A method of rolling water distribution in an irrigation canal system

Abstract

The present application belongs to the technical field of agricultural water resource management, and relates to a rolling water distribution method for irrigation canal systems, which comprises the following steps: 1) constructing a double-layer two-stage water distribution model; 2) calibrating the double-layer two-stage water distribution model obtained in step 1) to extract a water distribution parameter line of the double-layer two-stage water distribution model; and 3) based on the water distribution parameter line obtained in step 2), determining the water distribution amount of each branch canal in a time period in real time. The present application provides a rolling water distribution method for irrigation canal systems, which can improve the utilization efficiency of guaranteed water amount and reduce the overall water shortage risk of the system.

Description

Technical Field

[0001] This invention belongs to the field of agricultural water resources management technology, and relates to a rolling water distribution method, particularly a rolling water distribution method for irrigation canal systems. Background Technology

[0002] Improving agricultural water use efficiency is a crucial way to alleviate water shortages, and optimized canal system water allocation is a core means of improving agricultural water use efficiency. Optimized canal system water allocation refers to developing multi-level canal system operation plans based on real-time soil moisture status, reported irrigation water use plans, and forecasted water inflow information, while prioritizing crop water needs, canal safety, and ecological constraints. This optimizes the allocation of water at each level of the canal system, achieving optimal spatial and temporal distribution of agricultural water resources and ensuring a smooth canal water transport process. Implementing a suitable water allocation plan significantly reduces the risk of water shortages and water transport losses. Traditional static water allocation methods rely on fixed crop phenological characteristics to develop plans for the entire growing season, irrigation season, or a single water allocation process. This method cannot respond to short-term weather changes and dynamic fluctuations in crop water needs, exhibiting poor adaptability in the context of climate change. While some water allocation practices utilize short-term hydrological and meteorological forecasts to periodically develop and update future water allocation plans, the effectiveness of this method heavily depends on the accuracy of the forecast information. However, current hydrological and meteorological forecasts, especially long-term forecasts for the entire growing season or irrigation season, exhibit significant uncertainties in their accuracy. Therefore, against the backdrop of intensified climate change and rapid crop structure adjustment, existing canal water allocation methods are either completely unable to adapt to the dynamic changes in water demand or rely too much on the accuracy of water demand forecasts. There is a lack of a rolling water allocation decision-making method that balances the dynamic adaptability of water allocation plans and the sensitivity to forecast accuracy by comprehensively utilizing characteristic water volume information from historical water allocation processes and real-time forecast information, which restricts the dynamic balance and efficient utilization of agricultural water resources. Summary of the Invention

[0003] In order to solve the above-mentioned technical problems in the background art, the present invention provides a rolling water distribution method that can reduce the overall water shortage risk of the system and improve the utilization rate of irrigation contract water volume.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A rolling water distribution method for irrigation canal systems, characterized in that the rolling water distribution method for irrigation canal systems includes the following steps: 1) Construct a two-layer, two-stage water distribution model; 2) Calibrate the two-layer-two-stage water distribution model obtained in step 1) and extract the water distribution parameter lines of the two-layer-two-stage water distribution model; 3) Based on the water distribution parameter line obtained in step 2), the water distribution volume of each branch canal should be determined in real time during the current period.

[0005] Preferably, the two-layer-two-stage water allocation model used in step 1) of the present invention includes an upper-layer model and a lower-layer model; the upper-layer model is responsible for determining the target remaining water volume at the water source end; the lower-layer model is used to solve for the water allocation threshold and optimal water allocation volume for each water user that meets the conditions; the two stages are to divide a water allocation process into two stages, the facing period and the remaining period, to coordinate the decision-making on water allocation; the remaining period is the period from the facing period to the end of the reproductive period.

[0006] As a preferred embodiment, the upper-level model adopted in this invention is constructed by taking the target remaining water volume as the decision variable, taking the minimum sum of the water shortage indices of each branch canal as the objective function, and taking the range of values ​​of the target remaining water volume as the constraint condition. The lower-level model is constructed by using the initial water allocation threshold as the decision variable, minimizing the loss of water allocation benefits over many years as the objective function, and using the water allocation details, the range of water allocation threshold values ​​for each branch canal, and water balance as constraints.

