using sectors in the water resource optimization allocation area include the agricultural sector, the industrial sector, the residential sector, and the ecological environment sector. The historical indicators for the agricultural sector include the output value of the primary industry, water consumption per 10,000 yuan of agricultural output value, irrigated arable land area, and irrigation energy consumption. The historical indicators for the industrial sector include the output value of the secondary industry, water consumption per 10,000 yuan of industrial output value, the number of large-scale industrial enterprises, and industrial production energy consumption. The historical indicators for the residential sector include population density, per capita residential water consumption, GDP, and residential water energy consumption. The historical indicators for the ecological environment sector include urban green space area, rainfall, changes in groundwater reserves, and renewable energy utilization rate.

3. The water-energy coupled water resource planning method according to claim 2, characterized in that, In step 1, the specific steps for processing the historical indicator values ​​using the exponential smoothing forecasting method with volatility coefficient correction are as follows: Determine the smoothing coefficient, determine the initial smoothing value based on historical index values, iteratively calculate the smoothing value, and adjust the smoothing coefficient according to the iteration results and mean square error to complete the point prediction of the index value for the planning year. The point prediction is converted into an interval prediction by using the fluctuation coefficient correction method. The historical coefficient of variation of each historical indicator value is calculated. The historical coefficient of variation is the ratio of the standard deviation to the mean. The index values ​​obtained by the exponential smoothing prediction method are then corrected to the upper and lower limit interval values ​​of each indicator in the planning year by combining the historical coefficient of variation.

4. The water-energy coupled water resource planning method according to claim 3, characterized in that, In step 1, the construction and training steps of the water demand prediction model are as follows: Step 101: Collect historical indicator values ​​of the agricultural, industrial, residential, and ecological environmental sectors within the water resource optimization allocation area; standardize the collected historical indicator values, imputate missing values, and remove outliers to obtain a standardized dataset; divide the standardized dataset into a training set and a validation set according to a preset ratio. Step 102: Construct a BP neural network model, wherein the input layer of the BP neural network model corresponds to the historical index values ​​of each water-using sector, and the output layer corresponds to the water demand of each water-using sector; Step 103: Using the root mean square error as the evaluation criterion for the fitness function, a genetic algorithm is used to globally optimize the initial weights and thresholds of the BP neural network model, output the optimal initial weights and thresholds, and replace the random initial parameters in the BP neural network model with the optimal initial weights and thresholds. Step 104: Iteratively train the optimized BP neural network model based on the training set, and use the validation set to monitor the loss function value of the model in real time to determine whether the loss function value is less than the preset threshold or whether the current iteration number has reached the preset upper limit. If any of the judgment conditions are met, training will stop, and the water demand prediction model will be obtained after training.

5. The water-energy coupled water resource planning method according to claim 4, characterized in that, In step 2, the construction and quantitative analysis process of the water-energy system coupling risk analysis model is as follows: Step 201: Collect historical water availability data and historical energy availability data within the water resource optimization allocation area, and perform outlier removal and standardization processing in sequence to obtain water availability dataset and energy availability dataset; Step 202: For the preprocessed water resource availability dataset and energy availability dataset, the marginal distribution is fitted using normal distribution, log-normal distribution and Weibull distribution respectively. The root mean square error is used as the evaluation index of the fitting effect. The root mean square error corresponding to each probability distribution model is calculated. The probability distribution model with the smallest root mean square error is selected as the optimal marginal distribution for water resource availability and energy availability respectively. Step 203: Based on the optimal marginal distribution of water resource availability and energy availability, four Copula functions, namely Clayton, Frank, Gumbel, and Student-t, are used to construct a joint probability distribution model of water resource availability and energy availability. The Copula function with the smallest root mean square error and mean square error, and the lowest Akaike information criterion and Bayesian information criterion, is selected as the core mathematical tool to characterize the coupling relationship between water resources and energy, so as to construct a water-energy system coupling risk analysis model. Step 204: Pre-set the risk threshold for water-energy composite risk. Based on the water-energy system coupled risk analysis model, calculate the joint probability of water resource and energy availability. Combine the optimal marginal distribution of the water resource subsystem and the energy subsystem. Determine the feasible combination of water resource subsystem risk and energy subsystem risk through probability inverse reasoning. Then, combine the preset risk threshold with the feasible combination of water resource subsystem and energy subsystem risk to classify the water-energy system coupled risk into four categories: low, low-medium, medium-high, and high. Clarify the current risk level of the water-energy system. Map the core risk parameters corresponding to the risk level to the default probability of the opportunity constraint.

