Multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty

Abstract

The present invention is based on a multi-objective allocation method for water resources in large irrigation areas with forecast uncertainty, comprising the following steps: collecting domestic and industrial water consumption data and agricultural water consumption data within the irrigation area in the past five years; using a linear regression method to predict domestic and industrial water demand within the irrigation area, and using an irrigation quota method to predict the total agricultural water demand within the irrigation area; performing water balance preprocessing on multiple water sources and water-receiving areas in different areas within the irrigation area to obtain the monthly water shortage of each area in the irrigation area; using a linear regression method to predict domestic and industrial water demand that needs to be taken from large reservoirs, and using SVM deep learning to predict agricultural water demand that needs to be taken from large reservoirs; averaging the predictions of the two prediction methods to obtain the final water demand; establishing a water resource optimization model with multiple objective functions, such as maximizing the comprehensive benefit of large reservoirs, maximizing the water storage benefit at the end of the period, and minimizing the amount of abandoned water; and solving the model to obtain the optimal allocation plan for large reservoirs.

Description

Technical Field

[0001] The present invention belongs to the technical field of water resource scheduling, and in particular relates to a multi-objective water resource allocation method for a large irrigation area based on forecast uncertainty. Background Art

[0002] The optimal allocation of water resources is to rationally allocate various water sources under the principles of fairness, efficiency and sustainable utilization to meet the water needs of various water-using departments and maximize and optimize the overall benefits of water resource utilization. The uncertainty of reservoir runoff forecast adds challenges to the formulation of reservoir water resource allocation plans. To this end, the present invention conducts research on the optimal allocation of water resources from the perspective of regional-hydraulic project control, takes into account the uncertainty of runoff forecast in reservoir scheduling, and proposes a water resource optimization allocation model based on regional water balance and with reservoir optimization scheduling as the core, which provides a basis for large reservoirs to formulate regional water resource supply plan scheduling and improves the water resource utilization efficiency of large reservoir irrigation areas.

[0003] Therefore, there are still many urgent issues that need to be studied in terms of multi-objective allocation of water resources in large reservoir irrigation areas to achieve efficient utilization of water resources. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned existing background technology and provide a multi-objective optimization allocation method for water resources in large irrigation areas based on forecast uncertainty, so as to meet the needs of water resource allocation under current conditions and achieve optimal allocation of water resources in the entire region.

[0005] The water resource optimization configuration of the present invention starts from the entire irrigation area, comprehensively manages the water storage of all large, medium and small reservoirs, weirs and ponds in the irrigation area, and rationally allocates water sources to meet the water needs of various water uses such as life, industry, and agriculture in each area, thereby improving the comprehensive utilization efficiency of water resources. However, due to the issue of the ownership of management authority of small and medium-sized reservoirs and weirs in the region, it is difficult to achieve unified management. The present invention transforms the water resource optimization configuration problem into the optimization scheduling problem of large reservoirs, that is, first pre-processing the water sources and water needs of small and medium-sized reservoirs in the irrigation area through supply and demand balance analysis, predicting the water shortage in the irrigation area, and using it as the water supply of large reservoirs, and then optimizing the water resource scheduling of large reservoirs. At the same time, considering the uncertainty of the water inflow forecast of large reservoirs, a water resource optimization model with multiple objective functions such as maximizing the comprehensive benefit of the reservoir, maximizing the water storage benefit at the end of the period, and minimizing the amount of abandoned water is proposed to achieve optimal water resource configuration for the entire region.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty includes the following steps:

[0008] S1, divide the irrigation area into different areas and collect data on domestic, industrial and agricultural water consumption within the irrigation area every year for the past five years;

[0009] S2, using linear regression method to predict the domestic and industrial water demand in different areas within the irrigation area;

[0010] S3, using the irrigation quota method to predict the agricultural water demand of different areas within the irrigation area;

[0011] S4, pre-balancing the water supply and water demand of multiple water sources in different areas within the irrigation area according to the water balance pre-processing method, and obtaining the monthly water shortage of different areas in the irrigation area as the domestic and industrial water demand and agricultural water demand that need to be taken from large reservoirs;

[0012] S5, using linear regression to predict domestic and industrial water demand from large reservoirs;

[0013] S6, using SVM deep learning method to predict agricultural water demand that needs to be drawn from large reservoirs;

[0014] S7, averaging the domestic and industrial water demands that need to be taken from large reservoirs predicted by the water balance preprocessing method and the linear regression method, and averaging the agricultural water demands predicted by the water balance preprocessing method and the SVM deep learning method, to obtain the final domestic and industrial water demands that need to be taken from large reservoirs, and the agricultural water demands;

[0015] S8, considering the uncertainty of runoff forecast, establish a water resources optimization model with the maximum comprehensive benefits of large reservoirs, the maximum end-of-period water storage benefits, and the minimum amount of abandoned water as multi-objective functions;

[0016] S9, transform the multi-objective function into a single-objective function, use the genetic algorithm to solve the model, and obtain the optimal allocation plan for large reservoirs.

[0017] In the above technical solution, step S1 includes:

[0018] S11, divide the irrigation area into several areas according to water supply facilities, and collect and organize data on domestic and industrial water withdrawal;

[0019] S12, calculate the monthly and annual domestic and industrial water consumption of each area by area.

[0020] In the above technical solution, step S2 includes:

[0021] S21, based on the annual water consumption data for the past five years collected in step S1, perform a linear regression calculation on the total annual domestic and industrial water demand of each area within the irrigation area to obtain a predicted total domestic and industrial water demand of each area;

[0022] S22. Taking into account the cyclical nature of monthly domestic and industrial water consumption, the average proportion of each month's domestic and industrial water consumption in the past three years is calculated based on the monthly domestic and industrial water consumption of each area. Based on the total domestic and industrial water demand of each area calculated in step S21, the monthly domestic and industrial water demand of each area is calculated.