[0007] Preferably, the specific expression of the objective function of the upper-level model used in this invention is: The constraints of the upper-level model are: in: It is the sum of the water shortage indices of each branch canal, and is the function value of the upper-level model in the two-layer-two-stage model; This represents the net water distribution volume of the main canal at time t during the year y, expressed in cubic meters (m³). 3 ; This represents the water demand of the main canal during the t-period of year y, expressed in meters. 3 ; It is the target remaining water volume at the beginning of time period t, in meters. 3 ; This refers to the total contracted water volume for the reproductive period in year y, in meters (m³). 3 ; These are the year sequence number and the total number of years in the calculation series, respectively. These are the sequence number of the time period within the year and the total number of time periods within the year, respectively. These are the branch canal numbers and the total number of branch canals.

[0008] Preferably, the specific expression of the objective function of the lower-level model used in this invention is: in: It is the sum of the water distribution efficiency losses of each branch canal, and is the function value of the lower-level model in the two-layer-two-stage model; These are the deviations between the water allocation and demand during the current period, and the deviations between the remaining contracted water volume at the end of the period and its target value. This refers to the remaining contracted water volume at the beginning of period t in year y, in meters. 3 .

[0009] The water allocation details and constraints of the lower-level model are used to calculate the water allocation volume for this period, and are calculated based on three cases: A), B), or C). A): If the available water supply for the current period meets the water demand and the remaining contracted water supply is not lower than the water allocation limit threshold, then water will be allocated as needed. B): If the available water volume for the current period meets the water demand, but the remaining contracted water volume is lower than the water allocation limit threshold, and the water demand is limited according to the ratio of the remaining contracted water volume to the water allocation limit threshold, then the water allocation is limited according to that ratio. C): If the available water volume or the remaining contracted water volume in the current period does not meet the water demand, then the smaller value between the available water volume and the remaining contracted water volume is taken and the losses are deducted. The water is then allocated to each branch canal according to the weight of the water demand in the total water demand. The specific expression for calculating the water allocation volume for the current time period according to the water allocation rules is as follows: in: This is the water distribution threshold for the main canal at the beginning of time period t, in meters. 3 ; This represents the total water loss of the canal system during period t in year y; This refers to the remaining contracted water volume at the beginning of period t in year y, in meters. 3 ; This refers to the amount of water that the main canal can divert during the period t in year y, expressed in meters. 3 ; This is the gross water distribution volume of the main canal during period t in year y, in meters. 3 ; This refers to the gross water distribution volume of the main canal corresponding to the on-demand water distribution during period t in year y, in meters.3 .

[0010] The constraint condition for the range of restricted water distribution threshold values ​​for each branch canal in the lower-level model is that the initial restricted water distribution threshold for each time period is not less than the maximum value of the water distribution volume corresponding to the on-demand water distribution for each time period in previous years, and is not greater than the maximum value of the total contracted water volume in previous years; the initial restricted water distribution threshold for the later time period is not greater than the corresponding value for the previous time period. The specific expression for the range of restricted water distribution threshold values ​​for each branch canal in the lower-level model is: The constraint condition for the water balance of the canal system in the lower-level model is that the inlet water volume of each canal equals the sum of the outlet water volume and the water loss in the canal. The specific expression for the constraint condition of the water balance of the canal system in the lower-level model is: The constraint condition for water balance in the lower-level model is that the remaining contracted water volume at the beginning of the first time period is the total contracted water volume for that year, and the remaining contracted water volume at the beginning of this time period minus the gross water allocation volume within this time period is the remaining contracted water volume at the beginning of the next time period. The specific expression for the constraint condition for water balance in the lower-level model is: The specific expressions for the channel flow capacity and water volume upper limit constraints in the lower-level model are as follows: in: These are the empirical parameters for loss calculation; , These are the lengths of the k-th branch canal and the j-th segment of the canal, respectively. This is the net water volume of section j of the main canal during period t in year y, in meters. 3 ; These are the segment numbers of the main canal, with 1 representing the first segment and J representing the last segment. These are the starting and ending numbers of the branch canal directly connected to the j-th section of the main canal; , These are the design flow rates of the main canal and the k-branch canal, respectively, in cubic meters per second (m³). 3 ; It is the length of a time period.

[0011] Preferably, the calibration in step 2) of this invention is based on historical long-term water allocation, meteorological and crop planting conditions; the water allocation parameter lines of the two-layer-two-stage water allocation model include the target remaining packaged water volume parameter line and the water allocation threshold parameter lines of each branch canal; the target remaining packaged water volume parameter line is the water volume benchmark expected to be retained in response to the risk of the remaining period in the face of the time period; the water allocation threshold parameter lines of each branch canal are the critical values ​​of water allocation restrictions in the face of the time period.