6. The water-energy coupled water resource planning method according to claim 5, characterized in that, In step 3, the objective function in the multi-objective interval stochastic chance constrained programming model includes a first sub-objective, a second sub-objective, and a third sub-objective, wherein, The first sub-objective is to maximize the profits generated by water use in each water-using sector; The second sub-objective is to minimize the total energy consumption of the entire water-energy coupling system. The third sub-objective is to minimize the overall carbon emissions of the system operation; The first, second, and third sub-objectives are assigned corresponding weight coefficients, and the weight coefficients of the first, second, and third sub-objectives are dynamically adjusted according to different water-energy coupling risk scenarios. The sum of the products of the first, second, and third sub-objectives and their corresponding weight coefficients is taken as the comprehensive benefit of the system.

7. The water-energy coupled water resource planning method according to claim 3, characterized in that, In step 5, the solution steps of the intelligent hybrid algorithm are as follows: Using the weighted objective function corresponding to each deterministic sub-model as the fitness function, a global pre-search is performed on a single deterministic sub-model using the particle swarm optimization algorithm to obtain the initial optimal solution set of each deterministic sub-model. The initial optimal solution set is then merged with the random initial solution generated by the NSGA-II algorithm to form the initial population of the NSGA-II algorithm. The initial population is then subjected to non-dominated sorting, crowding calculation, selection, crossover, and mutation operations in sequence. After the iteration terminates, the solution of each deterministic sub-model is completed, and the final solution set of the corresponding deterministic sub-model is obtained.

8. A water-energy coupled water resource planning device, characterized in that, include: The water demand range prediction module is used to collect historical index values ​​corresponding to each water-using sector within the water resource optimization allocation area, process the historical index values ​​using the exponential smoothing prediction method with fluctuation coefficient correction, and obtain the upper and lower limit ranges of each index in the planning year; and input the upper and lower limit ranges of each index in the planning year into the water demand prediction model, and predict the water demand of each water-using sector through the water demand prediction model, and output the water demand range prediction results of each water-using sector. The coupling risk scenario quantification module is used to collect historical water resource availability data and historical energy availability data in the water resource optimization allocation area, and preprocess the data; construct a water-energy system coupling risk analysis model based on the Copula function to realize the quantitative analysis of water-energy system coupling risk; Combining the preset risk threshold with the quantitative analysis results, the coupling risk of the water-energy system is divided into four categories of coupling risk scenarios: low, low-medium, medium-high, and high. The current risk level of the water-energy system is clarified, and the risk parameters corresponding to the risk level are mapped to the default probability of the opportunity constraint. The optimization model construction module is used to construct a multi-objective interval stochastic chance constrained programming model based on the water demand interval prediction results of each water-using sector and the quantitative analysis results output by the coupled risk scenario quantification module. The core optimization objective is to maximize the comprehensive benefits of the system. At the same time, the system joint risk constraints, water supply and demand balance constraints, total water resource balance constraints, energy availability constraints and carbon emission control constraints are incorporated as constraints. The model decomposition module is used to decompose the multi-objective interval stochastic chance constrained programming model into a sub-model representing the most optimistic scenario and a sub-model representing the most conservative scenario using an interactive algorithm. The model solving and result integration module is used to solve the two sub-models separately using a hybrid intelligent algorithm. Based on the solution results, it integrates and outputs the regional water resource optimization allocation results adapted to four types of coupled risk scenarios: low, medium-low, medium-high, and high.

9. An electronic device comprising a central processing unit and a memory, characterized in that, The central processing unit is used to invoke and run a computer program stored in the memory to perform the steps of the water-energy coupled water resources planning method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores, in the form of computer-readable instructions, a computer program implementing the water-energy coupled water resources planning method according to any one of claims 1 to 7, which, when called by a computer, executes the steps included in the corresponding method.

Images