[0023] In the above technical solution, step S3 includes:

[0024] S31, collect irrigation quotas for each growth stage of early rice, mid-season rice, and late rice during the entire growth cycle under rainfall frequencies of 25%, 50%, and 75% based on a large number of experimental analyses conducted by the irrigation experimental station. The growth stages include the field soaking stage, the greening and transplanting stage, the early tillering stage, the late tillering stage, the jointing and booting stage, the heading and flowering stage, the milky and yellowing stage, and the yellow ripening stage;

[0025] S32, collects the main crop types in each area of ​​the irrigation district, including early rice, mid-season rice, late rice, planting area, time range corresponding to the crop growth stage, irrigation canal utilization coefficient, field water utilization coefficient, etc.;

[0026] S33, calculate the agricultural water demand D for each area in the irrigation area at each growth stage, and the calculation formula is as follows:

[0027]

[0028] Among them, A is the crop irrigation quota, S is the crop planting area, L0 is the field water utilization coefficient, and L1 is the irrigation canal utilization coefficient.

[0029] S34, calculating the monthly agricultural water demand of each area according to the corresponding time of each growth stage in different areas.

[0030] In the above technical solution, step S4 includes:

[0031] S41, collect water supply data from multiple water sources in the irrigation area for the past ten years, classify them into wet, normal and dry periods, and combine them with the water inflow characteristics of the forecast period to obtain the predicted monthly water supply;

[0032] S42, based on the monthly domestic and industrial water demand and agricultural water demand of each area calculated in step S2 and step S3, and the available water volume calculated in step S41, water balance calculation is performed according to the water supply priority of domestic and industrial water demand and industrial water demand, such as Figure 2 According to the steps shown, the domestic and industrial water shortages and agricultural water shortages in each area are obtained, and the total monthly water shortage is calculated as the domestic and industrial water demand and agricultural water demand that need to be taken from large reservoirs.

[0033] In the above technical solution, step S5 includes:

[0034] S51: Collect and compile monthly domestic and industrial water consumption data for large reservoirs over the past five years;

[0035] S52, perform linear regression calculation on the annual domestic and industrial water supply of large reservoirs to obtain the predicted domestic and industrial water demand;

[0036] S53. Based on the monthly domestic and industrial water consumption in the past three years, calculate the average proportion of each month to the annual domestic and industrial water consumption, and calculate the monthly domestic and industrial water demand.

[0037] In the above technical solution, step S6 includes:

[0038] S61: Collect and compile data on irrigation area in irrigation districts over the past ten years, and forecast annual monthly irrigation area, monthly agricultural water consumption in irrigation districts, and monthly agricultural water supply from large reservoirs;

[0039] S62, taking the monthly irrigation area and monthly agricultural water consumption as influencing factors, uses the data of the past ten years to train the SVM deep learning model for each month, and calculates the monthly agricultural water demand that needs to be taken from large reservoirs in the predicted year.

[0040] In the above technical solution, step S7 includes:

[0041] S71, calculating the domestic and industrial water demand of large reservoirs: averaging the monthly domestic and industrial water demand calculated in steps S4 and S5;

[0042] S72, calculate the agricultural water demand of large reservoirs: calculate the proportion of agricultural water consumption of large reservoirs to the total agricultural water demand each month according to step S4, calculate the proportion of agricultural water consumption of large reservoirs drawn from water to the total agricultural water demand each month according to step S6, and average the above two predicted proportions each month.

[0043] In the above technical solution, step S8 includes:

[0044] S81. Collect basic reservoir data, including domestic water prices, industrial water prices, agricultural water prices, monthly average electricity prices, water data from the past 30 years, water level and storage capacity relationship curves, tailwater level and discharge relationship curves, evaporation and leakage coefficients, and average hydropower station output coefficients;

[0045] S82: Establish the multi-objective function of the optimization scheduling model, and establish a water resources optimization model with the multi-objective functions of maximizing the comprehensive benefits of large reservoirs, maximizing the end-of-period water storage benefits, and minimizing the amount of abandoned water;

[0046] S83, establish the decision variables of the optimization scheduling model, select the power generation flow of large reservoirs as the decision variable, and propose that the minimum flow of river ecology and the maximum flow capacity of hydropower station are respectively the power generation flow Q t The lower and upper bounds of :

[0047] in: Q t They are the upper and lower limits of power generation flow in period t, in m 3 / s;

[0048] S84, establish the constraints of the optimal scheduling model, including water balance constraints, water level and storage capacity constraints, tailwater level discharge constraints, head constraints, hydropower generation constraints, water loss constraints, and reservoir water level constraints;

[0049] S85, establish the end-of-period water storage benefits, e.g. Figure 3 The steps shown are as follows: based on the analysis of reservoir water inflow data over the past thirty years, one or two years are selected as a cycle, the water inflow frequency corresponding to the annual water inflow is calculated, and a reservoir optimization scheduling model with the maximum comprehensive benefit is established. The normal operating water level of the discrete reservoir is calculated annually according to the set water level, the reservoir water inflow over the past thirty years, and the domestic and industrial water consumption and water demand, agricultural water consumption and water demand. The annual water inflow frequency is regarded as the weight of the annual optimization calculation scheme, and the weighted average is used to obtain the corresponding water storage benefit under the set water level; by setting different water levels, the above calculation is repeated to obtain the water storage benefits corresponding to different water levels; the discrete points are subjected to regression analysis to obtain the functional relationship between water level and water storage benefit.