[0012] Preferably, the specific implementation of the calibration in step 2) of this invention is as follows: 2.1) Set a set of target remaining water volume that meets the constraints in the upper-level model and pass it to the lower-level model; 2.2) Based on the target remaining water volume given by the upper-level model, the lower-level model independently solves the limiting water distribution threshold and optimal water distribution volume of each branch canal using a genetic algorithm or particle swarm optimization algorithm. 2.3) The lower-level model feeds back the optimal water distribution to the upper-level model to calculate the upper-level objective function value; 2.4) The upper-level model determines whether the change in the target value meets the convergence condition (usually required to be less than a certain threshold). If the convergence condition is not met, repeat steps 2.1) to 2.4) until convergence is achieved. If the convergence condition is met, the calibration of the two-layer-two-stage water distribution model is completed. Preferably, the convergence is when the deviation between two iterations is less than a given error.

[0013] As a preferred embodiment, step 3) of the present invention is specifically implemented as follows: 3.1) The irrigation water demand for the current period shall be calculated using the field water balance method; the meteorological data such as rainfall and evaporation required for the field water balance method shall be the latest forecast values ​​from nearby meteorological stations; 3.2) Using the target remaining water volume parameter line and combining it with the water source storage capacity, determine the amount of water that can be diverted during the current period; 3.3) Based on the results obtained in step 3.1) and step 3.2), and combined with the irrigation water demand and the remaining contracted water volume, the water distribution volume of each branch canal is calculated using the water distribution threshold parameter line and water distribution details of each branch canal during the current period. 3.4) Based on the results obtained in step 3.3), perform water allocation for the current time period.

[0014] Preferably, step 3) of the present invention further includes the following after step 3.4): 3.5) Repeat steps 3.1) to 3.4) until the water distribution period ends.

[0015] Compared with the prior art, the present invention has the following significant advantages: This invention provides a rolling water allocation method for irrigation canals, comprising the following steps: 1) constructing a two-layer, two-stage water allocation model; 2) calibrating the two-layer, two-stage water allocation model obtained in step 1) and extracting the water allocation parameter lines of the two-layer, two-stage water allocation model; 3) based on the water allocation parameter lines obtained in step 2), determining the water allocation volume of each branch canal in real time for the current period. Specifically, the rolling water allocation method for irrigation canals provided by this invention has the following advantages: First, it has strong risk hedging capabilities. By introducing a dual-parameter rule of target remaining water volume parameter lines and limiting water allocation threshold parameter lines, a mechanism is clearly established to offset future severe water shortages with current moderate water shortages, achieving optimized allocation and risk balance of water resources in the time dimension. Second, it makes full use of information, breaking through the limitations of traditional water allocation relying solely on forecasts. It deeply integrates the long-term historical water allocation patterns of the irrigation area (reflected through the two-layer, two-stage model) with real-time short-term forecast information, making decision-making more robust and significantly enhancing the ability to resist forecast error interference. Third, the rules are transparent and easy to operate. The proposed rolling water distribution parameter lines have clear physical meaning and explicit triggering conditions, forming an interpretable and operable scheduling procedure that is easy for irrigation district managers to understand and apply, and facilitates promotion. Fourth, it has wide applicability. The model is executed on a rolling basis. Based on the water distribution parameter lines calibrated by the two-layer-two-stage water distribution model, each decision only requires a short-term forecast, reducing the difficulty of obtaining high-precision medium- and long-term forecasts and enhancing applicability under variable climate conditions. Clearly, this invention innovatively constructs a two-layer, two-stage water allocation model. By coupling the upper and lower layers and the current and future stages, it comprehensively utilizes historical water allocation information and hydrological and meteorological forecast information to collaboratively optimize and determine the target remaining water volume and water allocation threshold for each time period. This achieves a dynamic balance and hedging of water shortage risk between the time period and the remaining period, realizes the mining and application of long-term historical water allocation patterns, strengthens the system's adaptability to forecast uncertainties, reduces excessive reliance on single forecast information, and ultimately improves the efficiency of water volume utilization, reduces the overall water shortage risk of the system, thereby improving the robustness of the water allocation scheme and the efficiency of water resource utilization. Attached Figure Description

[0016] Figure 1 This is a flowchart of the rolling water distribution parameter line extraction method provided by the present invention; Figure 2 This is a schematic diagram of the canal system structure of the Yongji irrigation area in the Hetao Irrigation District according to an embodiment of the present invention; Figure 3 This is the parameter line of the target remaining water volume in the Yongji irrigation area obtained by calibrating the two-layer-two-stage water distribution model of this invention; Figure 4The parameters for limiting water distribution thresholds in each branch canal of the Yongji irrigation area are obtained from the calibration of the two-layer-two-stage water distribution model of this invention. Detailed Implementation

[0017] See Figure 1 The present invention provides a method for rolling water distribution in an irrigation canal system, comprising the following steps: 1) Construct a two-layer, two-stage water allocation model. The two-layer, two-stage water allocation model consists of an upper-layer model and a lower-layer model. The upper-layer model is responsible for determining the target remaining water volume at the water source end (main canal), while the lower-layer model is used to solve for the water allocation threshold and optimal water allocation volume for each water user (branch canal) that meets the conditions. The two-stage refers to dividing a water allocation process into two stages: the facing period and the remaining period (the period from the facing period to the end of the reproductive period) to coordinate water allocation decisions.

[0018] The specific construction method for the upper-level model and the lower-level model is as follows: the upper-level model is constructed with the target remaining contracted water volume as the decision variable, the objective function being the minimum sum of the water shortage indices of each branch canal, and the constraints being the range of values ​​for the target remaining contracted water volume; the lower-level model is constructed with the initial water allocation threshold as the decision variable, the objective function being the minimum loss of water allocation benefits over many years, and the constraints being the water allocation details, the range of values ​​for the water allocation thresholds of each branch canal, and water balance.

[0019] The specific expression for the objective function of the upper-level model is: The constraints of the upper-level model are that the initial target remaining contracted water volume for each time period is non-negative and does not exceed the maximum total contracted water volume for the growing season over the years, and the initial target remaining contracted water volume for the next time period does not exceed the corresponding value for the previous time period. The specific expression is: in: The net water distribution volume of the main canal during the period t in year y is expressed in m³. 3 ; The unit is the water demand of the main canal during the period t in year y, expressed in m³. 3 ; This is the sum of the water shortage indices of each branch canal; The target remaining water volume at the beginning of time period t, in meters. 3 ; The total contracted water volume for the reproductive period in year y is expressed in meters (m³). 3 ; These are the year serial numbers, used to calculate the total number of years in the series; These are the sequence numbers of time periods within the year and the total number of time periods within the year, respectively. These are the branch canal numbers and the total number of branch canals.

[0020] The objective function of the lower-level model used in this invention is to minimize the sum of the two-stage water allocation losses, that is, to minimize the sum of the deviation between the current water allocation and demand and the remaining contracted water volume (remaining contracted water volume at the end of the period) and their target value. The specific expression is: The lower-level model's water allocation details constraints are used to calculate the water allocation volume for the current period, including the following three cases: A), B), and C). A) If the available water volume for the current period meets the water demand and the remaining contracted water volume is not lower than the water allocation limit threshold, then water will be allocated as needed; B) If the available water volume for the current period meets the water demand, but the remaining contracted water volume is lower than the water allocation limit threshold, and the water demand is limited according to the ratio of the remaining contracted water volume to the water allocation limit threshold, then the water allocation is limited according to that ratio. C) If the available water volume or the remaining contracted water volume in the current period does not meet the water demand, then the smaller value between the available water volume and the remaining contracted water volume is taken, and after deducting the losses, each branch canal is allocated water according to the weight of the water demand in the total water demand.

[0021] The specific expression for calculating the water distribution volume in the lower-level model water distribution details is: The lower-level model's value range constraints are: the initial water allocation threshold for each time period is not less than the maximum water allocation volume corresponding to the on-demand water allocation for each time period in previous years, and not greater than the maximum total contracted water volume in previous years; the initial water allocation threshold for the next time period is not greater than the corresponding value for the previous time period. The specific expression is: The water balance constraint of the lower-level model is that the inlet water volume of each channel equals the sum of the outlet water volume and the channel loss water volume. The specific expression is: The water balance constraint of the lower-level model is that the remaining contracted water volume at the beginning of the first time period is the total contracted water volume for that year. The remaining contracted water volume at the beginning of this time period minus the gross water allocation volume during this time period is the remaining contracted water volume at the beginning of the next time period. The specific expression is: The specific expressions for the channel flow capacity and water volume upper limit constraints of the lower-level model are: in: The function value of the lower-level model in the two-layer-two-stage model is the sum of the water distribution benefit losses of each branch canal; These are the deviations between the water allocation and demand during the current period, and the deviations between the remaining contracted water volume and its target value at the end of the period. The remaining contracted water volume at the beginning of period t in year y, in m³. 3 ; This is the water distribution threshold for the main canal at the beginning of time period t, in meters. 3 ; The volume of water that can be diverted from the main canal during the period t in year y is expressed in meters. 3 ; The gross water distribution volume of the main canal during period t in year y is expressed in m³. 3 ; This refers to the gross water distribution volume of the main canal corresponding to the on-demand water distribution during period t in year y, in meters. 3 ; The total water loss of the canal system during period t in year y is expressed in cubic meters. 3 ; These are the segment numbers of the main canal, with 1 representing the first segment and J representing the last segment. Empirical parameters for loss calculation; , These are the lengths of the k-th branch canal and the j-th segment of the canal, respectively. The net water volume of the j-th section of the main canal during the t-th period of year y (where 1 is the first section of the main canal and J is the last section of the main canal) is expressed in m³. 3 ; The starting and ending numbers of the branch canal directly connected to the j-th section of the main canal; , The design flow rates for the main canal and branch canal (k) are respectively, in cubic meters per second (m³). 3 ; The length of a time period.