[0050] F2=f(Z)

[0051] Among them, F2 is the water storage benefit of the reservoir, in ten thousand yuan; Z is the water level of the reservoir, in meter.

[0052] In the above technical solution, the objective function in step S82 is specifically:

[0053] (1) Objective function for maximizing the comprehensive benefits of the reservoir:

[0054]

[0055] E t =N t ×Δt h / 10000

[0056] Among them, maxF1 is the maximum comprehensive benefit of the reservoir, in ten thousand yuan; t and T are the time period number and the total number of time periods respectively; is the electricity price during period t, in RMB / kWh; are the water prices for life, industry and agriculture in period t, in yuan / m 3 ;E t is the power generation in period t, in 10,000 kW; are the water consumption for life, industry and agriculture in period t, in ten thousand m 3 . N t is the output of the hydropower station during period t, in kW; Δt h The number of hours in period t, in h.

[0057] (2) The objective function for maximizing the reservoir's final water storage benefit:

[0058] max F 2=f(Z T )

[0059] Among them, maxF2 is the maximum water storage benefit of the reservoir at the end of the period, in ten thousand yuan; Z T is the water level at the end of period T, in m; f(Z T ) is the relationship between the water level at the end of the period and the water storage benefit at the end of the period.

[0060] (3) Objective function for minimizing the amount of water discharged from the reservoir:

[0061]

[0062] Among them, W t q is the amount of water discarded during period t, W q The total amount of water discarded during the period, in ten thousand m 3 .

[0063] In the above technical solution, the constraint conditions in step S84 are as follows:

[0064] (1) Water balance constraints:

[0065]

[0066]

[0067]

[0068]

[0069]

[0070] Among them, V t is the reservoir capacity at the end of period t, in ten thousand m 3 ;I t The water inflow during period t, in ten thousand m 3 ; is the water consumption for power generation during period t, in ten thousand m3 ;loss t is the water loss during period t, in ten thousand m 3 ; They are respectively the water demand for life, industry and agriculture in period t, in ten thousand m 3 ; α1, α2, α3 are the demand satisfaction coefficients of domestic, industrial and agricultural water use respectively; Q t is the water consumption for power generation during period t, in ten thousand m 3 .

[0071] (2) Water level and storage capacity constraints:

[0072] Z t =f(V t )

[0073] Among them, Z t is the reservoir water level at the end of period t, in m.

[0074] (3) Tailwater level discharge constraint:

[0075] Zd t =f(Q t )

[0076] Among them, Zd t is the downstream water level of the reservoir at the end of period t, in meters.

[0077] (4) Head constraint:

[0078]

[0079]

[0080]

[0081] in, H is the average upstream and downstream water levels of the reservoir during period t, in m; t is the hydraulic head of the reservoir during period t, in m.

[0082] (5) Hydropower output constraints:

[0083] N t =KQ t H t

[0084] Among them, K is the comprehensive output coefficient of the hydropower station.

[0085] (6) Water loss constraints:

[0086]

[0087]

[0088]

[0089]

[0090] in, is the average storage capacity during period t, in ten thousand m; is the average area of ​​the reservoir during period t, in km 2 ; γ is the evaporation intensity of the reservoir, unit is mm; α and λ are the evaporation and leakage coefficients respectively.

[0091] (7) Reservoir boundary conditions:

[0092]

[0093] in, Z t are the upper and lower limits of the reservoir water level in period t, respectively, in m.

[0094] In the above technical solution, step S9 includes:

[0095] S91: The multi-objective function of maximizing the comprehensive benefit of the reservoir, maximizing the end-of-period water storage benefit, and minimizing the amount of abandoned water is converted into a single-objective function. With maximizing the comprehensive benefit of the reservoir as the goal, the constraint of minimizing the amount of abandoned water is converted into a maximum constraint. The comprehensive benefit of the reservoir and the end-of-period water storage benefit are weighted as equally important and converted into a single-objective function of maximizing the reservoir benefit:

[0096] (1) The minimum target of reservoir water discharge is converted into constraints:

[0097] W q ≤ε

[0098] Among them, ε is the maximum amount of water allowed to be abandoned, in units of 10,000 m 3 .

[0099] (2) The multiple objectives of maximizing the comprehensive benefits of the reservoir and the water storage benefits at the end of the period are transformed into single objectives of equal importance:

[0100]

[0101] Among them, B t is the comprehensive benefit of the reservoir in period t, in ten thousand yuan; B mo is the comprehensive benefit at the end of period T, which is expressed as the functional relationship between the water storage benefit at the end of the period and the water level at the end of the period.

[0102] S92, a genetic algorithm is used to solve the reservoir optimization scheduling model in multiple ways, and finally multiple optimal allocation schemes are obtained.

[0103] The beneficial effects of the present invention are: the present invention comprehensively considers the water resource supply and receiving ends, and when it is difficult to achieve unified management of small and medium-sized reservoirs and ponds in the irrigation area, the water resource optimization allocation problem is converted into an optimization scheduling problem of large reservoirs (steps S1-S7), and the total amount of domestic, industrial and agricultural water demand that needs to be taken from large reservoirs in the irrigation area is predicted. Taking into account the uncertainty of water inflow from large reservoirs, a reasonable allocation plan is formulated within the year to improve the utilization rate of regional and reservoir water resources as a whole. It is suitable for regional water resource allocation scenarios with large reservoirs as control reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 Flowchart of the present invention;

[0105] Figure 2 This is a flow chart of the water balance pretreatment of water sources and water used in the present invention;

[0106] Figure 3 This is a flow chart for calculating the water level-optimal benefit relationship in the present invention. DETAILED DESCRIPTION

[0107] The present invention will be further described in detail below with reference to the accompanying drawings and examples, which are not intended to limit the present invention but are merely examples. The advantages of the present invention will become clearer and easier to understand through the description.