[0022] 2) Calibrate the two-layer-two-stage water allocation model and extract water allocation parameter lines. Based on historical long-term water allocation, meteorological, and crop planting data, calibrate the two-layer-two-stage water allocation model and extract two types of water allocation parameter lines: the target remaining contracted water volume parameter line and the water allocation threshold parameter line for each branch canal. The target remaining contracted water volume parameter line is the water volume benchmark expected to be retained to cope with the risk of the remaining period in the current time period, and the water allocation threshold parameter line for each branch canal is the critical value of water allocation restriction in the current time period.

[0023] In step 2), the specific implementation method of calibration is as follows: 2.1) The upper-level model sets a set of target remaining water volume and passes it to the lower-level model; 2.2) The lower-level model generates N sets of water allocation thresholds based on the target remaining water volume given by the upper-level model, and optimizes the current optimal water allocation volume by combining the water allocation rules. 2.3) The lower-level model feeds back the current optimal water distribution to the upper-level model to calculate the upper-level objective function value; 2.4) The upper-level model adjusts the target remaining water volume setting value accordingly, and repeats steps 2.1) to 2.4) until the convergence condition is met. The collaborative solution between the upper and lower levels is achieved through iterative optimization.

[0024] 3) Calculate the irrigation water demand for the current period based on meteorological forecast information. Combined with the current water storage capacity, and according to the target remaining contracted water volume parameter line, the water allocation threshold parameter lines for each branch canal, and the water allocation details, determine the water allocation volume for each branch canal in real time. For example, the specific implementation of step 3) in this invention is as follows: 3.1) The field water balance method is used to calculate the irrigation water demand for the current period. The meteorological data such as rainfall and evaporation required for the field water balance method are the latest forecast values ​​from nearby meteorological stations. 3.2) Using the target remaining water volume parameter line and combining it with the water source storage capacity, determine the amount of water that can be diverted during the current period; 3.3) Based on the results obtained in step 3.1) and step 3.2), the water allocation threshold parameter line and water allocation rules are used to make real-time rolling decisions in combination with irrigation water demand and remaining contracted water volume to determine the water allocation volume of each branch canal in the current period. 3.4) Execute water allocation for the current period, update the remaining contracted water volume, enter the next period, and repeat steps 3.1) to 3.4) until the water allocation period ends.

[0025] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0026] See Figure 1 The present invention proposes a rolling water distribution method for irrigation canals, the specific steps of which are as follows: Step 1: Construct a two-layer, two-stage water allocation model. The upper-layer model is constructed using the target remaining contracted water volume as the decision variable, minimizing the sum of water shortage indices for each branch canal as the objective function, and the range of values ​​for the target remaining contracted water volume (non-negative, decreasing, and not exceeding the total contracted water volume for the growing season) as constraints. The lower-layer model is constructed using the initial water allocation threshold as the decision variable, minimizing the sum of the deviation between the current water allocation volume and demand for the current period and the deviation between the remaining contracted water volume (remaining contracted water volume at the end of the period) and its target value as the objective function, and the water allocation details, the range of limiting water allocation thresholds for each branch canal, canal system water balance, and canal system flow capacity as constraints.

[0027] Step 2: Calibrate the parameters of the two-layer, two-stage water distribution model. Input historical long-term water distribution, meteorological, and crop planting data, and achieve collaborative solution between the upper and lower layer models through iterative optimization to obtain the optimal target remaining water volume parameter line and the water distribution threshold parameter lines for each branch canal. Specific steps include: Step 2.1 The upper-level model sets a set of target remaining water volume that meets the constraints and passes it to the lower-level model.

[0028] Step 2.2 Based on the target remaining water volume given by the upper-level model, the lower-level model independently solves for the limiting water allocation threshold and optimal water allocation volume of each branch canal using optimization algorithms such as genetic algorithm and particle swarm optimization. There are three scenarios in the water allocation details that need to be considered one by one: First, if the available water volume in the current period meets the water demand and the remaining water volume is not lower than the limiting water allocation threshold, then water is allocated according to demand; second, if the available water volume in the current period meets the water demand and the remaining water volume is lower than the limiting water allocation threshold, but meets the water demand after limiting it according to the ratio of the remaining water volume to the limiting water allocation threshold, then water is allocated according to that ratio; third, if the available water volume or the remaining water volume in the current period does not meet the water demand, then the smaller value between the available water volume and the remaining water volume is taken, and losses are deducted, and water is allocated to each branch canal according to the weight of the water demand in relation to the total water demand.

[0029] Step 2.3 The lower-level model feeds back the current optimal water distribution to the upper-level model to calculate the upper-level objective function value.

[0030] Step 2.4 The upper-level model determines whether the change in the target value meets the convergence condition (usually required to be less than a certain threshold). If the convergence condition is not met, repeat steps 2.1) to 2.4) until convergence is achieved. If the convergence condition is met, the calibration of the two-layer-two-stage water distribution model is completed.

[0031] Step 3: Real-time rolling water allocation decision-making. Based on meteorological forecast information, calculate the irrigation water demand for the current period. Combined with the current water source storage, and according to the target remaining contracted water volume parameter line, the water allocation threshold parameter line for each branch canal, and the water allocation details, determine the water allocation volume for each branch canal in real time. Specific steps include: Step 3.1 The field water balance method is used to calculate the irrigation water demand for the current period. The meteorological data such as rainfall and evaporation required for the field water balance method are the latest forecast values ​​from nearby meteorological stations.

[0032] Step 3.2 Use the target remaining water volume parameter line and the water source storage to determine the amount of water that can be diverted in the current period.

[0033] Step 3.3 Combine the irrigation water demand and the remaining contracted water volume, and use the water distribution threshold parameter line and water distribution details of each branch canal to calculate the water distribution volume of each branch canal in the current time period.

[0034] Step 3.4 Execute water allocation for the current time period, update the remaining contracted water volume, and enter the next time period. Repeat steps 3.1) to 3.4) until the water allocation period ends.

[0035] Taking the water distribution decision-making of the Yongji irrigation area in the Hetao Irrigation District as an example, this paper illustrates the effectiveness and rationality of the present invention. The Hetao Irrigation District is a large irrigation district in Inner Mongolia Autonomous Region, China. The Yongji irrigation area is an important component of the Hetao Irrigation District, with a designed irrigation area of ​​2,037,100 mu (approximately 133,867 hectares). Its location within the Hetao Irrigation District and the generalized backbone canal system structure are as follows: Figure 2 As shown, the canal system of the Yongji irrigation area is generalized into a two-level canal system, with all its water diversion capacity coming from the Hetao Main Canal, and supplying water to six branch canals: Yonglan, Yonggang, Xile, Xinhua, Zhengshao, and Dashui. Based on the irrigation area, canal length, design flow rate, crop planting structure, water diversion capacity of the main canal head, the historical allocation of water volume to the six branch canals of the Yongji irrigation area, and meteorological data such as historical rainfall, the calculation period is from April 6th to September 12th, with a calculation step of days. The planting areas of major crops such as wheat, corn, and sunflowers are all from the River Irrigation District Management Bureau, and the crop irrigation water requirements are calculated using measured meteorological data from the Linhe Station from 2001 to 2020. Based on this information, a two-layer, two-stage water distribution model for the Yongji Irrigation Basin Hetao Main Canal to distribute water to six branch canals was constructed using the aforementioned method. The target remaining contracted water volume and the limiting water distribution threshold in the model were calibrated using the multi-population particle swarm optimization method proposed by the patent applicant. Water was distributed according to the water diversion volume available for the current time period, the remaining contracted water volume, and the limiting water distribution threshold of the six branch canals, based on the water distribution details A), B), or C) in step 3.

[0036] Using historical data from 2001 to 2010 (a total of 10 years), the two-layer, two-stage water distribution model was calibrated using step 2) to obtain the water distribution parameter curve, i.e. Figure 3 The remaining amount of water in the target package (blue line) and Figure 4 The six branch canals in the middle section have water allocation thresholds (red lines). In the two graphs, the black and green lines represent the upper and lower limits of the actual remaining contracted water volume, respectively. The gray line represents the annual variation of the actual remaining contracted water volume during the growing season in the Yongji irrigation area from 2001 to 2010. Figure 3 and Figure 4 ( Figure 4 In the diagram (where a is Yonglan, b is Yonggang, c is Xile, d is Xinhua, e is Zhengshao, and f is Da Tuishui), it can be seen that the decreasing characteristics of the target remaining contracted water volume parameter line and the water distribution threshold parameter lines of the 6 branch canals in each time period are consistent with the changing pattern of the total irrigation water volume required by the irrigation area gradually decreasing over time.