[0108] The inventive concept of the present invention is that the setting of the reservoir water level at the end of the scheduling period and the benefits of the scheduling period are mutually constrained. When allocating reservoir water resources, it is not only necessary to achieve efficient utilization of water resources within the scheduling period, but also to consider the reasonable setting of the reservoir water level at the end of the scheduling period. However, the accuracy of medium- and long-term water inflow forecasts for reservoir operation and management is low, making it difficult to meet the requirements for formulating water resource scheduling plans. Therefore, for the optimal allocation of water resources in large reservoir-irrigation areas, the present invention transforms the water resource optimization problem into the optimal scheduling problem of large reservoirs, and considers the uncertainty of reservoir water inflow forecasts, thereby improving the efficient utilization of water resources and reducing the risks caused by runoff forecast errors.

[0109] like Figure 1 As shown, the present invention provides a multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty. Specifically, it includes:

[0110] S1, divide the irrigation area into different areas and collect data on domestic, industrial and agricultural water consumption within the irrigation area in the past five years;

[0111] S2, using linear regression method to predict the domestic and industrial water demand in different areas within the irrigation area;

[0112] S3, using the irrigation quota method to predict the agricultural water demand of different areas within the irrigation area;

[0113] S4, according to the water balance pre-processing method, the water supply and water demand of multiple water sources in different areas of the irrigation area are pre-balanced to obtain the monthly water shortage of different areas in the irrigation area, which is used as the domestic and industrial water demand and agricultural water demand that need to be taken from large reservoirs; that is, water balance pre-processing is first performed on the irrigation area ponds, small reservoirs and medium-sized reservoirs in the irrigation area to obtain the monthly water shortage of each area in the irrigation area, which is used as the water demand from large reservoirs. Figure 2 shown.

[0114] S5, using linear regression to predict domestic and industrial water demand from large reservoirs;

[0115] S6, using SVM deep learning method to predict agricultural water demand that needs to be drawn from large reservoirs;

[0116] S7, averaging the domestic and industrial water demands that need to be taken from large reservoirs predicted by the water balance preprocessing method and the linear regression method, and averaging the agricultural water demands predicted by the water balance preprocessing method and the SVM deep learning method, to obtain the final domestic and industrial water demands that need to be taken from large reservoirs, and the agricultural water demands;

[0117] S8, considering the uncertainty of runoff forecast, establish a water resources optimization model with the maximum comprehensive benefits of large reservoirs, the maximum end-of-period water storage benefits, and the minimum amount of abandoned water as multi-objective functions;

[0118] S9, transform the multi-objective function into a single-objective function, use the genetic algorithm to solve the model, and obtain the optimal allocation plan for large reservoirs.

[0119] Taking the annual water resource allocation for Reservoir X in Irrigation District Y as an example, Reservoir X is a Class I Large (I) reservoir with multi-year regulation capacity. Its primary function is irrigation, with additional functions such as flood control, urban water supply, and power generation. The designed irrigation area is 2.6 million mu. The proposed method is used to implement multi-objective water resource allocation that takes into account forecast uncertainty.

[0120] (1) In the embodiment of the present invention, Figure 1 As shown, step S1 includes:

[0121] 1) Divide the irrigation area into six zones based on the main canals of the water supply channels in irrigation area Y, and collect and organize data on domestic, industrial, and agricultural water consumption;

[0122] 2) Calculate the actual domestic, industrial and agricultural water consumption in each area in each month and year.

[0123] (2) In the embodiment of the present invention, Figure 1 As shown, step S2 includes:

[0124] 1) Based on the annual water consumption data of the past five years collected in step S1, a linear regression calculation is performed on the annual domestic and industrial water demand of each area within the irrigation district to obtain the predicted domestic and industrial water demand of each area; the agricultural water demand of different areas within the irrigation district is predicted using the irrigation quota method;

[0125] 2) Based on the monthly domestic and industrial water consumption of each area in recent years, calculate the average proportion of each month in the annual domestic and industrial water consumption, and calculate the monthly domestic and industrial water demand of different areas.

[0126] (3) In the embodiment of the present invention, Figure 1 As shown, step S3 includes:

[0127] 1) Irrigation District Y is mainly planted with mid-season rice. Considering that the irrigation district is too large and the corresponding time periods of the same growth cycle in each rice field are different, the irrigation quotas for each growth stage of rice in each district were collected from the irrigation experimental station under rainfall frequencies of 25%, 50%, and 75%. The growth stages include the soaking stage, the greening and transplanting stage, the early tillering stage, the late tillering stage, the jointing and booting stage, the heading and flowering stage, the milky yellow stage, and the yellow ripe stage.

[0128] 2) Collect information on crop types, planting areas, crop growth cycles, irrigation canal utilization coefficients, and field water utilization coefficients for each area within irrigation district Y;

[0129] 3) Calculate the agricultural water demand D for each area in irrigation zone Y at each growth stage using the following formula:

[0130]

[0131] Among them, A is the crop irrigation quota, S is the crop planting area, L0 is the field water utilization coefficient, and L1 is the irrigation canal utilization coefficient.

[0132] 4) Calculate the agricultural water demand for each area in each month according to the corresponding time of each growth stage.

[0133] (4) In the embodiment of the present invention, Figure 2 As shown, step S4 includes:

[0134] S41, collects the water supply of each water source in the irrigation area in the past ten years, divides it into wet, normal and dry seasons, and combines it with the water inflow characteristics of the predicted annual water inflow to obtain the predicted water supply for each month.