[0037] Historical data from 2011 to 2020 (a total of 10 years, replacing forecast data during validation) were used for comparison and validation with traditional models (conventional water distribution models). The data were calibrated using a two-layer, two-stage water distribution model. Figure 3 , Figure 4 The water distribution parameter line, calculated using step 3), yields a multi-year average water shortage index sum of 31.44 for the main canals. Traditional conventional water distribution models make decisions based on forecast information for each period from the current time to the end of the decision period. The decision variable is the water distribution volume for each period from the current time to the end of the decision period, and the objective function is to minimize the sum of the main canal water shortage indices within the decision period. Calculations show that the multi-year average sum of the main canal water shortage indices for the conventional water distribution model is 40.81. Therefore, the water shortage index sum calculated by the method of this invention is smaller, and the water distribution effect obtained using this invention is superior to the conventional water distribution method.

Claims

1. A method for rolling water distribution in an irrigation canal system, characterized in that: The rolling water distribution method for the irrigation canal system includes the following steps: 1) Construct a two-layer, two-stage water distribution model; 2) Calibrate the two-layer-two-stage water distribution model obtained in step 1) and extract the water distribution parameter lines of the two-layer-two-stage water distribution model; 3) Based on the water distribution parameter line obtained in step 2), the water distribution volume of each branch canal should be determined in real time during the current period; The two-layer, two-stage water allocation model in step 1) includes an upper-layer model and a lower-layer model. The upper-layer model is responsible for determining the target remaining water volume at the water source end. The lower-layer model is used to solve for the water allocation threshold and optimal water volume for each water user that meets the conditions. The two stages are to divide a water allocation process into two stages: the facing period and the remaining period, to coordinate water allocation decisions. The remaining period is the period from the facing period to the end of the reproductive period. The upper-level model is constructed by taking the target remaining water volume as the decision variable, minimizing the sum of the water shortage indices of each branch canal as the objective function, and taking the range of values ​​of the target remaining water volume as the constraint condition. The lower-level model is constructed by using the initial water allocation threshold as the decision variable, minimizing the loss of water allocation benefits over many years as the objective function, and using the water allocation details, the range of water allocation threshold values ​​for each branch canal, and water balance as constraints. The specific expression for the objective function of the upper-level model is: The constraints of the upper-level model are: in: It is the sum of the water shortage indices of each branch canal, and is the function value of the upper-level model in the two-layer-two-stage model; This represents the net water distribution volume of the main canal at time t during the year y, expressed in cubic meters (m³). 3 ; This represents the water demand of the main canal during the t-period of year y, expressed in meters. 3 ; It is the target remaining water volume at the beginning of time period t, in meters. 3 ; This refers to the total contracted water volume for the reproductive period in year y, in meters (m³). 3 ; These are the year sequence number and the total number of years in the calculation series, respectively. These are the sequence number of the time period within the year and the total number of time periods within the year, respectively. These are the branch canal numbers and the total number of branch canals; The specific expression for the objective function of the lower-level model is: in: It is the sum of the water distribution efficiency losses of each branch canal, and is the function value of the lower-level model in the two-layer-two-stage model; These are the deviations between the water allocation and demand during the current period, and the deviations between the remaining contracted water volume at the end of the period and its target value. This refers to the remaining contracted water volume at the beginning of period t in year y, in meters. 3 .

2. The rolling water distribution method for irrigation canals according to claim 1, characterized in that: The constraints of the water allocation rules for the lower-level model are calculated based on three cases: A), B), or C). A): If the available water supply for the current period meets the water demand and the remaining contracted water supply is not lower than the water allocation limit threshold, then water will be allocated as needed. B): If the available water volume for the current period meets the water demand, but the remaining contracted water volume is lower than the water allocation limit threshold, and the water demand is limited according to the ratio of the remaining contracted water volume to the water allocation limit threshold, then the water allocation is limited according to that ratio. C): If the available water volume or the remaining contracted water volume in the current period does not meet the water demand, then the smaller value between the available water volume and the remaining contracted water volume is taken and the losses are deducted. The water is then allocated to each branch canal according to the weight of the water demand in the total water demand. The specific expression for calculating the water allocation volume for the current period according to the water allocation rules is: in: This is the water distribution threshold for the main canal at the beginning of time period t, in meters. 3 ; This represents the total water loss of the canal system during period t in year y; This refers to the remaining contracted water volume at the beginning of period t in year y, in meters. 3 ; This refers to the amount of water that the main canal can divert during the period t in year y, expressed in meters. 3 ; This is the gross water distribution volume of the main canal during period t in year y, in meters. 3 ; This refers to the gross water distribution volume of the main canal corresponding to the on-demand water distribution during period t in year y, in meters. 3 ; The constraint condition for the range of restricted water distribution threshold values ​​for each branch canal in the lower-level model is that the initial restricted water distribution threshold for each time period is not less than the maximum value of the water distribution volume corresponding to the on-demand water distribution for each time period in previous years, and is not greater than the maximum value of the total contracted water volume in previous years; the initial restricted water distribution threshold for the later time period is not greater than the corresponding value for the previous time period. The specific expression for the range of restricted water distribution threshold values ​​for each branch canal in the lower-level model is: The constraint condition for the water balance of the canal system in the lower-level model is that the inlet water volume of each canal equals the sum of the outlet water volume and the water loss in the canal. The specific expression for the constraint condition of the water balance of the canal system in the lower-level model is: The constraint condition for water balance in the lower-level model is that the remaining contracted water volume at the beginning of the first time period is the total contracted water volume for that year, and the remaining contracted water volume at the beginning of this time period minus the gross water allocation volume within this time period is the remaining contracted water volume at the beginning of the next time period. The specific expression for the constraint condition for water balance in the lower-level model is: The specific expressions for the channel flow capacity and water volume upper limit constraints in the lower-level model are as follows: in: These are the empirical parameters for loss calculation; , These are the lengths of the k-th branch canal and the j-th segment of the canal, respectively. This is the net water volume of section j of the main canal during period t in year y, in meters. 3 ; These are the segment numbers of the main canal, with 1 representing the first segment and J representing the last segment. These are the starting and ending numbers of the branch canal directly connected to the j-th section of the main canal; , These are the design flow rates of the main canal and the k-branch canal, respectively, in cubic meters per second (m³). 3 ; It is the length of a time period.

3. The rolling water distribution method for irrigation canals according to claim 2, characterized in that: The calibration in step 2) is based on historical long-term water allocation, meteorological and crop planting conditions; the water allocation parameter lines of the two-layer-two-stage water allocation model include the target remaining packaged water volume parameter line and the water allocation threshold parameter lines of each branch canal; the target remaining packaged water volume parameter line is the water volume benchmark expected to be retained in response to the risk of the remaining period in the current time period; the water allocation threshold parameter lines of each branch canal are the critical values ​​of water allocation restrictions in the current time period.

4. The rolling water distribution method for irrigation canals according to claim 3, characterized in that: The specific implementation method of calibration in step 2) is as follows: 2.1) Set a set of target remaining water volume that meets the constraints in the upper-level model and pass it to the lower-level model; 2.2) Based on the target remaining water volume given by the upper-level model, the lower-level model independently solves the limiting water distribution threshold and optimal water distribution volume of each branch canal using a genetic algorithm or particle swarm optimization algorithm. 2.3) The lower-level model feeds back the optimal water distribution obtained in step 2.2) to the upper-level model to calculate the upper-level objective function value; 2.4) The upper-level model determines whether the change in the objective function value meets the convergence condition. If the convergence condition is not met, repeat steps 2.1) to 2.4) until convergence is achieved. If the convergence condition is met, the calibration of the two-layer-two-stage water distribution model is completed.

5. The rolling water distribution method for irrigation canals according to claim 4, characterized in that: The specific implementation method of step 3) is as follows: 3.1) The irrigation water demand for the current period is calculated using the field water balance method; the meteorological data for rainfall and evaporation required for the field water balance method are the latest forecast values ​​from nearby meteorological stations; 3.2) Using the target remaining water volume parameter line and combining it with the water source storage capacity, determine the amount of water that can be diverted during the current period; 3.3) Based on the results obtained in step 3.1) and step 3.2), and combined with the irrigation water demand and the remaining contracted water volume, the water distribution volume of each branch canal is calculated using the water distribution threshold parameter line and water distribution details of each branch canal during the current period. 3.4) Based on the results obtained in step 3.3), perform water allocation for the current time period.

6. The rolling water distribution method for irrigation canals according to claim 5, characterized in that: Step 3) after step 3.4) further includes: 3.5) Repeat steps 3.1) to 3.4) until the water distribution period ends.

Images