[0135] S42, based on the domestic and industrial water demand and agricultural water demand of each area in each month calculated in step S2 and step S3, and the available water volume calculated in step S41, perform water balance calculation according to the water supply priority of domestic, industrial and industrial water types, and obtain the domestic, industrial and agricultural water shortages in each area. The domestic, industrial and agricultural water shortages for each month are statistically obtained as the domestic, industrial and agricultural water volume that needs to be taken from large reservoirs.

[0136] (5) In the embodiment of the present invention, Figure 1 As shown, step S5 includes:

[0137] 1) Collect and compile monthly domestic and industrial water consumption data for large reservoirs over the past five years;

[0138] 2) Perform linear regression calculations on the annual domestic and industrial water demands of large reservoirs to obtain predicted domestic and industrial water demands;

[0139] 3) Based on the monthly domestic and industrial water consumption in recent years, calculate the average proportion of each month in the annual domestic and industrial water consumption, and calculate the domestic and industrial water demand for each month.

[0140] (6) In the embodiment of the present invention, Figure 1 As shown, step S6 includes:

[0141] 1) Collect and compile data on the irrigated area in the irrigation district over the past ten years, and predict the irrigated area Y in each month of the year, the agricultural water consumption in each month of the irrigation district, and the agricultural water supply from large reservoirs in each month;

[0142] 2) Taking the monthly irrigation area and agricultural water consumption in the irrigation district as influencing factors, the historical data were used to establish a SVM deep learning model for each month for training. The predicted annual agricultural water supply required from large reservoirs was calculated to meet the agricultural water demand requirements.

[0143] (7) In the embodiment of the present invention, Figure 1 As shown, step S7 includes:

[0144] 1) Taking the monthly weighted average of the domestic and industrial water demands calculated in steps S4 and S5, the final domestic and industrial water supply required to be drawn from large reservoirs is obtained to meet the domestic and industrial water demand requirements;

[0145] 2) Calculate the proportion of agricultural water demand that needs to be drawn from large reservoirs in each month according to step 4, and calculate the proportion of agricultural water demand that needs to be drawn from large reservoirs in each month according to step 6. Average the two predicted proportions for each month to obtain the final agricultural water demand that needs to be drawn from large reservoirs.

[0146] The present invention predicts that the domestic water consumption and industrial water consumption of Reservoir X are 86.38 million m 3 48.77 million m 3 Using the water inflow and irrigation quota when the rainfall frequency is P = 25%, P = 50% and P = 75%, it is predicted that the agricultural water demand of reservoir X is 230.15 million m 3 , 249.92 million m 3 264.2 million m 3 .

[0147] (8) In the embodiment of the present invention, Figure 1 As shown, step S8 includes:

[0148] 1) Collect basic information on Reservoir X, including domestic water prices, industrial water prices, agricultural water prices, monthly average electricity prices, water data from the past 30 years, water level and storage capacity relationship curves, tailwater level and discharge relationship curves, evaporation and leakage coefficients, and average hydropower station output coefficients;

[0149] 2) Establish a water resources optimization model with the multi-objective function of maximizing the comprehensive benefits of large reservoirs, maximizing the end-of-period water storage benefits, and minimizing the amount of abandoned water. Establish the multi-objective function of the optimization scheduling model as follows:

[0150] a) Objective function for maximizing the comprehensive benefits of the reservoir:

[0151]

[0152] E t =N t ×Δt h / 10000

[0153] Among them, maxF1 is the maximum comprehensive benefit of the reservoir, in ten thousand yuan; t and T are the time period number and the total number of time periods respectively; is the electricity price during period t, in RMB / kWh; are the water prices for life, industry and agriculture in period t, in yuan / m 3 ;E t is the power generation in period t, in 10,000 kW; are the water consumption for life, industry and agriculture in period t, in ten thousand m 3 . N t is the output of the hydropower station during period t, in kW; Δt h The number of hours in period t, in h.

[0154] b) The objective function for maximizing the reservoir's final water storage benefit:

[0155] max F2=f(Z T )

[0156] Among them, maxF2 is the maximum water storage benefit of the reservoir at the end of the period, in ten thousand yuan; Z T is the water level at the end of period T, in m; f(Z T ) is the relationship between the water level at the end of the period and the water storage benefit at the end of the period.

[0157] c) Minimum objective function of reservoir water discharge:

[0158]

[0159] Among them, W t q 、W q They are the amount of water discarded in period t and the total amount of water discarded in the period, in units of 10,000 m 3 .

[0160] 3) Establish the decision variables of the optimization scheduling model, select the power generation flow of large reservoirs as the decision variable, and propose the minimum flow of river ecology and the maximum flow capacity of hydropower station as the power generation flow Q t The lower and upper bounds of :

[0161]

[0162] in, Q t They are the upper and lower limits of power generation flow in period t, in m 3 / s;

[0163] 4) Establish the constraints of the optimal scheduling model, including water balance constraints, water level and storage capacity constraints, tailwater level and discharge constraints, head constraints, hydropower generation constraints, water loss constraints, and reservoir water level constraints;

[0164] a) Water balance constraints:

[0165]

[0166]

[0167]

[0168]

[0169]

[0170] Among them, V t is the reservoir capacity at the end of period t, in ten thousand m 3 ;I t The water inflow during period t, in ten thousand m 3 ; is the water consumption for power generation during period t, in ten thousand m3 ;loss t is the water loss during period t, in ten thousand m 3 ; They are respectively the water demand for life, industry and agriculture in period t, in ten thousand m 3 ; α1, α2, α3 are the demand satisfaction coefficients of domestic, industrial and agricultural water use respectively; Q t is the water consumption for power generation during period t, in ten thousand m 3 .

[0171] b) Water level and storage capacity constraints:

[0172] Z t =f(V t )

[0173] Among them, Z t is the reservoir water level at the end of period t, in m.

[0174] c) Tailwater level discharge constraint:

[0175] Zd t =f(Q t )

[0176] Among them, Zd t is the downstream water level of the reservoir at the end of period t, in meters.

[0177] d) Head constraint:

[0178]

[0179]

[0180]

[0181] in, H is the average upstream and downstream water levels of the reservoir during period t, in m; t is the hydraulic head of the reservoir during period t, in m.

[0182] e) Hydropower output constraints:

[0183] N t =KQ t H t

[0184] Among them, K is the comprehensive output coefficient of the hydropower station.

[0185] f) Water loss constraints:

[0186]

[0187]

[0188]

[0189]

[0190] in, is the average storage capacity during period t, in ten thousand m 3 ; is the average area of ​​the reservoir during period t, in km 2 ; γ is the evaporation intensity of the reservoir, unit is mm; α and λ are the evaporation and leakage coefficients respectively.

[0191] g) Reservoir boundary conditions:

[0192]

[0193] in, Z t are the upper and lower limits of the reservoir water level in period t, respectively, in m.

[0194] 5) If Figure 3 As shown in , the end-of-period water storage benefit is established. The end-of-period water storage benefit is the comprehensive benefit within a certain period after the end of the scheduling period. For multi-year regulation reservoirs, a two-year period is selected based on the reservoir inflow data analysis of the past thirty years. The long sequence analysis method is used to calculate the water inflow frequency corresponding to the annual water inflow, and establish an optimal scheduling model for the reservoir with the maximum comprehensive benefit. The normal operating water level of the discrete reservoir is calculated according to the set water level, the reservoir inflow in the past thirty years and the water demand for domestic, industrial and agricultural water use. The maximum comprehensive benefit of the reservoir at this water level is calculated each year. The annual water inflow frequency is regarded as the weight of the annual optimization calculation plan, and the weighted average is used to obtain the corresponding water storage benefit under the set water level. By setting different water levels, repeating the above calculation to obtain the water storage benefits corresponding to different water levels, and the discrete points are subjected to regression analysis to obtain the functional relationship between water level and water storage benefit.

[0195] F2=f(Z)

[0196] Among them, F2 is the water storage efficiency of the reservoir; Z is the water level of the reservoir.

[0197] (9) In the embodiment of the present invention, Figure 1 As shown, step S8 includes:

[0198] 1) The multi-objective function of maximizing the comprehensive benefit of the reservoir, maximizing the end-of-period water storage benefit, and minimizing the amount of abandoned water is converted into a single-objective function. With the maximization of the comprehensive benefit of the reservoir as the goal, the constraint of minimizing the amount of abandoned water is converted into a maximum constraint. The comprehensive benefit of the reservoir and the end-of-period water storage benefit are weighted with equal importance and converted into a single-objective function of maximizing the reservoir benefit:

[0199] a) The minimum amount of water discharged from the reservoir is converted into constraints:

[0200] W q ≤ε

[0201] Among them, ε is the maximum amount of water allowed to be abandoned, in units of 10,000 m 3 .

[0202] b) The multiple objectives of maximizing the comprehensive benefits of the reservoir and the water storage benefits at the end of the period are converted into single objectives of equal importance:

[0203]

[0204] Among them, B t is the comprehensive benefit of the reservoir in period t, in ten thousand yuan; B mo is the comprehensive benefit at the end of period T, which is expressed as the functional relationship between the water storage benefit at the end of the period and the water level at the end of the period.

[0205] 2) Select the water inflow when the rainfall frequency is P = 25%, P = 50% and P = 75%, and use the genetic algorithm to solve the reservoir optimization operation model. Finally, the optimal allocation scheme of reservoir X under various rainfall frequencies is obtained, as follows:

[0206] a) When the rainfall frequency P = 25%, the benefit during the period is 45.62 million yuan, the water level at the end of the period is 119.01m, and the comprehensive benefit is 116.90 million yuan. Compared with the scheduling method using the scheduling chart, the water level at the end of the period is 2m lower, and the comprehensive benefit increases by 610,000 yuan;

[0207] b) When the rainfall frequency P = 50%, the benefit during the period is 40.84 million yuan, the water level at the end of the period is 117.58m, and the comprehensive benefit is 108.54 million yuan. Compared with the scheduling method using the scheduling diagram, the water level at the end of the period is 0.95m lower, and the comprehensive benefit increases by 300,000 yuan;

[0208] c) When the rainfall frequency P = 75%, the benefit during the period is 37.75 million yuan, and the water level at the end of the period is 119.01m, with a comprehensive benefit of 116.90 million yuan. Compared with the scheduling method using the scheduling diagram, the water level at the end of the period is raised by 0.25m, and the comprehensive benefit increases by 130,000 yuan.

[0209] The working principle and process of the present invention are as follows: the present invention separates the irrigation area and the reservoir, considering both the supply and the demand sides. First, the irrigation area is divided into zones; the water inflow and water demand within the irrigation area are predicted and water balance preprocessing is performed; then, the reservoir's historical water supply over many years is analyzed and predicted using statistical analysis methods; the predicted water demand from the two prediction methods is weighted averaged as the reservoir's water supply; considering the uncertainty of reservoir runoff forecasts, a water resource optimization model is established with multiple objective functions, namely, maximizing the comprehensive benefits of large reservoirs, maximizing the end-of-period water storage benefits, and minimizing the amount of abandoned water; then, the multi-objective function is converted into a single objective function, and the model is solved using a genetic algorithm to obtain the optimal allocation plan for large reservoirs.

[0210] Other parts not described in detail are prior art.

Claims

1. A multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty, characterized by: The steps include: S1, divide the irrigation area into different areas and collect data on domestic, industrial and agricultural water consumption within the irrigation area every year for the past five years; S2, using linear regression method to predict the domestic and industrial water demand in different areas within the irrigation area; S3, using the irrigation quota method to predict the agricultural water demand of different areas within the irrigation area; S4, using the water balance pre-processing method to pre-balance the water supply and water demand of multiple water sources in different areas within the irrigation area, and obtain the monthly water shortage of different areas within the irrigation area, which is used as the domestic and industrial water demand and agricultural water demand that need to be taken from large reservoirs; S5, using linear regression to predict domestic and industrial water demand from large reservoirs; S6, using SVM deep learning method to predict agricultural water demand that needs to be drawn from large reservoirs; S7, averaging the domestic and industrial water demands that need to be taken from large reservoirs predicted by the water balance preprocessing method and the linear regression method, and averaging the agricultural water demands predicted by the water balance preprocessing method and the SVM deep learning method, to obtain the final domestic and industrial water demands that need to be taken from large reservoirs, and the agricultural water demands; S8, considering the uncertainty of runoff forecast, establish a water resources optimization model with the maximum comprehensive benefits of large reservoirs, the maximum end-of-period water storage benefits, and the minimum amount of abandoned water as multi-objective functions; The step S8 comprises: S81: Collect basic reservoir data, including domestic water prices, industrial water prices, agricultural water prices, monthly average electricity prices, water data from the past 30 years, water level and storage capacity relationship curves, tailwater level and discharge relationship curves, evaporation and leakage coefficients, and average hydropower station output coefficients; S82: Establish the multi-objective function of the optimization scheduling model, and establish a water resources optimization scheduling model with the multi-objective function of maximizing the comprehensive benefits of large reservoirs, maximizing the end-of-period water storage benefits, and minimizing the amount of abandoned water; S83, establish the decision variables of the optimization scheduling model, select the power generation flow of large reservoirs as the decision variable, and propose that the minimum flow of river ecology and the maximum flow capacity of hydropower station are respectively the power generation flow Q t The lower and upper bounds of : in: Q t They are the upper and lower limits of power generation flow in period t, in m 3 / s; S84, establish the constraints of the optimal scheduling model, including water balance constraints, water level and storage capacity constraints, tailwater level discharge constraints, head constraints, hydropower generation constraints, water loss constraints, and reservoir water level constraints; S85: Establish the end-of-period water storage benefits. Based on an analysis of reservoir inflow data over the past thirty years, select a one- or two-year cycle, calculate the water inflow frequency corresponding to the annual water inflow, and establish a reservoir optimization scheduling model for maximum comprehensive benefits. Discrete the normal operating water level of the reservoir. Based on the set water level, reservoir inflow over the past thirty years, domestic and industrial water consumption and demand, and agricultural water consumption and demand, calculate the maximum comprehensive benefit of the reservoir at that water level each year. Treat the annual water inflow frequency as the weight of the annual optimization calculation scheme, and calculate the corresponding water storage benefit at the set water level using a weighted average. Repeat the above calculations to obtain the water storage benefits corresponding to different water levels by setting different water levels. Use regression analysis on the discrete points to obtain the functional relationship between water level and water storage benefit. F2=f(Z) Among them, F2 is the water storage benefit of the reservoir, in ten thousand yuan; Z is the water level of the reservoir, in meters; S9, transform the multi-objective function into a single-objective function, use the genetic algorithm to solve the model, and obtain the optimal allocation plan for large reservoirs.

2. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1 is characterized in that: The step S1 comprises: S11, divide the irrigation area into several areas according to water supply facilities, and collect and organize data on domestic and industrial water withdrawal; S12, calculate the monthly domestic and industrial water consumption of each area over the past five years.

3. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1 or 2, characterized in that: Step S2 includes: S21, based on the annual water consumption data for the past five years collected in step S1, perform a linear regression calculation on the total annual domestic and industrial water demand of each area within the irrigation area to obtain a predicted total domestic and industrial water demand of each area; S22. Taking into account the cyclical nature of monthly domestic and industrial water consumption, the average monthly proportion of annual domestic and industrial water consumption is calculated based on the monthly domestic and industrial water consumption of each area in the past three years. Based on the total domestic and industrial water demand of each area calculated in step S21, the monthly domestic and industrial water demand of each area is calculated.

4. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 3 is characterized in that: Step S3 includes: S31. Collect irrigation quota data for each growth stage of early rice, mid-season rice, and late rice during the entire growth cycle under conditions of rainfall frequencies of 25%, 50%, and 75%, as summarized by the irrigation experimental station through experimental analysis. The growth stages include the field soaking stage, the greening and transplanting stage, the early tillering stage, the late tillering stage, the jointing and booting stage, the heading and flowering stage, the milk-yellow stage, and the yellow-ripening stage. S32, collects the main crop types in each area of ​​the irrigation district, including early rice, mid-season rice, late rice, planting area, time range corresponding to the crop growth stage, irrigation canal utilization coefficient, and field water utilization coefficient; S33, calculate the agricultural water demand D at each growth stage in different areas within the irrigation area. The calculation formula is as follows: Among them, A is the crop irrigation quota, S is the crop planting area, L0 is the field water utilization coefficient, and L1 is the irrigation canal utilization coefficient; S34, calculating the monthly agricultural water demand of different areas according to the corresponding time of each growth stage in different areas.

5. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 4 is characterized in that: The step S4 comprises: S41: Collect water supply data from multiple water sources in the irrigation area for the past ten years, classify them into wet, normal, and dry periods, and combine them with the water inflow characteristics of the forecast period to obtain the predicted monthly water supply; S42. Based on the monthly domestic and industrial water demand and agricultural water demand of each area calculated in steps S2 and S3, and the available water supply calculated in step S41, a water balance calculation is performed according to the water supply priority of domestic and industrial water demand and agricultural water demand, and the domestic and industrial water shortage and agricultural water shortage of each area are obtained. The total monthly water shortage is calculated as the domestic and industrial water demand and agricultural water demand that need to be taken from large reservoirs.

6. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1 is characterized in that: The step S5 comprises: S51: Collect and compile monthly domestic and industrial water consumption data for large reservoirs over the past five years; S52, perform linear regression calculation on the annual domestic and industrial water demand of large reservoirs to obtain the predicted domestic and industrial water demand; S53. Based on the monthly domestic and industrial water consumption in the past three years, calculate the average proportion of each month to the annual domestic and industrial water consumption, and calculate the monthly domestic and industrial water demand.

7. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1 is characterized in that: The step S6 comprises: S61: Collect and compile data on the irrigated area of ​​irrigation districts over the past ten years, the irrigated area of ​​irrigation districts per month in the forecast year, the monthly agricultural water consumption of irrigation districts, and the monthly agricultural water supply of large reservoirs; S62, taking the monthly irrigation area and monthly agricultural water consumption in the irrigation area as influencing factors, uses data from the past ten years to train a SVM deep learning model for each month, and calculates the monthly agricultural water supply that needs to be taken from large reservoirs in the predicted year.

8. The multi-objective water resource allocation method for large reservoir irrigation areas based on forecast uncertainty according to claim 1 is characterized in that: The objective function in step S82 is specifically: (1) Objective function for maximizing the comprehensive benefits of the reservoir: E t =n t ×Δt h / 10000 Among them, maxF1 is the maximum comprehensive benefit of the reservoir, in ten thousand yuan; t and T are the time period number and the total number of time periods respectively; is the electricity price during period t, in RMB / kWh; are the water prices for life, industry and agriculture in period t, in yuan / m 3 ;E t is the power generation in period t, in ten thousand kWh; W t g 、 are the water consumption for life, industry and agriculture in period t, in ten thousand m 3 ; N t is the output of the hydropower station during period t, in kW; Δt h is the number of hours in period t, in h; (2) The objective function for maximizing the reservoir's final water storage benefit: maxF 2=f(Z T ) Among them, maxF2 is the maximum water storage benefit of the reservoir at the end of the period, in ten thousand yuan; Z T is the water level at the end of period T, in m; f(Z T ) is the relationship between the water level at the end of the period and the water storage efficiency at the end of the period; (3) Objective function for minimizing the amount of water discharged from the reservoir: Among them, W t q is the amount of water discarded during period t, W q The total amount of water discarded during the period, in ten thousand m 3 .

9. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1 or 8, characterized in that: The constraints in step S84 are as follows: (1) Water balance constraints: Among them, V t is the reservoir capacity at the end of period t, in ten thousand m 3 ;I t The water inflow during period t, in ten thousand m 3 ; is the water consumption for power generation during period t, in ten thousand m 3 ;loss t is the water loss during period t, in ten thousand m 3 ; They are respectively the water demand for life, industry and agriculture in period t, in ten thousand m 3 ; α1, α2, α3 are the demand satisfaction coefficients of domestic, industrial and agricultural water use respectively; Q t is the power generation flow in period t, in m 3 / s; (2) Water level and storage capacity constraints: With t =f(V t ) Among them, Z t is the reservoir water level at the end of period t, in m; (3) Tailwater level discharge constraint: Zd t =f(Q t ) Among them, Zd t is the downstream water level of the reservoir at the end of period t, in meters; (4) Head constraint: in, H is the average upstream and downstream water levels of the reservoir during period t, in m; t is the hydraulic head of the reservoir during period t, in m; (5) Hydropower output constraints: N t Result t H t Among them, K is the comprehensive output coefficient of the hydropower station; (6) Water loss constraints: in, is the average storage capacity during period t, in ten thousand m 3 ; is the average area of ​​the reservoir during period t, in km 2 ;γ is the evaporation intensity of the reservoir, unit is mm; α and λ are the evaporation and leakage coefficients respectively; (7) Boundary conditions: in, Z t They are the upper and lower limits of the water level of the reservoir at section t, in meters.

10. The multi-objective water resource allocation method for large irrigation areas based on forecast uncertainty according to claim 1, characterized in that: The step S9 includes: S91: The multi-objective function of maximizing the comprehensive benefit of the reservoir, maximizing the end-of-period water storage benefit, and minimizing the amount of abandoned water is converted into a single-objective function. With maximizing the comprehensive benefit of the reservoir as the goal, the constraint of minimizing the amount of abandoned water is converted into a maximum constraint. The comprehensive benefit of the reservoir and the end-of-period water storage benefit are weighted as equally important and converted into a single-objective function of maximizing the reservoir benefit: (1) The minimum target of reservoir water discharge is converted into constraints: IN q ≤ε Among them, ε is the maximum amount of water allowed to be abandoned, in units of 10,000 m 3 ; (2) The multiple objectives of maximizing the comprehensive benefits of the reservoir and the water storage benefits at the end of the period are transformed into single objectives of equal importance: Among them, B t is the comprehensive benefit of the reservoir in period t, in ten thousand yuan; B mo is the comprehensive benefit at the end of period T, expressed as the functional relationship between the water storage benefit at the end of the period and the water level at the end of the period; S92, a genetic algorithm is used to solve the reservoir optimization scheduling model in multiple ways, and finally multiple optimal allocation schemes are obtained.

Images